id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.03305 | Yoshita Baruah | Yoshita Baruah | Geomagnetic Storms and Satellite Orbital Decay | 7 pages, 2 figures, IAUS-390 Proceedings | null | null | null | astro-ph.SR astro-ph.EP physics.space-ph | http://creativecommons.org/licenses/by/4.0/ | Energetic events on the Sun, particularly coronal mass ejections and high speed streams, regulate the near Earth space environment and give rise to space weather. A major terrestrial manifestation of such events are geomagnetic storms. A geomagnetic storm results in dissipation of energy from the solar wind into the atmosphere, leading to Joule heating and thermospheric expansion. This has serious consequences on Low Earth Orbit satellite lifetimes. Our work demonstrates the impact of different kinds of geomagnetic storms on satellite orbits. We also briefly discuss about some physical attributes of satellites that can make them prone to higher orbital decay. Our work highlights the importance of monitoring and predicting space weather, and assessing their impacts on space-based human technologies. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:45:11 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Baruah",
"Yoshita",
""
]
] |
2506.03306 | Santosh Harish | Santosh Harish, Jeyhan S. Kartaltepe, Daizhong Liu, Anton M. Koekemoer, Caitlin M. Casey, Maximilien Franco, Hollis B. Akins, Olivier Ilbert, Marko Shuntov, Nicole E. Drakos, Mike Engesser, Andreas L. Faisst, Ghassem Gozaliasl, Crystal L. Martin, Michaela Hirschmann, Vasily Kokorev, Erini Lambrides, Henry Joy McCracken, Jed McKinney, Louise Paquereau, Jason Rhodes and Brant E. Robertson | COSMOS-Web: MIRI Data Reduction and Number Counts at 7.7$\mu$m using JWST | Submitted to ApJ, comments are welcome! | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | The COSMOS-Web survey is the largest JWST Cycle 1 General Observer program covering a contiguous ~0.54 deg$^2$ area with NIRCam imaging in four broad-band filters and a non-contiguous ~0.2 deg$^2$ with parallel MIRI imaging in a single broad-band filter, F770W. Here we present a comprehensive overview of the MIRI imaging observations, the data reduction procedure, the COSMOS-Web MIRI photometric catalog, and the first data release including the entire COSMOS-Web MIRI coverage. Data reduction is predominantly based on the JWST Science Calibration Pipeline with an additional step involving custom background subtraction to mitigate the presence of strong instrumental features and sky background in the MIRI images. We reach 5$\sigma$ (point source) limiting depths ($m_{F770W}$~25.51 based on $r$~0.3'' circular apertures) that are significantly better than initial expectations. We create a COSMOS-Web MIRI catalog based on the images presented in this release and compare the F770W flux densities with the Spitzer/IRAC CH4 measurements from the COSMOS2020 catalog for CH4 detections with S/N $>5$. We find that these are in reasonable agreement with a small median offset of $<0.05$ mag. We also derive robust 7.7$\mu$m number counts spanning five orders of magnitude in flux ($\sim$0.2-2300 $\mu$Jy) $\unicode{x2013}$ making COSMOS-Web the only JWST survey to date to efficiently sample such a large flux range $\unicode{x2013}$ which is in good agreement with estimates from other JWST and IRAC surveys. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:46:23 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Harish",
"Santosh",
""
],
[
"Kartaltepe",
"Jeyhan S.",
""
],
[
"Liu",
"Daizhong",
""
],
[
"Koekemoer",
"Anton M.",
""
],
[
"Casey",
"Caitlin M.",
""
],
[
"Franco",
"Maximilien",
""
],
[
"Akins",
"Hollis B.",
""
],
[
"Ilbert",
"Olivier",
""
],
[
"Shuntov",
"Marko",
""
],
[
"Drakos",
"Nicole E.",
""
],
[
"Engesser",
"Mike",
""
],
[
"Faisst",
"Andreas L.",
""
],
[
"Gozaliasl",
"Ghassem",
""
],
[
"Martin",
"Crystal L.",
""
],
[
"Hirschmann",
"Michaela",
""
],
[
"Kokorev",
"Vasily",
""
],
[
"Lambrides",
"Erini",
""
],
[
"McCracken",
"Henry Joy",
""
],
[
"McKinney",
"Jed",
""
],
[
"Paquereau",
"Louise",
""
],
[
"Rhodes",
"Jason",
""
],
[
"Robertson",
"Brant E.",
""
]
] |
2506.03307 | Kristen Goebel | Kristen Goebel, William Solow, Paola Pesantez-Cabrera, Markus Keller, Alan Fern | Budgeted Online Active Learning with Expert Advice and Episodic Priors | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces a novel approach to budgeted online active learning from finite-horizon data streams with extremely limited labeling budgets. In agricultural applications, such streams might include daily weather data over a growing season, and labels require costly measurements of weather-dependent plant characteristics. Our method integrates two key sources of prior information: a collection of preexisting expert predictors and episodic behavioral knowledge of the experts based on unlabeled data streams. Unlike previous research on online active learning with experts, our work simultaneously considers query budgets, finite horizons, and episodic knowledge, enabling effective learning in applications with severely limited labeling capacity. We demonstrate the utility of our approach through experiments on various prediction problems derived from both a realistic agricultural crop simulator and real-world data from multiple grape cultivars. The results show that our method significantly outperforms baseline expert predictions, uniform query selection, and existing approaches that consider budgets and limited horizons but neglect episodic knowledge, even under highly constrained labeling budgets. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:48:16 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Goebel",
"Kristen",
""
],
[
"Solow",
"William",
""
],
[
"Pesantez-Cabrera",
"Paola",
""
],
[
"Keller",
"Markus",
""
],
[
"Fern",
"Alan",
""
]
] |
2506.03309 | Song Zuo | Santiago Balseiro, Kshipra Bhawalkar, Yuan Deng, Zhe Feng, Jieming Mao, Aranyak Mehta, Vahab Mirrokni, Renato Paes Leme, Di Wang, Song Zuo | Position Auctions in AI-Generated Content | null | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider an extension to the classic position auctions in which sponsored creatives can be added within AI generated content rather than shown in predefined slots. New challenges arise from the natural requirement that sponsored creatives should smoothly fit into the context. With the help of advanced LLM technologies, it becomes viable to accurately estimate the benefits of adding each individual sponsored creatives into each potential positions within the AI generated content by properly taking the context into account. Therefore, we assume one click-through rate estimation for each position-creative pair, rather than one uniform estimation for each sponsored creative across all positions in classic settings. As a result, the underlying optimization becomes a general matching problem, thus the substitution effects should be treated more carefully compared to standard position auction settings, where the slots are independent with each other.
In this work, we formalize a concrete mathematical model of the extended position auction problem and study the welfare-maximization and revenue-maximization mechanism design problem. Formally, we consider two different user behavior models and solve the mechanism design problems therein respectively. For the Multinomial Logit (MNL) model, which is order-insensitive, we can efficiently implement the optimal mechanisms. For the cascade model, which is order-sensitive, we provide approximately optimal solutions. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:48:57 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Balseiro",
"Santiago",
""
],
[
"Bhawalkar",
"Kshipra",
""
],
[
"Deng",
"Yuan",
""
],
[
"Feng",
"Zhe",
""
],
[
"Mao",
"Jieming",
""
],
[
"Mehta",
"Aranyak",
""
],
[
"Mirrokni",
"Vahab",
""
],
[
"Leme",
"Renato Paes",
""
],
[
"Wang",
"Di",
""
],
[
"Zuo",
"Song",
""
]
] |
2506.03310 | Guillermo Marco Rem\'on | Guillermo Marco, Julio Gonzalo, V\'ictor Fresno | The Reader is the Metric: How Textual Features and Reader Profiles Explain Conflicting Evaluations of AI Creative Writing | Camera-ready version, 14 pages, 3 figures. Accepted to Findings of the Association for Computational Linguistics (ACL) 2025. Code & data: https://github.com/grmarco/the-reader-is-the-metric | null | null | null | cs.CL cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent studies comparing AI-generated and human-authored literary texts have produced conflicting results: some suggest AI already surpasses human quality, while others argue it still falls short. We start from the hypothesis that such divergences can be largely explained by genuine differences in how readers interpret and value literature, rather than by an intrinsic quality of the texts evaluated. Using five public datasets (1,471 stories, 101 annotators including critics, students, and lay readers), we (i) extract 17 reference-less textual features (e.g., coherence, emotional variance, average sentence length...); (ii) model individual reader preferences, deriving feature importance vectors that reflect their textual priorities; and (iii) analyze these vectors in a shared "preference space". Reader vectors cluster into two profiles: 'surface-focused readers' (mainly non-experts), who prioritize readability and textual richness; and 'holistic readers' (mainly experts), who value thematic development, rhetorical variety, and sentiment dynamics. Our results quantitatively explain how measurements of literary quality are a function of how text features align with each reader's preferences. These findings advocate for reader-sensitive evaluation frameworks in the field of creative text generation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:50:22 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Marco",
"Guillermo",
""
],
[
"Gonzalo",
"Julio",
""
],
[
"Fresno",
"Víctor",
""
]
] |
2506.03311 | Haim Avron | Haim Avron, Uria Mor | Demystifying Tubal Tensor Algebra | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Developed in a series of seminal papers in the early 2010s, the tubal tensor framework provides a clean and effective algebraic setting for tensor computations, supporting matrix-mimetic features such as a tensor Singular Value Decomposition and Eckart-Young-like optimality results. It has proven to be a powerful tool for analyzing inherently multilinear data arising in hyperspectral imaging, medical imaging, neural dynamics, scientific simulations, and more. At the heart of tubal tensor algebra lies a special tensor-tensor product: originally the t-product, later generalized into a full family of products via the $\star_M$-product. Though initially defined through the multiplication of a block-circulant unfolding of one tensor by a matricization of another, it was soon observed that the t-product can be interpreted as standard matrix multiplication where the scalars are tubes-i.e., real vectors twisted ``inward.'' Yet, a fundamental question remains: why is this the ``right'' way to define a tensor-tensor product in the tubal setting? In this paper, we show that the t-product and its $\star_M$ generalization arise naturally when viewing third-order tensors as matrices of tubes, together with a small set of desired algebraic properties. Furthermore, we prove that the $\star_M$-product is, in fact, the only way to define a tubal product satisfying these properties. Thus, while partly expository in nature - aimed at presenting the foundations of tubal tensor algebra in a cohesive and accessible way - this paper also addresses theoretical gaps in the tubal tensor framework, proves new results, and provides justification for the tubal tensor framework central constructions, thereby shedding new light on it. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:51:29 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Avron",
"Haim",
""
],
[
"Mor",
"Uria",
""
]
] |
2506.03312 | Celia Chen | Celia Chen, Scotty Beland, Ingo Burghardt, Jill Byczek, William J. Conway, Eric Cotugno, Sadaf Davre, Megan Fletcher, Rajesh Kumar Gnanasekaran, Kristin Hamilton, Marilyn Harbert, Jordan Heustis, Tanaya Jha, Emily Klein, Hayden Kramer, Alex Leitch, Jessica Perkins, Casi Sherman, Celia Sterrn, Logan Stevens, Rebecca Zarrella, Jennifer Golbeck | Cross-Platform Violence Detection on Social Media: A Dataset and Analysis | In Proceedings of the 17th ACM Web Science Conference (WebSci '25). 9 pages | In Proceedings of the 17th ACM Web Science Conference (WebSci '25). ACM, New Brunswick, NJ, USA, 9 pages (2025) | 10.1145/3717867.3717877 | null | cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Violent threats remain a significant problem across social media platforms. Useful, high-quality data facilitates research into the understanding and detection of malicious content, including violence. In this paper, we introduce a cross-platform dataset of 30,000 posts hand-coded for violent threats and sub-types of violence, including political and sexual violence. To evaluate the signal present in this dataset, we perform a machine learning analysis with an existing dataset of violent comments from YouTube. We find that, despite originating from different platforms and using different coding criteria, we achieve high classification accuracy both by training on one dataset and testing on the other, and in a merged dataset condition. These results have implications for content-classification strategies and for understanding violent content across social media. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:54:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Celia",
""
],
[
"Beland",
"Scotty",
""
],
[
"Burghardt",
"Ingo",
""
],
[
"Byczek",
"Jill",
""
],
[
"Conway",
"William J.",
""
],
[
"Cotugno",
"Eric",
""
],
[
"Davre",
"Sadaf",
""
],
[
"Fletcher",
"Megan",
""
],
[
"Gnanasekaran",
"Rajesh Kumar",
""
],
[
"Hamilton",
"Kristin",
""
],
[
"Harbert",
"Marilyn",
""
],
[
"Heustis",
"Jordan",
""
],
[
"Jha",
"Tanaya",
""
],
[
"Klein",
"Emily",
""
],
[
"Kramer",
"Hayden",
""
],
[
"Leitch",
"Alex",
""
],
[
"Perkins",
"Jessica",
""
],
[
"Sherman",
"Casi",
""
],
[
"Sterrn",
"Celia",
""
],
[
"Stevens",
"Logan",
""
],
[
"Zarrella",
"Rebecca",
""
],
[
"Golbeck",
"Jennifer",
""
]
] |
2506.03313 | Francisco de Melo Vir\'issimo | Francisco de Melo Vir\'issimo | Low-EFFourth: A computational framework for generating and studying multilevel model ensembles in low-dimensional systems | LEF4 and its documentation are freely available for download on Zenodo, see https://doi.org/10.5281/zenodo.15566109 | null | null | null | physics.geo-ph nlin.CD physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | This paper introduces Low-EFFourth (LEF4), a MATLAB-based computational framework designed for generating and studying multilevel model ensembles in continuous dynamical systems. Initially developed to address questions in climate modelling, LEF4 can also be used in other disciplines such as epidemiology, economics, and engineering, with minimal modifications to the code. The framework provides an efficient and flexible approach for investigating uncertainties arising from initial conditions, model parameters, numerical methods, and model formulation. This preprint serves as a formal reference for the LEF4 codebase and provides a concise technical and conceptual overview of its purpose, structure, applications and development pipeline. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:58:42 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Viríssimo",
"Francisco de Melo",
""
]
] |
2506.03314 | Fien Apers | Fien Apers, Miguel Montero, Irene Valenzuela | Backtracking AdS flux vacua | 24 pages + appendices + references, 2 figures | null | null | CERN-TH-2025-109, IFT-25-062 | hep-th | http://creativecommons.org/licenses/by/4.0/ | We introduce an algorithm (dubbed "flux backtracking") to reverse-engineer the brane picture from an AdS flux vacuum. Given an AdS flux vacuum as input, the algorithm outputs a singularity in 10 or 11 dimensions. This singularity has the property that when probed with the appropriate stack of branes (and after taking the near-horizon limit), one recovers the initial AdS vacuum. After testing the procedure on a number of known AdS/CFT pairs, we apply it to AdS flux vacua without known holographic dual, notably the scale-separated DGKT solution. In this case, flux backtracking produces a certain strongly coupled singularity in massive IIA; we conjecture that the worldvolume CFT of D4-branes probing this singularity should be the holographic CFT dual to DGKT (if it exists). Applying the procedure to the DGKT-related scale-separated AdS$_4$ solutions without Romans mass, we find instead a conical and weakly coupled singularity. We also comment on the results and limitations of applying the procedure to KKLT. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 18:59:18 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Apers",
"Fien",
""
],
[
"Montero",
"Miguel",
""
],
[
"Valenzuela",
"Irene",
""
]
] |
2506.03315 | Kai Sauerwald | Kai Sauerwald, Kenneth Skiba, Eduardo Ferm\'e and Thomas Meyer | Axiomatics of Restricted Choices by Linear Orders of Sets with Minimum as Fallback | null | null | null | null | cs.AI cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study how linear orders can be employed to realise choice functions for which the set of potential choices is restricted, i.e., the possible choice is not possible among the full powerset of all alternatives. In such restricted settings, constructing a choice function via a relation on the alternatives is not always possible. However, we show that one can always construct a choice function via a linear order on sets of alternatives, even when a fallback value is encoded as the minimal element in the linear order. The axiomatics of such choice functions are presented for the general case and the case of union-closed input restrictions. Restricted choice structures have applications in knowledge representation and reasoning, and here we discuss their applications for theory change and abstract argumentation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:03:12 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Sauerwald",
"Kai",
""
],
[
"Skiba",
"Kenneth",
""
],
[
"Fermé",
"Eduardo",
""
],
[
"Meyer",
"Thomas",
""
]
] |
2506.03316 | Alex Krasnok | Deepanshu Trivedi, Laraib Niaz, and Alex Krasnok | Selective Addressing of Coupled Qubits via Complex Frequency Zero Targeting | null | null | null | null | quant-ph physics.app-ph physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Achieving precise, individual control over qubits within scalable quantum processors is critically hampered by parasitic couplings and spectral crowding, leading to detrimental crosstalk. While optimal absorption strategies based on time-reversal symmetry have shown promise for single emitters, their applicability is limited in realistic multi-qubit systems where realistic losses break time-reversal symmetry. This work introduces a robust approach using complex frequency (CF) pulses specifically tailored to the complex reflection zeros of the complete, coupled, and explicitly lossy qubit-waveguide system. This method circumvents the limitations of idealized time-reversal arguments by directly engaging with the dissipative system's true response characteristics. We first develop a theoretical framework for a system of three coupled two-level emitters, employing Heisenberg equations to derive the system's response and design appropriate CF pulses that inherently account for the system's dissipative nature. The efficacy and practicality of this approach are then validated through comprehensive transient simulations for a realistic model of three Josephson junction-based transmon qubits, explicitly including intrinsic qubit losses. Our results demonstrate that CF pulses can selectively excite a target qubit with significantly suppressed crosstalk to neighboring qubits, markedly outperforming conventional Gaussian pulses of comparable energy. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:05:15 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Trivedi",
"Deepanshu",
""
],
[
"Niaz",
"Laraib",
""
],
[
"Krasnok",
"Alex",
""
]
] |
2506.03317 | Aydogan Ozcan | Yuntian Wang, Zafer Yilmaz, Yuhang Li, Edward Liu, Eric Ahlberg, Farid Ghahari, Ertugrul Taciroglu, Aydogan Ozcan | Structural Vibration Monitoring with Diffractive Optical Processors | 33 Pages, 8 Figures, 1 Table | null | null | null | physics.optics cs.CV cs.LG physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Structural Health Monitoring (SHM) is vital for maintaining the safety and longevity of civil infrastructure, yet current solutions remain constrained by cost, power consumption, scalability, and the complexity of data processing. Here, we present a diffractive vibration monitoring system, integrating a jointly optimized diffractive layer with a shallow neural network-based backend to remotely extract 3D structural vibration spectra, offering a low-power, cost-effective and scalable solution. This architecture eliminates the need for dense sensor arrays or extensive data acquisition; instead, it uses a spatially-optimized passive diffractive layer that encodes 3D structural displacements into modulated light, captured by a minimal number of detectors and decoded in real-time by shallow and low-power neural networks to reconstruct the 3D displacement spectra of structures. The diffractive system's efficacy was demonstrated both numerically and experimentally using millimeter-wave illumination on a laboratory-scale building model with a programmable shake table. Our system achieves more than an order-of-magnitude improvement in accuracy over conventional optics or separately trained modules, establishing a foundation for high-throughput 3D monitoring of structures. Beyond SHM, the 3D vibration monitoring capabilities of this cost-effective and data-efficient framework establish a new computational sensing modality with potential applications in disaster resilience, aerospace diagnostics, and autonomous navigation, where energy efficiency, low latency, and high-throughput are critical. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:06:04 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"Yuntian",
""
],
[
"Yilmaz",
"Zafer",
""
],
[
"Li",
"Yuhang",
""
],
[
"Liu",
"Edward",
""
],
[
"Ahlberg",
"Eric",
""
],
[
"Ghahari",
"Farid",
""
],
[
"Taciroglu",
"Ertugrul",
""
],
[
"Ozcan",
"Aydogan",
""
]
] |
2506.03318 | Francis Papa | Francis P. Papa | Validation of Quantum Elliptic Curve Point Addition Circuits | null | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Specific quantum algorithms exist to-in theory-break elliptic curve cryptographic protocols. Implementing these algorithms requires designing quantum circuits that perform elliptic curve arithmetic. To accurately judge a cryptographic protocol's resistance against future quantum computers, researchers figure out minimal resource-count circuits for performing these operations while still being correct. To assure the correctness of a circuit, it is integral to restore all ancilla qubits used to their original states. Failure to do so could result in decoherence of the computation's final result. Through rigorous classical simulation and unit testing, I surfaced four inconsistencies in the state-of-the-art quantum circuit for elliptic curve point addition where the circuit diagram states the qubits are returned in the original ($|0\rangle$) state, but the intermediate values are not uncomputed. I provide fixes to the circuit without increasing the leading-order gate cost. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:06:18 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Papa",
"Francis P.",
""
]
] |
2506.03319 | Daniel Cranston | Nicolas Bousquet and Daniel W. Cranston | A Linear Kernel for Independent Set Reconfiguration in Planar Graphs | 20 pages, 8 figures | null | null | null | math.CO cs.CC | http://creativecommons.org/licenses/by/4.0/ | Fix a positive integer $r$, and a graph $G$ that is $K_{3,r}$-minor-free. Let $I_s$ and $I_t$ be two independent sets in $G$, each of size $k$. We begin with a ``token'' on each vertex of $I_s$ and seek to move all tokens to $I_t$, by repeated ``token jumping'', removing a single token from one vertex and placing it on another vertex. We require that each intermediate arrangement of tokens again specifies an independent set of size $k$. Given $G$, $I_s$, and $I_t$, we ask whether there exists a sequence of token jumps that transforms $I_s$ into $I_t$. When $k$ is part of the input, this problem is known to be PSPACE-complete. However, it was shown by Ito, Kami\'nski, and Ono (2014) to be fixed-parameter tractable. That is, when $k$ is fixed, the problem can be solved in time polynomial in the order of $G$.
Here we strengthen the upper bound on the running time in terms of $k$ by showing that the problem has a kernel of size linear in $k$. More precisely, we transform an arbitrary input problem on a $K_{3,r}$-minor-free graph into an equivalent problem on a ($K_{3,r}$-minor-free) graph with order $O(k)$. This answers positively a question of Bousquet, Mouawad, Nishimura, and Siebertz (2024) and improves the recent quadratic kernel of Cranston, M\"{u}hlenthaler, and Peyrille (2024+). For planar graphs, we further strengthen this upper bound to get a kernel of size at most $42k$. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:06:30 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bousquet",
"Nicolas",
""
],
[
"Cranston",
"Daniel W.",
""
]
] |
2506.03320 | Jack Bell | Jack Bell, Luigi Quarantiello, Eric Nuertey Coleman, Lanpei Li, Malio Li, Mauro Madeddu, Elia Piccoli, Vincenzo Lomonaco | The Future of Continual Learning in the Era of Foundation Models: Three Key Directions | 16 pages, 1 figure, accepted at TCAI workshop 2025 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Continual learning--the ability to acquire, retain, and refine knowledge over time--has always been fundamental to intelligence, both human and artificial. Historically, different AI paradigms have acknowledged this need, albeit with varying priorities: early expert and production systems focused on incremental knowledge consolidation, while reinforcement learning emphasised dynamic adaptation. With the rise of deep learning, deep continual learning has primarily focused on learning robust and reusable representations over time to solve sequences of increasingly complex tasks. However, the emergence of Large Language Models (LLMs) and foundation models has raised the question: Do we still need continual learning when centralised, monolithic models can tackle diverse tasks with access to internet-scale knowledge? We argue that continual learning remains essential for three key reasons: (i) continual pre-training is still necessary to ensure foundation models remain up to date, mitigating knowledge staleness and distribution shifts while integrating new information; (ii) continual fine-tuning enables models to specialise and personalise, adapting to domain-specific tasks, user preferences, and real-world constraints without full retraining, avoiding the need for computationally expensive long context-windows; (iii) continual compositionality offers a scalable and modular approach to intelligence, enabling the orchestration of foundation models and agents to be dynamically composed, recombined, and adapted. While continual pre-training and fine-tuning are explored as niche research directions, we argue it is continual compositionality that will mark the rebirth of continual learning. The future of AI will not be defined by a single static model but by an ecosystem of continually evolving and interacting models, making continual learning more relevant than ever. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:06:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bell",
"Jack",
""
],
[
"Quarantiello",
"Luigi",
""
],
[
"Coleman",
"Eric Nuertey",
""
],
[
"Li",
"Lanpei",
""
],
[
"Li",
"Malio",
""
],
[
"Madeddu",
"Mauro",
""
],
[
"Piccoli",
"Elia",
""
],
[
"Lomonaco",
"Vincenzo",
""
]
] |
2506.03321 | Victor Cid | Victor H. Cid, James Mork | Enhancing Automatic PT Tagging for MEDLINE Citations Using Transformer-Based Models | 26 pages, 8 tables, 3 figures | null | null | null | cs.DL cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We investigated the feasibility of predicting Medical Subject Headings (MeSH) Publication Types (PTs) from MEDLINE citation metadata using pre-trained Transformer-based models BERT and DistilBERT. This study addresses limitations in the current automated indexing process, which relies on legacy NLP algorithms. We evaluated monolithic multi-label classifiers and binary classifier ensembles to enhance the retrieval of biomedical literature. Results demonstrate the potential of Transformer models to significantly improve PT tagging accuracy, paving the way for scalable, efficient biomedical indexing. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:06:51 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cid",
"Victor H.",
""
],
[
"Mork",
"James",
""
]
] |
2506.03322 | Jazm\'in Ord\'o\~nez-Toro | Jazm\'in Ord\'o\~nez-Toro (1), Sergio A. Dzib (2), Laurent Loinard (1 and 3 and 4), Gisela Ortiz-Le\'on (5), Marina A. Kounkel (6), Phillip A. B. Galli (7), Josep M. Masqu\'e (8), Luis H. Quiroga-Nu\~nez (9), Sundar Srinivasan (1), S.-N. X. Medina (2 and 10), Luis F. Rodr\'iguez (1) ((1) Instituto de Radioastronom\'ia y Astrof\'isica, Universidad Nacional Aut\'onoma de M\'exico, M\'exico, (2) Max Planck Institut f\"ur Radioastronomie, Germany, (3) Black Hole Initiative at Harvard University, USA, (4) David Rockefeller Center for Latin American Studies, Harvard University, USA, (5) Instituto Nacional de Astrof\'isica, \'Optica y Electr\'onica, M\'exico, (6) Department of Physics and Astronomy, University of North Florida, USA, (7) Instituto de Astronomia, Geof\'isica e Ci\^encias Atmosf\'ericas, Universidade de S\~ao Paulo, Brazil, (8) Departamento de Astronom\'ia, Universidad de Guanajuato, M\'exico, (9) Department of Aerospace, Physics and Space Sciences, Florida Institute of Technology, USA, (10) German Aerospace Center, Germany) | Dynamical mass of the Serpens intermediate-mass young stellar system EC\,95 with DYNAMO--VLBA | Accepted for publication in MNRAS. 13 pages, 11 figures | null | null | null | astro-ph.SR astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We present dynamical mass measurements, obtained from multi-epoch Very Long Baseline Interferometry (VLBI) observations, for the young multiple stellar system EC\,95, located in the core of the Serpens star-forming region. Our dataset includes both archival data and more recent observations obtained as part of the \textit{Dynamical Masses of Young Stellar Multiple Systems with the VLBA project} (DYNAMO--VLBA), totalling 32 epochs over 12 years of observation. The observations span more than half of the EC\,95AB orbit, which has an estimated period of $21.6\pm0.1$ years, and yield masses of $2.15\pm0.10$ M$_\odot$ for EC\,95A and $2.00\pm0.12$ M$_\odot$ for EC\,95B. Furthermore, for the first time, we have estimated the mass of the third component (EC\,95C) using four available radio detections as well as an infrared observation. We find it to be 0.26 $^{+0.53}_{-0.46}$ M$_\odot$, suggesting that E\,95C is a low-mass T Tauri star. We conclude that EC\,95 is a hierarchical triple system comprised of a tight central binary proto-Herbig\,AeBe system and a lower mass companion on a wider orbit. These results contribute to improve our understanding of the stellar dynamics in this system and provide valuable insights into its triple nature. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:10:48 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ordóñez-Toro",
"Jazmín",
"",
"Instituto de Radioastronomía y Astrofísica, Universidad Nacional Autónoma de México, México"
],
[
"Dzib",
"Sergio A.",
"",
"Max Planck Institut für Radioastronomie, Germany"
],
[
"Loinard",
"Laurent",
"",
"1 and 3 and 4"
],
[
"Ortiz-León",
"Gisela",
"",
"Instituto Nacional de Astrofísica, Óptica y Electrónica, México"
],
[
"Kounkel",
"Marina A.",
"",
"Department of Physics and Astronomy, University of North Florida, USA"
],
[
"Galli",
"Phillip A. B.",
"",
"Instituto de Astronomia, Geofísica e Ciências Atmosféricas, Universidade de São Paulo, Brazil"
],
[
"Masqué",
"Josep M.",
"",
"Departamento de Astronomía, Universidad de Guanajuato, México"
],
[
"Quiroga-Nuñez",
"Luis H.",
"",
"Department of Aerospace, Physics and Space Sciences, Florida Institute of Technology, USA"
],
[
"Srinivasan",
"Sundar",
"",
"Instituto de Radioastronomía y Astrofísica, Universidad Nacional Autónoma de México, México"
],
[
"Medina",
"S. -N. X.",
"",
"2 and 10"
],
[
"Rodríguez",
"Luis F.",
"",
"Instituto de Radioastronomía y Astrofísica, Universidad Nacional Autónoma de México, México"
]
] |
2506.03323 | Giuseppe Di Guglielmo | Flor Sanders and Gaurav Agarwal and Luca Carloni and Giuseppe Di Guglielmo and Andy C. Y. Li and Gabriel N. Perdue | Fast Machine Learning for Quantum Control of Microwave Qudits on Edge Hardware | null | null | null | FERMILAB-PUB-25-0356-ETD | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Quantum optimal control is a promising approach to improve the accuracy of quantum gates, but it relies on complex algorithms to determine the best control settings. CPU or GPU-based approaches often have delays that are too long to be applied in practice. It is paramount to have systems with extremely low delays to quickly and with high fidelity adjust quantum hardware settings, where fidelity is defined as overlap with a target quantum state. Here, we utilize machine learning (ML) models to determine control-pulse parameters for preparing Selective Number-dependent Arbitrary Phase (SNAP) gates in microwave cavity qudits, which are multi-level quantum systems that serve as elementary computation units for quantum computing. The methodology involves data generation using classical optimization techniques, ML model development, design space exploration, and quantization for hardware implementation. Our results demonstrate the efficacy of the proposed approach, with optimized models achieving low gate trace infidelity near $10^{-3}$ and efficient utilization of programmable logic resources. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:14:18 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Sanders",
"Flor",
""
],
[
"Agarwal",
"Gaurav",
""
],
[
"Carloni",
"Luca",
""
],
[
"Di Guglielmo",
"Giuseppe",
""
],
[
"Li",
"Andy C. Y.",
""
],
[
"Perdue",
"Gabriel N.",
""
]
] |
2506.03324 | Ethan Che | Ethan Che and Hakan Ceylan and James McInerney and Nathan Kallus | Optimization of Epsilon-Greedy Exploration | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern recommendation systems rely on exploration to learn user preferences for new items, typically implementing uniform exploration policies (e.g., epsilon-greedy) due to their simplicity and compatibility with machine learning (ML) personalization models. Within these systems, a crucial consideration is the rate of exploration - what fraction of user traffic should receive random item recommendations and how this should evolve over time. While various heuristics exist for navigating the resulting exploration-exploitation tradeoff, selecting optimal exploration rates is complicated by practical constraints including batched updates, time-varying user traffic, short time horizons, and minimum exploration requirements. In this work, we propose a principled framework for determining the exploration schedule based on directly minimizing Bayesian regret through stochastic gradient descent (SGD), allowing for dynamic exploration rate adjustment via Model-Predictive Control (MPC). Through extensive experiments with recommendation datasets, we demonstrate that variations in the batch size across periods significantly influence the optimal exploration strategy. Our optimization methods automatically calibrate exploration to the specific problem setting, consistently matching or outperforming the best heuristic for each setting. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:14:53 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Che",
"Ethan",
""
],
[
"Ceylan",
"Hakan",
""
],
[
"McInerney",
"James",
""
],
[
"Kallus",
"Nathan",
""
]
] |
2506.03325 | Mark Langhirt | Mark Langhirt, Charles Holland, Ying-Tsong Lin | A three-dimensional energy flux acoustic propagation model | 54 pages, 18 figures | null | null | null | physics.comp-ph physics.ao-ph | http://creativecommons.org/licenses/by/4.0/ | This paper extends energy flux methods to handle three-dimensional ocean acoustic environments, the implemented solution captures horizontally refracted incoherent acoustic intensity, and its required computational effort is predominantly independent of range and frequency. Energy flux models are principally derived as incoherent solutions for acoustic propagation in bounded waveguides. The angular distribution of incoherent acoustic intensity may be derived from Wentzel-Kramers-Brillouin modes transformed to the continuous angular domain via the ray-mode analogy. The adiabatic approximation maps angular distributions of acoustic intensity as waveguide properties vary along a range-dependent environment, and the final solution integrates a modal intensity kernel over propagation angles. Additional integration kernels can be derived that modulate the incoherent field by specific physical wave phenomena such as geometric spreading, refractive focusing, and boundary attenuation and interference. This three-dimensional energy flux model is derived from a double-mode-sum cross-product, is integrated over solid-angles, incorporates a bi-variate convergence factor, accounts for acoustic energy escaping the computational domain through transparent transverse boundaries, and accumulates bottom attenuation along transverse cycle trajectories. Transmission loss fields compare favorably with analytic, ray tracing, and parabolic equation solutions for the canonical ASA wedge problem, and three-dimensional adiabatic ray trajectories for the ideal wedge are demonstrated. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:15:16 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Langhirt",
"Mark",
""
],
[
"Holland",
"Charles",
""
],
[
"Lin",
"Ying-Tsong",
""
]
] |
2506.03326 | Roberto Montemanni | Roberto Montemanni, Derek H. Smith | On Solving the Shortest Paths with Exclusive-Disjunction Arc Pairs Conflicts | null | null | null | null | math.OC | http://creativecommons.org/licenses/by/4.0/ | A variant of the well-known Shortest Path Problem is studied in this paper, where pairs of conflicting arcs are provided, and for each conflicting pair a penalty is paid once neither or both of the arcs are selected. This configures a set of soft-constraints. The problem, which can be used to model real applications, looks for a path from a given origin to a given destination that minimizes the cost of the arcs traversed plus the penalties incurred. In this paper, we consider a compact mixed integer linear program representing the problem and we solve it with the open-source solver CP-SAT, part of the Google OR-Tools computational suite. An experimental campaign on the instances available from the literature indicates that the approach we propose achieves results comparable with those of state-of-the-art solvers, notwithstanding it is a compact model, while the other approaches require the generation of dynamic constraints in order for the models to be competitive. Some best-known results have been improved in this study, and some instances have been closed for the first time. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:16:32 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Montemanni",
"Roberto",
""
],
[
"Smith",
"Derek H.",
""
]
] |
2506.03327 | Paul A. Wiegert | Paul Wiegert, Denis Vida, David L. Clark, Auriane Egal, Richard Wainscoat, Robert Weryk | A limit on the mass of the Taurid Resonant Swarm at sub-100 meter sizes | 9 pages, accepted by the Planetary Science Journal | null | null | null | astro-ph.EP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We report on a pencil-beam survey of the Taurid Swarm, a possible concentration of bodies in the Taurid meteoroid stream associated with the 7:2 mean-motion resonance with Jupiter. Canada-France-Hawaii Telescope MegaCam observations reaching apparent magnitudes of 24.5 in the gri filter were taken over three nights. Rates of motion on the sky allowed for the quick elimination of main-belt objects from the over 1000 moving sources seen. Eight candidates with on-sky rates of motion consistent with Taurids were detected, but seven were subsequently shown to be non-Taurids (Hungarias, Mars-crossers, etc). One object might be a 60 m class Taurid but not enough data was collected and its orbit remains ambiguous. Our results are consistent with no Taurid Swarm members observed, and an upper limit of fewer than 3e3 - 3e4 objects down to H=25.6 +/- 0.3 (diameter of 34-76 m assuming a 2P/Encke-like albedo) at the 95% confidence level. While meteor observations have confirmed the Taurid Swarm's existence at meter and smaller sizes, our results indicate that the current mass budget of the swarm at 100 m sizes does not require an outsize parent to explain it. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:18:54 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wiegert",
"Paul",
""
],
[
"Vida",
"Denis",
""
],
[
"Clark",
"David L.",
""
],
[
"Egal",
"Auriane",
""
],
[
"Wainscoat",
"Richard",
""
],
[
"Weryk",
"Robert",
""
]
] |
2506.03328 | Tugba Erpek | Yalin E. Sagduyu, Tugba Erpek, Sastry Kompella, Kemal Davaslioglu | Relay Selection and User Equipment Admission in Resource-Efficient NextG Sidelink Communications | null | null | null | null | cs.NI cs.DC cs.IT eess.SP math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | 5G/6G sidelink communications addresses the challenge of connecting outer UEs, which are unable to directly access a base station (gNodeB), through inner UEs that act as relays to connect to the gNodeB. The key performance indicators include the achievable rates, the number of outer UEs that can connect to a gNodeB, and the latency experienced by outer UEs in establishing connections. We consider problem of determining the assignment of outer UEs to inner UEs based on the channel, interference, and traffic characteristics. We formulate an optimization problem to maximize a weighted sum rate of UEs, where weights can represent priority, waiting time, and queue length. This optimization accommodates constraints related to channel and interference characteristics that influence the rates at which links can successfully carry assigned traffic. While an exhaustive search can establish an upper bound on achievable rates by this non-convex optimization problem, it becomes impractical for larger number of outer UEs due to scalability issues related to high computational complexity. To address this, we present a greedy algorithm that incrementally selects links to maximize the sum rate, considering already activated links. This algorithm, although effective in achieving high sum rates, may inadvertently overlook some UEs, raising concerns about fairness. To mitigate this, we introduce a fairness-oriented algorithm that adjusts weights based on waiting time or queue length, ensuring that UEs with initially favorable conditions do not unduly disadvantage others over time. We show that this strategy not only improves the average admission ratio of UEs but also ensures a more equitable distribution of service among them, thereby providing a balanced and fair solution to sidelink communications. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:19:04 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Sagduyu",
"Yalin E.",
""
],
[
"Erpek",
"Tugba",
""
],
[
"Kompella",
"Sastry",
""
],
[
"Davaslioglu",
"Kemal",
""
]
] |
2506.03329 | Seongmin Kim | Seongmin Kim and In-Saeng Suh | Optimization of Functional Materials Design with Optimal Initial Data in Surrogate-Based Active Learning | null | null | null | null | cs.CE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The optimization of functional materials is important to enhance their properties, but their complex geometries pose great challenges to optimization. Data-driven algorithms efficiently navigate such complex design spaces by learning relationships between material structures and performance metrics to discover high-performance functional materials. Surrogate-based active learning, continually improving its surrogate model by iteratively including high-quality data points, has emerged as a cost-effective data-driven approach. Furthermore, it can be coupled with quantum computing to enhance optimization processes, especially when paired with a special form of surrogate model ($i.e.$, quadratic unconstrained binary optimization), formulated by factorization machine. However, current practices often overlook the variability in design space sizes when determining the initial data size for optimization. In this work, we investigate the optimal initial data sizes required for efficient convergence across various design space sizes. By employing averaged piecewise linear regression, we identify initiation points where convergence begins, highlighting the crucial role of employing adequate initial data in achieving efficient optimization. These results contribute to the efficient optimization of functional materials by ensuring faster convergence and reducing computational costs in surrogate-based active learning. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:21:08 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Kim",
"Seongmin",
""
],
[
"Suh",
"In-Saeng",
""
]
] |
2506.03330 | Roberto Montemanni | Roberto Montemanni, Derek H. Smith | On Solving the Knapsack Problem with Conflicts | null | null | null | null | math.OC | http://creativecommons.org/licenses/by/4.0/ | A variant of the well-known Knapsack Problem is studied in this paper, where pairs of items are conflicting, and cannot be selected at the same time. This configures a set of hard constraints. The problem, which can be used to model real applications, looks for a selection of items such that the total profit is maximized, the capacity of the container is respected, and no conflict is violated. In this paper, we consider a previously known mixed integer linear program representing the problem and we solve it with the open-source solver CP-SAT, part of the Google OR-Tools computational suite. An experimental campaign on the instances available from the literature and adopted in the last decade, indicate that the approach we propose achieves results comparable with, and often better than, those of state-of-the-art solvers, notwithstanding its intrinsic conceptual and implementation simplicity. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:22:04 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Montemanni",
"Roberto",
""
],
[
"Smith",
"Derek H.",
""
]
] |
2506.03331 | Masaya Kitajima | Masaya Kitajima | Generalized Hardy's identity for the astroid-type p-circle lattice point problem | 13 pages, 1 figure | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $r$ be a positive real number and $p$ satisfy $(2/p)\in\mathbb{N}$. Then, we consider the lattice point problem of the closed curves astroid-type $p$-circle $\{x\in\mathbb{R}^{2}|\ |x_{1}|^{p}+|x_{2}|^{p}=r^{p}\}$ which generalize the circle. In investigating the asymptotic behavior of the error term in the area approximation of the circle, G.H. Hardy conjectured an infimum for the evaluation in 1917. One of the grounds for this conjecture is the Hardy's identity, which is a series representation of the term, consisting of the Bessel function of order one and a certain number-theoretic function.
In order to investigate an infimum in the error evaluation of the astroid-type $p$-circle, which is unknown in previous studies, in this paper, we derive generalized Hardy's identity for the figures by using generalized Bessel functions. Furthermore, the differential formula for the functions, which is important for the proof of this identity, is closely related to the Erd\'{e}lyi-Kober operator, and this formula and operator are expected to be useful in our future research. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:25:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Kitajima",
"Masaya",
""
]
] |
2506.03332 | Yifei Ming | Yifei Ming, Zixuan Ke, Xuan-Phi Nguyen, Jiayu Wang, and Shafiq Joty | Helpful Agent Meets Deceptive Judge: Understanding Vulnerabilities in Agentic Workflows | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Agentic workflows -- where multiple large language model (LLM) instances interact to solve tasks -- are increasingly built on feedback mechanisms, where one model evaluates and critiques another. Despite the promise of feedback-driven improvement, the stability of agentic workflows rests on the reliability of the judge. However, judges may hallucinate information, exhibit bias, or act adversarially -- introducing critical vulnerabilities into the workflow. In this work, we present a systematic analysis of agentic workflows under deceptive or misleading feedback. We introduce a two-dimensional framework for analyzing judge behavior, along axes of intent (from constructive to malicious) and knowledge (from parametric-only to retrieval-augmented systems). Using this taxonomy, we construct a suite of judge behaviors and develop WAFER-QA, a new benchmark with critiques grounded in retrieved web evidence to evaluate robustness of agentic workflows against factually supported adversarial feedback. We reveal that even strongest agents are vulnerable to persuasive yet flawed critiques -- often switching correct answers after a single round of misleading feedback. Taking a step further, we study how model predictions evolve over multiple rounds of interaction, revealing distinct behavioral patterns between reasoning and non-reasoning models. Our findings highlight fundamental vulnerabilities in feedback-based workflows and offer guidance for building more robust agentic systems. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:26:23 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ming",
"Yifei",
""
],
[
"Ke",
"Zixuan",
""
],
[
"Nguyen",
"Xuan-Phi",
""
],
[
"Wang",
"Jiayu",
""
],
[
"Joty",
"Shafiq",
""
]
] |
2506.03333 | Juan Sebastian Rojas | Juan Sebastian Rojas, Chi-Guhn Lee | A Differential Perspective on Distributional Reinforcement Learning | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | To date, distributional reinforcement learning (distributional RL) methods have exclusively focused on the discounted setting, where an agent aims to optimize a potentially-discounted sum of rewards over time. In this work, we extend distributional RL to the average-reward setting, where an agent aims to optimize the reward received per time-step. In particular, we utilize a quantile-based approach to develop the first set of algorithms that can successfully learn and/or optimize the long-run per-step reward distribution, as well as the differential return distribution of an average-reward MDP. We derive proven-convergent tabular algorithms for both prediction and control, as well as a broader family of algorithms that have appealing scaling properties. Empirically, we find that these algorithms consistently yield competitive performance when compared to their non-distributional equivalents, while also capturing rich information about the long-run reward and return distributions. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:26:25 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Rojas",
"Juan Sebastian",
""
],
[
"Lee",
"Chi-Guhn",
""
]
] |
2506.03334 | Oem Trivedi | Oem Trivedi and Abraham Loeb | Could Planck Star Remnants be Dark Matter? | 12 pages with 2 figures, comments are very welcome! | null | null | null | gr-qc astro-ph.CO hep-th | http://creativecommons.org/licenses/by/4.0/ | We explore the end state of gravitational collapse under quantum gravity effects and propose that Planck Star Remnants (PSR), formed via nonsingular bounces, could serve as viable dark matter candidates. Within the framework of Loop Quantum Cosmology, we model the collapse of a homogeneous matter distribution and show that the classical singularity is replaced by a quantum bounce at the Planck density. By analytically matching the Friedmann Lemaitre Robertson Walker (FLRW) interior to an exterior Schwarzschild spacetime using the Israel junction conditions, we demonstrate that the bounce remains causally hidden from external observers, avoiding any observable re-expansion. This naturally leads to the formation of stable, non-radiating PSR, whose radius coincides with the Schwarzschild radius when the black hole mass approaches the Planck mass as a result of Hawking evaporation. We suggest that such remnants may originate from evaporating primordial black holes in the early universe, and estimate the relic abundance needed for PSR to account for the observed dark matter density. We also discuss some crucial differences between PSR and previous proposals of Planck mass relics. The scenario is shown to be consistent with existing astrophysical and cosmological constraints, offering a unified framework connecting quantum gravitational collapse, and the nature of dark matter. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:27:47 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Trivedi",
"Oem",
""
],
[
"Loeb",
"Abraham",
""
]
] |
2506.03335 | Dheeraj Khanna | Dheeraj Khanna, Jerrin Bright, Yuhao Chen and John S. Zelek | SportMamba: Adaptive Non-Linear Multi-Object Tracking with State Space Models for Team Sports | Paper accepted at CVSports IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW'25). The paper has 8 pages, including 6 Figures and 5 Tables | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Multi-object tracking (MOT) in team sports is particularly challenging due to the fast-paced motion and frequent occlusions resulting in motion blur and identity switches, respectively. Predicting player positions in such scenarios is particularly difficult due to the observed highly non-linear motion patterns. Current methods are heavily reliant on object detection and appearance-based tracking, which struggle to perform in complex team sports scenarios, where appearance cues are ambiguous and motion patterns do not necessarily follow a linear pattern. To address these challenges, we introduce SportMamba, an adaptive hybrid MOT technique specifically designed for tracking in dynamic team sports. The technical contribution of SportMamba is twofold. First, we introduce a mamba-attention mechanism that models non-linear motion by implicitly focusing on relevant embedding dependencies. Second, we propose a height-adaptive spatial association metric to reduce ID switches caused by partial occlusions by accounting for scale variations due to depth changes. Additionally, we extend the detection search space with adaptive buffers to improve associations in fast-motion scenarios. Our proposed technique, SportMamba, demonstrates state-of-the-art performance on various metrics in the SportsMOT dataset, which is characterized by complex motion and severe occlusion. Furthermore, we demonstrate its generalization capability through zero-shot transfer to VIP-HTD, an ice hockey dataset. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:28:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Khanna",
"Dheeraj",
""
],
[
"Bright",
"Jerrin",
""
],
[
"Chen",
"Yuhao",
""
],
[
"Zelek",
"John S.",
""
]
] |
2506.03336 | Laura Balzer PhD | Kirsten E. Landsiedel, Rachel Abbott, Atukunda Mucunguzi, Florence Mwangwa, Elijah Kakande, Edwin D. Charlebois, Carina Marquez, Moses R. Kamya, Laura B. Balzer | Causal Inference with Missing Exposures, Missing Outcomes, and Dependence | 15 pages (double spaced) of main text + 4 pages (double spaced) of supplementary material + 5 figures | null | null | null | stat.ME | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Missing data are ubiquitous in public health research. The missing-completely-at-random (MCAR) assumption is often unrealistic and can lead to meaningful bias when violated. The missing-at-random (MAR) assumption tends to be more reasonable, but guidance on conducting causal analyses under MAR is limited when there is missingness on multiple variables. We present a series of causal graphs and identification results to demonstrate the handling of missing exposures and outcomes in observational studies. For estimation and inference, we highlight the use of targeted minimum loss-based estimation (TMLE) with Super Learner to flexibly and robustly address confounding, missing data, and dependence. Our work is motivated by SEARCH-TB's investigation of the effect of alcohol consumption on the risk of incident tuberculosis (TB) infection in rural Uganda. This study posed notable challenges due to confounding, missingness on the exposure (alcohol use), missingness on the baseline outcome (defining who was at risk of TB), missingness on the outcome at follow-up (capturing who acquired TB), and clustering within households. Application to real data from SEARCH-TB highlighted the real-world consequences of the discussed methods. Estimates from TMLE suggested that alcohol use was associated with a 49% increase in the relative risk (RR) of incident TB infection (RR=1.49, 95%CI: 1.39-1.59). These estimates were notably larger and more precise than estimates from inverse probability weighting (RR=1.13, 95%CI: 1.00-1.27) and unadjusted, complete case analyses (RR=1.18, 95%CI: 0.89-1.57). Our work demonstrates the utility of causal models for describing the missing data mechanism and TMLE for flexible inference. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:28:57 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Landsiedel",
"Kirsten E.",
""
],
[
"Abbott",
"Rachel",
""
],
[
"Mucunguzi",
"Atukunda",
""
],
[
"Mwangwa",
"Florence",
""
],
[
"Kakande",
"Elijah",
""
],
[
"Charlebois",
"Edwin D.",
""
],
[
"Marquez",
"Carina",
""
],
[
"Kamya",
"Moses R.",
""
],
[
"Balzer",
"Laura B.",
""
]
] |
2506.03337 | Yide Ran | Yide Ran, Wentao Guo, Jingwei Sun, Yanzhou Pan, Xiaodong Yu, Hao Wang, Jianwen Xie, Yiran Chen, Denghui Zhang, Zhaozhuo Xu | Mitigating Non-IID Drift in Zeroth-Order Federated LLM Fine-Tuning with Transferable Sparsity | 56 pages, 11 figures | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Federated Learning enables collaborative fine-tuning of Large Language Models (LLMs) across decentralized Non-Independent and Identically Distributed (Non-IID) clients, but such models' massive parameter sizes lead to significant memory and communication challenges. This work introduces Meerkat, a sparse zeroth-order optimization (ZO) method designed for federated LLM fine-tuning. By limiting fine-tuning to a transferable, static, extremely sparse subset of parameters, Meerkat achieves remarkable communication efficiency, enabling cost-effective high-frequency synchronization. With theoretical analysis and experiments, we show that this high-frequency communication effectively mitigates Non-IID data challenges and leads to superior performance compared to full-parameter ZO. Furthermore, experiment results show that Meerkat outperforms existing sparsity baselines with better performance at the same communication frequency. To further handle Non-IID drift, Meerkat leverages traceable local updates and forms a virtual path for each client. This virtual path mechanism reveals the GradIP phenomenon: the inner products between LLM pre-training gradients maintained by server and client gradients estimated via ZO converges for extreme Non-IID clients but oscillates for IID ones. This distinct behavior provides a signal for identifying clients with extreme data heterogeneity. Using this signal, Meerkat-vp is proposed to analyze GradIP trajectories to identify extreme Non-IID clients and applies early stopping to enhance aggregated model quality. Experiments confirm that Meerkat and Meerkat-vp significantly improve the efficiency and effectiveness of ZO federated LLM fine-tuning. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:29:50 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ran",
"Yide",
""
],
[
"Guo",
"Wentao",
""
],
[
"Sun",
"Jingwei",
""
],
[
"Pan",
"Yanzhou",
""
],
[
"Yu",
"Xiaodong",
""
],
[
"Wang",
"Hao",
""
],
[
"Xie",
"Jianwen",
""
],
[
"Chen",
"Yiran",
""
],
[
"Zhang",
"Denghui",
""
],
[
"Xu",
"Zhaozhuo",
""
]
] |
2506.03338 | Alexander Ibrahim | Alexander Ibrahim, Pierre-Nicholas Roy | Three-body potential energy surface for parahydrogen | null | J. Chem. Phys. 156, 044301 (2022) | 10.1063/5.0076494 | null | physics.chem-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a 3D isotropic ab initio three-body (para-H$_2$)$_3$ interaction potential energy surface (PES). The electronic structure calculations are carried out at the correlated coupled-cluster theory level, with single, double, and perturbative triple excitations. The calculations use an augmented correlation-consistent triple zeta basis set and a supplementary midbond function. We construct the PES using the Reproducing-Kernel Hilbert Space toolkit [J. Chem. Inf. Model. 57, 1923 (2017)] with phenomenological and empirical adjustments to account for short-range and long-range behaviour. The (para-H$_2$)$_3$ interaction energies deviate drastically from the Axilrod-Teller-Muto (ATM) potential at short intermolecular separations. We find that the configuration of three para-H$_2$ molecules at the corners of an equilateral triangle is responsible for the majority of the (para-H$_2$)$_3$ interaction energy contribution in a hexagonal-close-packed lattice. In cases where two para-H$_2$ molecules are close to one another while the third is far away, the (para-H$_2$)$_3$ interaction PES takes the form of a modified version of the ATM potential. We expect the combination of this PES together with a first principles para-H$_2$--para-H$_2$ Adiabatic Hindered Rotor potential to outperform a widely-used effective pair potential for condensed many-body systems of para-H$_2$. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:31:04 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ibrahim",
"Alexander",
""
],
[
"Roy",
"Pierre-Nicholas",
""
]
] |
2506.03339 | Maximilian Balthasar Mansky | Maximilian Balthasar Mansky, Tobias Rohe, Dmytro Bondarenko, Linus Menzel, Claudia Linnhoff-Popien | Clique detection using symmetry-restricted quantum circuits | 4 pages, 4 figures, submitted to QML@QCE 25. arXiv admin note: substantial text overlap with arXiv:2505.12764 | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show the application of permutation-invariant quantum circuits to the clique problem. The experiment asks to label a clique through identification of the nodes in a larger subgraph. The permutation-invariant quantum circuit outperforms a cyclic-invariant alternative as well as a standard quantum machine learning ansatz. We explain the behavior through the intrinsic symmetry of the problem, in the sense that the problem is symmetric under permutation of both the feature and the label. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:31:26 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mansky",
"Maximilian Balthasar",
""
],
[
"Rohe",
"Tobias",
""
],
[
"Bondarenko",
"Dmytro",
""
],
[
"Menzel",
"Linus",
""
],
[
"Linnhoff-Popien",
"Claudia",
""
]
] |
2506.03340 | Zihui Xue | Zihui Xue, Mi Luo, Kristen Grauman | Seeing the Arrow of Time in Large Multimodal Models | Project website: https://vision.cs.utexas.edu/projects/SeeAoT | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Arrow of Time (AoT)-time's irreversible flow shaping physical events-is fundamental to video comprehension, yet remains a significant challenge for modern large multimodal models (LMMs). Current LMMs struggle to perceive and utilize temporal directionality in video when responding to language queries, obstructing deeper temporal understanding. We tackle this deficiency by first providing a critical analysis of existing benchmarks and models. We then introduce ArrowRL, a reinforcement learning (RL)-based training strategy with an innovative reverse reward that instills AoT awareness by encouraging divergent video interpretations between forward and reversed visual frames. For rigorous evaluation, we additionally develop AoTBench, a new multi-faceted benchmark probing temporally challenging questions. Experiments show ArrowRL greatly advances temporal perception: it not only achieves substantial improvements on our challenging AoTBench but also demonstrably boosts performance on standard video question answering (VQA) benchmarks (with peak accuracy gains reaching over 20% and 10% respectively). This validates ArrowRL's effectiveness and highlights the critical need for dedicated AoT understanding in LMMs. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:32:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xue",
"Zihui",
""
],
[
"Luo",
"Mi",
""
],
[
"Grauman",
"Kristen",
""
]
] |
2506.03341 | Juliana Boasso | Hugo Aimar and Juliana Boasso | A measure theoretic approach to Lipschitz regularity and its Haar type wavelet analysis | 23 pages, 2 figures, 2 tables | null | null | null | math.CA math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this note, we extend the characterization of dyadic Lipschitz regularity of functions to non-atomic probability spaces, using generalized Haar systems. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:32:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Aimar",
"Hugo",
""
],
[
"Boasso",
"Juliana",
""
]
] |
2506.03342 | Andreas Celary | Andreas Celary, Paul Kr\"uhner, Zehra Eksi | Reproducing kernel Hilbert space methods for modelling the discount curve | null | null | null | null | q-fin.MF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the theory of bond discounts, defined as the difference between the terminal payoff of the contract and its current price. Working in the setting of finite-dimensional realizations in the HJM framework, under suitable notions of no-arbitrage, the admissible discount curves take the form of polynomial, exponential functions. We introduce reproducing kernels that are admissible under no-arbitrage as a tractable regression basis for the estimation problem in calibrating the model to market data. We provide a thorough numerical analysis using real-world treasury data. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:32:26 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Celary",
"Andreas",
""
],
[
"Krühner",
"Paul",
""
],
[
"Eksi",
"Zehra",
""
]
] |
2506.03343 | Samuel Hopkins | Sam Hopkins and Joel B. Lewis | Upho lattices II: ways of realizing a core | 20 pages; 7 figures | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | A poset is called upper homogeneous, or "upho," if all of its principal order filters are isomorphic to the whole poset. In previous work of the first author, it was shown that each (finite-type N-graded) upho lattice has associated to it a finite graded lattice, called its core, which determines the rank generating function of the upho lattice. In that prior work the question of which finite graded lattices arise as cores was explored. Here, we study the question of in how many different ways a given finite graded lattice can be realized as the core of an upho lattice. We show that if the finite lattice has no nontrivial automorphisms, then it is the core of finitely many upho lattices. We also show that the number of ways a finite lattice can be realized as a core is unbounded, even when restricting to rank-two lattices. We end with a discussion of a potential algorithm for listing all the ways to realize a given finite lattice as a core. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:33:45 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Hopkins",
"Sam",
""
],
[
"Lewis",
"Joel B.",
""
]
] |
2506.03344 | Matt Brigida | Matt Brigida and Kathleen Maceyka | Hedging Deposit Run Risk Prior to the 2023 Regional Banking Crisis | null | null | null | null | q-fin.GN | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this analysis we determine factors driving the cross-sectional variation in uninsured deposits during the interest rate raising cycle of 2022 to 2023. The goal of our analysis is to determine whether banks proactively managed deposit run risk prior to the hiking cycle which produced the 2023 Regional Banking Crisis. We find evidence that interest rate forward, futures, and swap use affected the change in a bank uninsured deposits over the period. Interest rate option use, however, has no effect on the change in uninsured deposits. Similarly, bank equity levels were uncorrelated with uninsured deposit changes. We conclude we find no evidence of banks managing run risk via their balance sheet prior to the 2023 Regional Banking Crisis. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:34:18 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Brigida",
"Matt",
""
],
[
"Maceyka",
"Kathleen",
""
]
] |
2506.03345 | Xin Zhang | Chien-Fu (Frank) Huang, Katherine Sieg, Leonid Karlinksy, Nash Flores, Rebekah Sheraw, Xin Zhang | Semiconductor SEM Image Defect Classification Using Supervised and Semi-Supervised Learning with Vision Transformers | Published at 36th Annual SEMI Advanced Semiconductor Manufacturing Conference (ASMC) 2025 | null | 10.1109/ASMC64512.2025.11010396 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Controlling defects in semiconductor processes is important for maintaining yield, improving production cost, and preventing time-dependent critical component failures. Electron beam-based imaging has been used as a tool to survey wafers in the line and inspect for defects. However, manual classification of images for these nano-scale defects is limited by time, labor constraints, and human biases. In recent years, deep learning computer vision algorithms have shown to be effective solutions for image-based inspection applications in industry. This work proposes application of vision transformer (ViT) neural networks for automatic defect classification (ADC) of scanning electron microscope (SEM) images of wafer defects. We evaluated our proposed methods on 300mm wafer semiconductor defect data from our fab in IBM Albany. We studied 11 defect types from over 7400 total images and investigated the potential of transfer learning of DinoV2 and semi-supervised learning for improved classification accuracy and efficient computation. We were able to achieve classification accuracies of over 90% with less than 15 images per defect class. Our work demonstrates the potential to apply the proposed framework for a platform agnostic in-house classification tool with faster turnaround time and flexibility. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:34:54 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chien-Fu",
"",
"",
"Frank"
],
[
"Huang",
"",
""
],
[
"Sieg",
"Katherine",
""
],
[
"Karlinksy",
"Leonid",
""
],
[
"Flores",
"Nash",
""
],
[
"Sheraw",
"Rebekah",
""
],
[
"Zhang",
"Xin",
""
]
] |
2506.03346 | Evan Johnson | Evan C. Johnson and Sean R. Connolly | Negligible effects of environmental fluctuations on the maintenance of coral biodiversity: A test of five storage effects | null | null | null | null | q-bio.PE | http://creativecommons.org/licenses/by/4.0/ | The storage effect is a general explanation for ecological coexistence, wherein different species specialize on different states of a fluctuating environment, e.g., hot vs. cold years. Despite the storage effect's prominence in theoretical ecology, we lack evidence on whether it maintains biodiversity in nature. Here, we examine five storage effects in a community of 11 coral species from the Great Barrier Reef, using detailed size-structured demographic data collected over five years. We parameterize integral projection models, simulate coral communities, and quantify coexistence mechanisms through Modern Coexistence Theory. Results show that storage effects promote coexistence but are weak compared to fitness differences. Despite coral communities exhibiting theoretical prerequisites for strong temporal niche partitioning, the storage effect plays only a minor role in maintaining coral biodiversity. This aligns with growing evidence that storage effects are weak across ecosystems. Coral coexistence likely depends more on spatial processes, including microhabitat partitioning and asymmetric dispersal. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:36:54 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Johnson",
"Evan C.",
""
],
[
"Connolly",
"Sean R.",
""
]
] |
2506.03347 | Paul Zivich | Paul N Zivich, Haidong Lu | Constructing g-computation estimators: two case studies in selection bias | null | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | G-computation is a useful estimation method that can be adapted to address various biases in epidemiology. However, these adaptations may not be obvious for some complex causal structures. This challenge is an example of the much wider issue of translating a causal diagram into a novel estimation strategy. To highlight these challenges, we consider two recent cases from the selection bias literature: treatment-induced selection and co-occurrence of biases that lack a joint adjustment set. For each case study, we show how g-computation can be adapted, described how to implement that adaptation, show some general statistical properties, and illustrate the estimator using simulation. To simplify both the theoretical study and practical application of our estimators, we express the proposed g-computation estimators as stacked estimating equations. These examples illustrate how epidemiologists can translate identification results into an estimation strategy and study the theoretical and finite-sample properties of a novel estimator. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:38:43 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Zivich",
"Paul N",
""
],
[
"Lu",
"Haidong",
""
]
] |
2506.03348 | Davide Racco | Peter W. Graham, Davide Racco | Revisiting Isocurvature Bounds on the Minimal QCD Axion | 16 + 7 pages, 7 + 6 figures | null | null | ZU-TH 38/25 | hep-ph astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | The QCD axion has important connections to early universe cosmology. For example, it is often said that isocurvature limits rule out a combination of high axion decay constant, $f_a$, and high inflationary Hubble scale, $H_I$. High scales are theoretically motivated, so it is important to ask how robust this constraint is. We demonstrate that this constraint is naturally evaded when the quartic coupling of the complex $U(1)_\mathrm{PQ}$-breaking field is small. In this case, $f_a$ changes from a larger value during inflation to a smaller value in the later universe, suppressing isocurvature perturbations. Importantly, we show that in large parts of parameter space this solution is not jeopardised by overproduction of the axion through parametric resonance. The isocurvature bounds are thus dependent on UV physics. We have found that, even for the minimal QCD axion, large parts of UV parameter space at both high $f_a$ and high $H_I$ are in fact allowed, not ruled out by isocurvature constraints. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:40:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Graham",
"Peter W.",
""
],
[
"Racco",
"Davide",
""
]
] |
2506.03349 | Lupin Chun-Che Lin Lupin | Lupin Chun-Che Lin, Chin-Ping Hu, Chien-Chang Yen, Kuo-Chuan Pan, C. Y. Hui, Kwan-Lok Li, Yu-Chiung Lin, Yi-Sheng Huang, and Albert K. H. Kong | Effectiveness of Stacks in the Stacked Hilbert-Huang Transform | 17 pages, 12 figures (2 in appendix) and 1 table; accepted by Physical Review D | null | 10.1103/jtc1-3cp2 | null | astro-ph.IM | http://creativecommons.org/licenses/by-sa/4.0/ | The Hilbert-Huang transform (HHT) consists of empirical mode decomposition (EMD), which is a template-free method that represents the combination of different intrinsic modes on a time-frequency map (i.e., the Hilbert spectrum). The application of HHT involves introducing trials by imposing white noise on the signal and then calculating the ensemble mean process of the corresponding EMD to demonstrate its significance on the Hilbert spectrum. In this study, we develop a stacked Hilbert-Huang Transform (sHHT) method that generates the Hilbert spectrum for each trial and compiles all results to enhance the strength of the real instantaneous frequency of the main signal on the time-frequency map. This new approach is more sensitive to detecting/tracing the nonlinear and transient features of a signal embedded in astronomical databases than the conventional HHT, particularly when the signal experiences dramatic frequency changes in a short time. We analytically investigate the consistency of HHT and sHHT and perform numerical simulations to examine the dispersion of the instantaneous frequency obtained through sHHT and compare its advantages and effectiveness with those of conventional HHT. To confirm the feasibility of the sHHT, we demonstrate its application in verifying the signal of superorbital modulation in X-ray and binary black hole mergers in gravitational waves. | [
{
"version": "v1",
"created": "Mon, 02 Jun 2025 05:32:59 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Lin",
"Lupin Chun-Che",
""
],
[
"Hu",
"Chin-Ping",
""
],
[
"Yen",
"Chien-Chang",
""
],
[
"Pan",
"Kuo-Chuan",
""
],
[
"Hui",
"C. Y.",
""
],
[
"Li",
"Kwan-Lok",
""
],
[
"Lin",
"Yu-Chiung",
""
],
[
"Huang",
"Yi-Sheng",
""
],
[
"Kong",
"Albert K. H.",
""
]
] |
2506.03350 | Eliot Jones | Eliot Krzysztof Jones, Alexander Robey, Andy Zou, Zachary Ravichandran, George J. Pappas, Hamed Hassani, Matt Fredrikson, and J. Zico Kolter | Adversarial Attacks on Robotic Vision Language Action Models | null | null | null | null | cs.RO cs.AI | http://creativecommons.org/licenses/by/4.0/ | The emergence of vision-language-action models (VLAs) for end-to-end control is reshaping the field of robotics by enabling the fusion of multimodal sensory inputs at the billion-parameter scale. The capabilities of VLAs stem primarily from their architectures, which are often based on frontier large language models (LLMs). However, LLMs are known to be susceptible to adversarial misuse, and given the significant physical risks inherent to robotics, questions remain regarding the extent to which VLAs inherit these vulnerabilities. Motivated by these concerns, in this work we initiate the study of adversarial attacks on VLA-controlled robots. Our main algorithmic contribution is the adaptation and application of LLM jailbreaking attacks to obtain complete control authority over VLAs. We find that textual attacks, which are applied once at the beginning of a rollout, facilitate full reachability of the action space of commonly used VLAs and often persist over longer horizons. This differs significantly from LLM jailbreaking literature, as attacks in the real world do not have to be semantically linked to notions of harm. We make all code available at https://github.com/eliotjones1/robogcg . | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:43:58 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Jones",
"Eliot Krzysztof",
""
],
[
"Robey",
"Alexander",
""
],
[
"Zou",
"Andy",
""
],
[
"Ravichandran",
"Zachary",
""
],
[
"Pappas",
"George J.",
""
],
[
"Hassani",
"Hamed",
""
],
[
"Fredrikson",
"Matt",
""
],
[
"Kolter",
"J. Zico",
""
]
] |
2506.03351 | Gissell Estrada-Rodriguez | Gissell Estrada-Rodriguez, Heiko Gimperlein | Macroscopic boundary conditions for a fractional diffusion equation in chemotaxis | null | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we examine boundary effects in a fractional chemotactic equation derived from a kinetic transport model describing cell movement in response to chemical gradients (chemotaxis). Specifically, we analyze reflecting boundary conditions within a nonlocal fractional framework. Using boundary layer methods and perturbation theory, we derive first-order approximations for interior and boundary layer solutions under symmetric reflection conditions. This work provides fundamental insights into the complex interplay between fractional dynamics, chemotactic transport phenomena, and boundary interactions, opening future research in biological and physical applications involving nonlocal processes. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:45:27 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Estrada-Rodriguez",
"Gissell",
""
],
[
"Gimperlein",
"Heiko",
""
]
] |
2506.03352 | Alexander Ibrahim | Alexander Ibrahim, Pierre-Nicholas Roy | Equation of state of solid parahydrogen using ab initio two-body and three-body interaction potentials | null | J. Chem. Phys. 157, 174503 (2022) | 10.1063/5.0120169 | null | physics.chem-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present the equation of state (EOS) of solid parahydrogen between $ 0.024 \, {\r{A}}^{-3} $ and $ 0.1 \, {\r{A}}^{-3} $ at $ T = 4.2 $ K, calculated using path-integral Monte Carlo simulations, with ab initio two-body and three-body interaction potentials. We correct for finite size simulation errors using potential tail corrections. Trotter factorization errors are accounted for, either via extrapolation, or by using a suitably small imaginary time step. We incorporate the three-body interaction using two methods; the full inclusion method, where pair and three-body interactions are used in both Monte Carlo sampling and in the energy estimators, and the perturbative method, where three-body interactions are omitted from sampling but are still present in energy estimations. Both treatments of the three-body interaction return very similar total energies and pressures. The presence of three-body interactions has only minor effects on the structural properties of the solid. Whereas the pair interaction, on its own, significantly overestimates the pressure of solid parahydrogen, the additional presence of the three-body interaction causes a severe underestimation of the pressure. Our findings suggest that accurate simulations of solid parahydrogen require four-body and possibly higher-order many-body interactions. It may also be the case that static interaction potentials are entirely unsuitable for simulations of solid parahydrogen at high densities. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:47:04 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ibrahim",
"Alexander",
""
],
[
"Roy",
"Pierre-Nicholas",
""
]
] |
2506.03353 | Eric Falcon | Marlone Vernet and Eric Falcon | Thermodynamics and Statistical Equilibrium of Large-Scale Hydroelastic Wave Turbulence | in press in Physical Review Letters | null | null | null | physics.flu-dyn cond-mat.stat-mech nlin.CD physics.ao-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Understanding how statistical equilibrium can occur in out-of-equilibrium systems is of paramount interest, as it would enable the use of statistical mechanics tools to these systems. Here, we report the first experimental evidence of statistical equilibrium of the large scales of hydroelastic turbulent waves driven by small-scale random forcing. The wave field statistics at scales larger than the forcing scale, resolved in space and time, align well with the predictions of Rayleigh-Jeans equilibrium spectra over more than a decade. We measure zero net energy flux in this regime, as expected. We also determine the effective temperature, entropy, and heat capacity of this nonequilibrium system, demonstrating that classical thermodynamic concepts apply to describe large scales in statistical equilibrium of turbulent systems. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:47:44 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Vernet",
"Marlone",
""
],
[
"Falcon",
"Eric",
""
]
] |
2506.03354 | Mat\'ias Bravo | Mat\'ias Bravo, Claudia del P. Lagos, Katy L. Proctor, \'Angel Chandro-G\'omez, Chris Power | The galaxy-AGN scaling relations over 13 billion years in SHARK v2.0 (I): SMBH masses | Submitted to ApJ. 24 pages, 14 figures. Comments are welcome! | null | null | null | astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The presence of strong correlations between super-massive black hole (SMBH) masses and galaxy properties like stellar mass have been well-established in the local Universe, but how these scaling relations evolve with cosmic time is yet to be settled in both observations and theoretical models. Recent works have also highlighted the role of galaxy morphology on the scatter of the SMBH-galaxy mass scaling relations, while the impact of other galaxy properties remains poorly studied, like the role of galaxy environment. We use the state-of-the-art SHARK v2.0 semi-analytic model to explore the evolution of these galaxy-SMBH scaling relations to expand the available predictions from theoretical models to contrast with existing and upcoming observations. We find the relations between SMBH masses and both total and bulge stellar mass predicted by SHARK v2.0 to be in good overall agreement with observational measurements across a wide range of redshift and stellar masses. These scaling relations show a significant evolution as a function of cosmic time in SHARK v2.0, with SMBH masses $\sim1$ dex lower at $z=0$ compared to $z=9$ at fixed stellar mass and the scatter increasing by a factor of $\sim2-5$ towards low redshift. Both relations show a strong dependence with galaxy morphology and the main source for SMBH growth (gas accretion or mergers), with smaller trends with star formation rate, galaxy sizes, and environment. We find that galaxy morphology alone explains most of the scatter around both scaling relations, with other galaxy properties tying to the SMBH scaling relations through their correlations with morphology. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:51:26 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bravo",
"Matías",
""
],
[
"Lagos",
"Claudia del P.",
""
],
[
"Proctor",
"Katy L.",
""
],
[
"Chandro-Gómez",
"Ángel",
""
],
[
"Power",
"Chris",
""
]
] |
2506.03355 | Elias Abad Rocamora | Elias Abad Rocamora, Christian Schlarmann, Naman Deep Singh, Yongtao Wu, Matthias Hein, Volkan Cevher | Robustness in Both Domains: CLIP Needs a Robust Text Encoder | null | null | null | null | cs.LG cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | Adversarial input attacks can cause a significant shift of CLIP embeddings. This can affect the downstream robustness of models incorporating CLIP in the pipeline, such as text-to-image generative models or large vision language models. While some efforts have been done towards making the CLIP image encoders robust, the robustness of text encoders remains unexplored. In this work, we cover this gap in the literature. We propose LEAF: an efficient adversarial finetuning method for the text domain, with the ability to scale to large CLIP models. Our models significantly improve the zero-shot adversarial accuracy in the text domain, while maintaining the vision performance provided by robust image encoders. When combined with text-to-image diffusion models, we can improve the generation quality under adversarial noise. When employing our robust CLIP encoders in multimodal retrieval tasks, we improve the recall under adversarial noise over standard CLIP models. Finally, we show that robust text encoders facilitate better reconstruction of input text from its embedding via direct optimization. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:57:09 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Rocamora",
"Elias Abad",
""
],
[
"Schlarmann",
"Christian",
""
],
[
"Singh",
"Naman Deep",
""
],
[
"Wu",
"Yongtao",
""
],
[
"Hein",
"Matthias",
""
],
[
"Cevher",
"Volkan",
""
]
] |
2506.03356 | Artur Grigorev | Artur Grigorev and David Lillo-Trynes and Adriana-Simona Mihaita | Spatial Association Between Near-Misses and Accident Blackspots in Sydney, Australia: A Getis-Ord $G_i^*$ Analysis | null | null | null | null | eess.SY cs.CY cs.SY stat.AP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Road safety management teams utilize on historical accident logs to identify blackspots, which are inherently rare and sparse in space and time. Near-miss events captured through vehicle telematics and transmitted in real-time by connected vehicles reveal a unique potential of prevention due to their high frequency nature and driving engagement on the road. There is currently a lack of understanding of the high potential of near-miss data in real-time to proactively detect potential risky driving areas, in advance of a fatal collision. This paper aims to spatially identify clusters of reported accidents (A) versus high-severity near-misses (High-G) within an urban environment (Sydney, Australia) and showcase how the presence of near-misses can significantly lead to future crashes in identified risky hotspots. First, by utilizing a 400m grid framework, we identify significant crash hotspots using the Getis-Ord $G_i^*$ statistical approach. Second, we employ a Bivariate Local Moran's I (LISA) approach to assess and map the spatial concordance and discordance between official crash counts (A) and High-G counts from nearmiss data (High-G). Third, we classify areas based on their joint spatial patterns into: a) High-High (HH) as the most riskiest areas in both historical logs and nearmiss events, High-Low (HL) for high crash logs but low nearmiss records, c) Low-High (LH) for low past crash records but high nearmiss events, and d) Low-Low (LL) for safe areas. Finally, we run a feature importance ranking on all area patterns by using a contextual Point of Interest (POI) count features and we showcase which factors are the most critical to the occurrence of crash blackspots. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 19:58:56 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Grigorev",
"Artur",
""
],
[
"Lillo-Trynes",
"David",
""
],
[
"Mihaita",
"Adriana-Simona",
""
]
] |
2506.03357 | Aldan Creo | Aldan Creo, H\'ector Cerezo-Costas, Pedro Alonso-Doval, Maximiliano Hormaz\'abal-Lagos | Ask a Local: Detecting Hallucinations With Specialized Model Divergence | Supplementary materials: https://github.com/ACMCMC/ask-a-local | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hallucinations in large language models (LLMs) - instances where models generate plausible but factually incorrect information - present a significant challenge for AI.
We introduce "Ask a Local", a novel hallucination detection method exploiting the intuition that specialized models exhibit greater surprise when encountering domain-specific inaccuracies. Our approach computes divergence between perplexity distributions of language-specialized models to identify potentially hallucinated spans. Our method is particularly well-suited for a multilingual context, as it naturally scales to multiple languages without the need for adaptation, relying on external data sources, or performing training. Moreover, we select computationally efficient models, providing a scalable solution that can be applied to a wide range of languages and domains.
Our results on a human-annotated question-answer dataset spanning 14 languages demonstrate consistent performance across languages, with Intersection-over-Union (IoU) scores around 0.3 and comparable Spearman correlation values. Our model shows particularly strong performance on Italian and Catalan, with IoU scores of 0.42 and 0.38, respectively, while maintaining cross-lingual effectiveness without language-specific adaptations. We release our code and architecture to facilitate further research in multilingual hallucination detection. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:00:49 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Creo",
"Aldan",
""
],
[
"Cerezo-Costas",
"Héctor",
""
],
[
"Alonso-Doval",
"Pedro",
""
],
[
"Hormazábal-Lagos",
"Maximiliano",
""
]
] |
2506.03358 | Cl\'ement W. Royer | Albert S. Berahas and Michael J. O'Neill and Cl\'ement W. Royer | A line search framework with restarting for noisy optimization problems | null | null | null | null | math.OC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Nonlinear optimization methods are typically iterative and make use of gradient information to determine a direction of improvement and function information to effectively check for progress. When this information is corrupted by noise, designing a convergent and practical algorithmic process becomes challenging, as care must be taken to avoid taking bad steps due to erroneous information. For this reason, simple gradient-based schemes have been quite popular, despite being outperformed by more advanced techniques in the noiseless setting. In this paper, we propose a general algorithmic framework based on line search that is endowed with iteration and evaluation complexity guarantees even in a noisy setting. These guarantees are obtained as a result of a restarting condition, that monitors desirable properties for the steps taken at each iteration and can be checked even in the presence of noise. Experiments using a nonlinear conjugate gradient variant and a quasi-Newton variant illustrate that restarting can be performed without compromising practical efficiency and robustness. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:01:10 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Berahas",
"Albert S.",
""
],
[
"O'Neill",
"Michael J.",
""
],
[
"Royer",
"Clément W.",
""
]
] |
2506.03359 | Shahab Dabironezare | Shahab O. Dabironezare, Giulia Conenna, Daan Roos, Dimitry Lamers, Daniela Perez Capelo, Hendrik M. Veen, David J. Thoen, Vishal Anvekar, Stephen J. C. Yates, Willem Jellema, Robert Huiting, Lorenza Ferrari, Carole Tucker, Sven L. van Berkel, Peter K. Day, Henry G. Leduc, Charles M. Bradford, Nuria Llombart, Jochem J. A. Baselmans | Lens based Kinetic Inductance Detectors with Distributed Dual Polarised Absorbers for Far Infra-red Space-based Astronomy | This work has been submitted to the IEEE for possible publication | null | null | null | astro-ph.IM physics.ins-det | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Future space-based far infra-red astronomical observations require background limited detector sensitivities and scalable focal plane array solutions to realise their vast potential in observation speed. In this work, a focal plane array of lens absorber coupled Kinetic Inductance Detectors (KIDs) is proposed to fill this role. The figures of merit and design guidelines for the proposed detector concept are derived by employing a previously developed electromagnetic spectral modelling technique. Two designs operating at central frequencies of $6.98$ and $12$ THz are studied. A prototype array of the former is fabricated, and its performance is experimentally determined and validated. Specifically, the optical coupling of the detectors to incoherent distributed sources (i.e. normalised throughput) is quantified experimentally with good agreement with the estimations provided by the model. The coupling of the lens absorber prototypes to an incident plane wave, i.e. aperture efficiency, is also indirectly validated experimentally matching the expected value of $54\%$ averaged over two polarisation. The noise equivalent power of the KIDs are also measured with limiting value of $8\times10^{-20}$ $\mathrm{W/\sqrt{Hz}}$. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:05:00 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Dabironezare",
"Shahab O.",
""
],
[
"Conenna",
"Giulia",
""
],
[
"Roos",
"Daan",
""
],
[
"Lamers",
"Dimitry",
""
],
[
"Capelo",
"Daniela Perez",
""
],
[
"Veen",
"Hendrik M.",
""
],
[
"Thoen",
"David J.",
""
],
[
"Anvekar",
"Vishal",
""
],
[
"Yates",
"Stephen J. C.",
""
],
[
"Jellema",
"Willem",
""
],
[
"Huiting",
"Robert",
""
],
[
"Ferrari",
"Lorenza",
""
],
[
"Tucker",
"Carole",
""
],
[
"van Berkel",
"Sven L.",
""
],
[
"Day",
"Peter K.",
""
],
[
"Leduc",
"Henry G.",
""
],
[
"Bradford",
"Charles M.",
""
],
[
"Llombart",
"Nuria",
""
],
[
"Baselmans",
"Jochem J. A.",
""
]
] |
2506.03360 | Zihui Ma | Zihui Ma, Lingyao Li, Juan Li, Wenyue Hua, Jingxiao Liu, Qingyuan Feng, Yuki Miura | A Multimodal, Multilingual, and Multidimensional Pipeline for Fine-grained Crowdsourcing Earthquake Damage Evaluation | null | null | null | null | cs.CL cs.CY cs.SI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Rapid, fine-grained disaster damage assessment is essential for effective emergency response, yet remains challenging due to limited ground sensors and delays in official reporting. Social media provides a rich, real-time source of human-centric observations, but its multimodal and unstructured nature presents challenges for traditional analytical methods. In this study, we propose a structured Multimodal, Multilingual, and Multidimensional (3M) pipeline that leverages multimodal large language models (MLLMs) to assess disaster impacts. We evaluate three foundation models across two major earthquake events using both macro- and micro-level analyses. Results show that MLLMs effectively integrate image-text signals and demonstrate a strong correlation with ground-truth seismic data. However, performance varies with language, epicentral distance, and input modality. This work highlights the potential of MLLMs for disaster assessment and provides a foundation for future research in applying MLLMs to real-time crisis contexts. The code and data are released at: https://github.com/missa7481/EMNLP25_earthquake | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:07:25 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ma",
"Zihui",
""
],
[
"Li",
"Lingyao",
""
],
[
"Li",
"Juan",
""
],
[
"Hua",
"Wenyue",
""
],
[
"Liu",
"Jingxiao",
""
],
[
"Feng",
"Qingyuan",
""
],
[
"Miura",
"Yuki",
""
]
] |
2506.03361 | Julia Shapiro | Giuseppe Cotardo, Gretchen L. Matthews, Alberto Ravagnani, Julia Shapiro | Multishot Capacity of Networks with Restricted Adversaries | null | null | null | null | cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | We investigate adversarial network coding and decoding, focusing on the multishot regime and when the adversary is restricted to operate on a vulnerable region of the network. Errors can occur on a proper subset of the network edges and are modeled via an adversarial channel. The paper contains both bounds and capacity-achieving schemes for the Diamond Network, the Mirrored Diamond Network, and generalizations of these networks. We also initiate the study of the capacity of 3-level networks in the multishot setting by computing the multishot capacity of the Butterfly Network, considered in [IEEE Transactions on Information Theory, vol. 69, no. 6, 2023], which is a variant of the network introduced by Ahlswede, Cai, Li and Yeung in 2000. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:12:52 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cotardo",
"Giuseppe",
""
],
[
"Matthews",
"Gretchen L.",
""
],
[
"Ravagnani",
"Alberto",
""
],
[
"Shapiro",
"Julia",
""
]
] |
2506.03362 | Yifei Dong | Yifei Dong, Yan Zhang, Sylvain Calinon, and Florian T. Pokorny | Robustness-Aware Tool Selection and Manipulation Planning with Learned Energy-Informed Guidance | null | null | null | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | Humans subconsciously choose robust ways of selecting and using tools, based on years of embodied experience -- for example, choosing a ladle instead of a flat spatula to serve meatballs. However, robustness under uncertainty remains underexplored in robotic tool-use planning. This paper presents a robustness-aware framework that jointly selects tools and plans contact-rich manipulation trajectories, explicitly optimizing for robustness against environmental disturbances. At the core of our approach is a learned, energy-based robustness metric, which guides the planner towards robust manipulation behaviors. We formulate a hierarchical optimization pipeline that first identifies a tool and configuration that optimizes robustness, and then plans a corresponding manipulation trajectory that maintains robustness throughout execution. We evaluate our approach across three representative tool-use tasks. Simulation and real-world results demonstrate that our approach consistently selects robust tools and generates disturbance-resilient manipulation plans. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:14:49 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Dong",
"Yifei",
""
],
[
"Zhang",
"Yan",
""
],
[
"Calinon",
"Sylvain",
""
],
[
"Pokorny",
"Florian T.",
""
]
] |
2506.03363 | Divya Shyamal | Divya Shyamal, Jiaqi Zhang, Caroline Uhler | Probabilistic Factorial Experimental Design for Combinatorial Interventions | null | null | null | null | cs.LG stat.ME stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A combinatorial intervention, consisting of multiple treatments applied to a single unit with potentially interactive effects, has substantial applications in fields such as biomedicine, engineering, and beyond. Given $p$ possible treatments, conducting all possible $2^p$ combinatorial interventions can be laborious and quickly becomes infeasible as $p$ increases. Here we introduce probabilistic factorial experimental design, formalized from how scientists perform lab experiments. In this framework, the experimenter selects a dosage for each possible treatment and applies it to a group of units. Each unit independently receives a random combination of treatments, sampled from a product Bernoulli distribution determined by the dosages. Additionally, the experimenter can carry out such experiments over multiple rounds, adapting the design in an active manner. We address the optimal experimental design problem within an intervention model that imposes bounded-degree interactions between treatments. In the passive setting, we provide a closed-form solution for the near-optimal design. Our results prove that a dosage of $\tfrac{1}{2}$ for each treatment is optimal up to a factor of $1+O(\tfrac{\ln(n)}{n})$ for estimating any $k$-way interaction model, regardless of $k$, and imply that $O\big(kp^{3k}\ln(p)\big)$ observations are required to accurately estimate this model. For the multi-round setting, we provide a near-optimal acquisition function that can be numerically optimized. We also explore several extensions of the design problem and finally validate our findings through simulations. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:15:08 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Shyamal",
"Divya",
""
],
[
"Zhang",
"Jiaqi",
""
],
[
"Uhler",
"Caroline",
""
]
] |
2506.03364 | Orchid Chetia Phukan | Orchid Chetia Phukan, Girish, Mohd Mujtaba Akhtar, Swarup Ranjan Behera, Priyabrata Mallick, Pailla Balakrishna Reddy, Arun Balaji Buduru, Rajesh Sharma | Towards Source Attribution of Singing Voice Deepfake with Multimodal Foundation Models | Accepted to INTERSPEECH 2025 | null | null | null | eess.AS cs.MM cs.SD | http://creativecommons.org/licenses/by/4.0/ | In this work, we introduce the task of singing voice deepfake source attribution (SVDSA). We hypothesize that multimodal foundation models (MMFMs) such as ImageBind, LanguageBind will be most effective for SVDSA as they are better equipped for capturing subtle source-specific characteristics-such as unique timbre, pitch manipulation, or synthesis artifacts of each singing voice deepfake source due to their cross-modality pre-training. Our experiments with MMFMs, speech foundation models and music foundation models verify the hypothesis that MMFMs are the most effective for SVDSA. Furthermore, inspired from related research, we also explore fusion of foundation models (FMs) for improved SVDSA. To this end, we propose a novel framework, COFFE which employs Chernoff Distance as novel loss function for effective fusion of FMs. Through COFFE with the symphony of MMFMs, we attain the topmost performance in comparison to all the individual FMs and baseline fusion methods. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:16:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Phukan",
"Orchid Chetia",
""
],
[
"Girish",
"",
""
],
[
"Akhtar",
"Mohd Mujtaba",
""
],
[
"Behera",
"Swarup Ranjan",
""
],
[
"Mallick",
"Priyabrata",
""
],
[
"Reddy",
"Pailla Balakrishna",
""
],
[
"Buduru",
"Arun Balaji",
""
],
[
"Sharma",
"Rajesh",
""
]
] |
2506.03365 | Artur Grigorev | Artur Grigorev and Adriana-Simona Mihaita | Urban Visibility Hotspots: Quantifying Building Vertex Visibility from Connected Vehicle Trajectories using Spatial Indexing | null | null | null | null | eess.SY cs.CV cs.SY stat.CO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Effective placement of Out-of-Home advertising and street furniture requires accurate identification of locations offering maximum visual exposure to target audiences, particularly vehicular traffic. Traditional site selection methods often rely on static traffic counts or subjective assessments. This research introduces a data-driven methodology to objectively quantify location visibility by analyzing large-scale connected vehicle trajectory data (sourced from Compass IoT) within urban environments. We model the dynamic driver field-of-view using a forward-projected visibility area for each vehicle position derived from interpolated trajectories. By integrating this with building vertex locations extracted from OpenStreetMap, we quantify the cumulative visual exposure, or ``visibility count'', for thousands of potential points of interest near roadways. The analysis reveals that visibility is highly concentrated, identifying specific ``visual hotspots'' that receive disproportionately high exposure compared to average locations. The core technical contribution involves the construction of a BallTree spatial index over building vertices. This enables highly efficient (O(logN) complexity) radius queries to determine which vertices fall within the viewing circles of millions of trajectory points across numerous trips, significantly outperforming brute-force geometric checks. Analysis reveals two key findings: 1) Visibility is highly concentrated, identifying distinct 'visual hotspots' receiving disproportionately high exposure compared to average locations. 2) The aggregated visibility counts across vertices conform to a Log-Normal distribution. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:16:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Grigorev",
"Artur",
""
],
[
"Mihaita",
"Adriana-Simona",
""
]
] |
2506.03366 | Matthieu Pinaud | Matthieu F. Pinaud | Manifolds of mappings associated with real-valued function spaces and natural mappings between them | 35 pages | null | null | null | math.DG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $M$ be a compact smooth manifold with corners and $N$ be a finite dimensional smooth manifold without boundary which admits local addition. We define a smooth manifold structure to general sets of continuous mapings $\mathcal{F}(M,N)$ whenever functions spaces $\mathcal{F}(U,\mathbb{R})$ on open subsets $U\subseteq [0,\infty)^n$ are given, subject to simple axioms. Construction and properties of spaces of sections and smoothness of natural mappings between spaces $\mathcal{F}(M,N)$ are discussed, like superposition operators $\mathcal{F}(M,f):\mathcal{F}(M,N_1)\to \mathcal{F}(M,N_2)$, $\eta \mapsto f\circ \eta$ for smooth maps $f:N_1\to N_2$. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:19:44 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Pinaud",
"Matthieu F.",
""
]
] |
2506.03367 | Noman Bashir | Wacuka Ngata and Noman Bashir and Michelle Westerlaken and Laurent Liote and Yasra Chandio and Elsa Olivetti | The Cloud Next Door: Investigating the Environmental and Socioeconomic Strain of Datacenters on Local Communities | to appear at ACM COMPASS 2025 as a Late Breaking Work | null | 10.1145/3715335.3736324 10.1145/3715335.3736324 | null | cs.DC cs.CY cs.ET | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Datacenters have become the backbone of modern digital infrastructure, powering the rapid rise of artificial intelligence and promising economic growth and technological progress. However, this expansion has brought growing tensions in the local communities where datacenters are already situated or being proposed. While the mainstream discourse often focuses on energy usage and carbon footprint of the computing sector at a global scale, the local socio-environmental consequences -- such as health impacts, water usage, noise pollution, infrastructural strain, and economic burden -- remain largely underexplored and poorly addressed. In this work, we surface these community-level consequences through a mixed-methods study that combines quantitative data with qualitative insights. Focusing on Northern Virginia's ``Data Center Valley,'' we highlight how datacenter growth reshapes local environments and everyday life, and examine the power dynamics that determine who benefits and who bears the costs. Our goal is to bring visibility to these impacts and prompt more equitable and informed decisions about the future of digital infrastructure. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:21:53 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ngata",
"Wacuka",
""
],
[
"Bashir",
"Noman",
""
],
[
"Westerlaken",
"Michelle",
""
],
[
"Liote",
"Laurent",
""
],
[
"Chandio",
"Yasra",
""
],
[
"Olivetti",
"Elsa",
""
]
] |
2506.03368 | Thilo Kuessner | Thilo Kuessner, Janusz Przewocki, Andreas Zastrow | On measure homology of mildly wild spaces | 70 pages | null | null | null | math.AT math.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove injectivity of the canonical map from singular homology to measure homology for certain ``mildly wild" spaces, that is, certain spaces not having the homotopy type of a CW-complex, but having countable fundamental groups. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:25:50 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Kuessner",
"Thilo",
""
],
[
"Przewocki",
"Janusz",
""
],
[
"Zastrow",
"Andreas",
""
]
] |
2506.03369 | Akshit Kumar | Omar Besbes, Yash Kanoria, Akshit Kumar | Impact of Rankings and Personalized Recommendations in Marketplaces | null | null | null | null | econ.TH cs.CY cs.IR | http://creativecommons.org/licenses/by/4.0/ | Individuals often navigate several options with incomplete knowledge of their own preferences. Information provisioning tools such as public rankings and personalized recommendations have become central to helping individuals make choices, yet their value proposition under different marketplace environments remains unexplored. This paper studies a stylized model to explore the impact of these tools in two marketplace settings: uncapacitated supply, where items can be selected by any number of agents, and capacitated supply, where each item is constrained to be matched to a single agent. We model the agents utility as a weighted combination of a common term which depends only on the item, reflecting the item's population level quality, and an idiosyncratic term, which depends on the agent item pair capturing individual specific tastes. Public rankings reveal the common term, while personalized recommendations reveal both terms. In the supply unconstrained settings, both public rankings and personalized recommendations improve welfare, with their relative value determined by the degree of preference heterogeneity. Public rankings are effective when preferences are relatively homogeneous, while personalized recommendations become critical as heterogeneity increases. In contrast, in supply constrained settings, revealing just the common term, as done by public rankings, provides limited benefit since the total common value available is limited by capacity constraints, whereas personalized recommendations, by revealing both common and idiosyncratic terms, significantly enhance welfare by enabling agents to match with items they idiosyncratically value highly. These results illustrate the interplay between supply constraints and preference heterogeneity in determining the effectiveness of information provisioning tools, offering insights for their design and deployment in diverse settings. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:26:14 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Besbes",
"Omar",
""
],
[
"Kanoria",
"Yash",
""
],
[
"Kumar",
"Akshit",
""
]
] |
2506.03370 | Leonid Ryvkin | Leonid Ryvkin | Comparison of different Unique hard attention transformer models by the formal languages they can recognize | null | null | null | null | cs.LG cs.CL cs.FL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This note is a survey of various results on the capabilities of unique hard attention transformers encoders (UHATs) to recognize formal languages. We distinguish between masked vs. non-masked, finite vs. infinite image and general vs. bilinear attention score functions. We recall some relations between these models, as well as a lower bound in terms of first-order logic and an upper bound in terms of circuit complexity. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:28:51 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ryvkin",
"Leonid",
""
]
] |
2506.03371 | Xiaonan Wang | Xiaonan Wang, Bo Shao and Hansaem Kim | Toward Reliable VLM: A Fine-Grained Benchmark and Framework for Exposure, Bias, and Inference in Korean Street Views | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in vision-language models (VLMs) have enabled accurate image-based geolocation, raising serious concerns about location privacy risks in everyday social media posts. However, current benchmarks remain coarse-grained, linguistically biased, and lack multimodal and privacy-aware evaluations. To address these gaps, we present KoreaGEO Bench, the first fine-grained, multimodal geolocation benchmark for Korean street views. Our dataset comprises 1,080 high-resolution images sampled across four urban clusters and nine place types, enriched with multi-contextual annotations and two styles of Korean captions simulating real-world privacy exposure. We introduce a three-path evaluation protocol to assess ten mainstream VLMs under varying input modalities and analyze their accuracy, spatial bias, and reasoning behavior. Results reveal modality-driven shifts in localization precision and highlight structural prediction biases toward core cities. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:28:55 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"Xiaonan",
""
],
[
"Shao",
"Bo",
""
],
[
"Kim",
"Hansaem",
""
]
] |
2506.03372 | Levi Smith | Ali Dehghanian and Thomas Darcie and Levi Smith | Genetic Algorithm-Based Inverse Design of Guided Wave Planar Terahertz Filters | 10 pages, 8 figures | null | null | null | physics.app-ph physics.optics | http://creativecommons.org/licenses/by/4.0/ | We present a genetic algorithm (GA)-based inverse design framework for synthesizing high-performance planar terahertz (THz) filters integrated with coplanar striplines (CPSs). The method efficiently explores high-dimensional design spaces to generate filter geometries matching user-defined S-parameter magnitude and phase responses, while enforcing structural connectivity for compatibility with terahertz system-on-chip (TSoC) platforms. To accelerate optimization, filter performance is evaluated using the ABCD matrix method, providing a significant computational advantage over full-wave simulations. Final validation is performed through finite element method (FEM) simulations. As a proof of concept, we design band-stop filters with center frequencies of 0.6, 0.8, and 1.0 THz, each with a 150 GHz target bandwidth, and demonstrate tunable rejection depths within a constant physical footprint. Optimization is guided by minimizing the root-mean-square error (RMSE) between simulated and target S-parameters. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:29:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Dehghanian",
"Ali",
""
],
[
"Darcie",
"Thomas",
""
],
[
"Smith",
"Levi",
""
]
] |
2506.03373 | Muhammad Shaban | Muhammad Shaban, Yuzhou Chang, Huaying Qiu, Yao Yu Yeo, Andrew H. Song, Guillaume Jaume, Yuchen Wang, Luca L. Weishaupt, Tong Ding, Anurag Vaidya, Abdallah Lamane, Daniel Shao, Mohammed Zidane, Yunhao Bai, Paige McCallum, Shuli Luo, Wenrui Wu, Yang Wang, Precious Cramer, Chi Ngai Chan, Pierre Stephan, Johanna Schaffenrath, Jia Le Lee, Hendrik A. Michel, Caiwei Tian, Cristina Almagro-Perez, Sophia J. Wagner, Sharifa Sahai, Ming Y. Lu, Richard J. Chen, Andrew Zhang, Mark Edward M. Gonzales, Ahmad Makky, Jia-Ying Joey Lee, Hao Cheng, Nourhan El Ahmar, Sayed Matar, Maximilian Haist, Darci Phillips, Yuqi Tan, Garry P. Nolan, W. Richard Burack, Jacob D. Estes, Jonathan T.C. Liu, Toni K Choueiri, Neeraj Agarwal, Marc Barry, Scott J. Rodig, Long Phi Le, Georg Gerber, Christian M. Sch\"urch, Fabian J. Theis, Youn H Kim, Joe Yeong, Sabina Signoretti, Brooke E. Howitt, Lit-Hsin Loo, Qin Ma, Sizun Jiang, Faisal Mahmood | A Foundation Model for Spatial Proteomics | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Foundation models have begun to transform image analysis by acting as pretrained generalist backbones that can be adapted to many tasks even when post-training data are limited, yet their impact on spatial proteomics, imaging that maps proteins at single-cell resolution, remains limited. Here, we introduce KRONOS, a foundation model built for spatial proteomics. KRONOS was trained in a self-supervised manner on over 47 million image patches covering 175 protein markers, 16 tissue types, and 8 fluorescence-based imaging platforms. We introduce key architectural adaptations to address the high-dimensional, multi-channel, and heterogeneous nature of multiplex imaging. We demonstrate that KRONOS learns biologically meaningful representations across multiple scales, ranging from cellular and microenvironment to tissue levels, enabling it to address diverse downstream tasks, including cell phenotyping, region classification, and patient stratification. Evaluated across 11 independent cohorts, KRONOS achieves state-of-the-art performance across cell phenotyping, treatment response prediction, and retrieval tasks, and is highly data-efficient. KRONOS also introduces the paradigm of segmentation-free patch-level processing for efficient and scalable spatial proteomics analysis, allowing cross-institutional comparisons, and as an image reverse search engine for spatial patterns. Together, these results position KRONOS as a flexible and scalable tool for spatial proteomics. The model is publicly accessible at https://github.com/mahmoodlab/KRONOS. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:30:25 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Shaban",
"Muhammad",
""
],
[
"Chang",
"Yuzhou",
""
],
[
"Qiu",
"Huaying",
""
],
[
"Yeo",
"Yao Yu",
""
],
[
"Song",
"Andrew H.",
""
],
[
"Jaume",
"Guillaume",
""
],
[
"Wang",
"Yuchen",
""
],
[
"Weishaupt",
"Luca L.",
""
],
[
"Ding",
"Tong",
""
],
[
"Vaidya",
"Anurag",
""
],
[
"Lamane",
"Abdallah",
""
],
[
"Shao",
"Daniel",
""
],
[
"Zidane",
"Mohammed",
""
],
[
"Bai",
"Yunhao",
""
],
[
"McCallum",
"Paige",
""
],
[
"Luo",
"Shuli",
""
],
[
"Wu",
"Wenrui",
""
],
[
"Wang",
"Yang",
""
],
[
"Cramer",
"Precious",
""
],
[
"Chan",
"Chi Ngai",
""
],
[
"Stephan",
"Pierre",
""
],
[
"Schaffenrath",
"Johanna",
""
],
[
"Lee",
"Jia Le",
""
],
[
"Michel",
"Hendrik A.",
""
],
[
"Tian",
"Caiwei",
""
],
[
"Almagro-Perez",
"Cristina",
""
],
[
"Wagner",
"Sophia J.",
""
],
[
"Sahai",
"Sharifa",
""
],
[
"Lu",
"Ming Y.",
""
],
[
"Chen",
"Richard J.",
""
],
[
"Zhang",
"Andrew",
""
],
[
"Gonzales",
"Mark Edward M.",
""
],
[
"Makky",
"Ahmad",
""
],
[
"Lee",
"Jia-Ying Joey",
""
],
[
"Cheng",
"Hao",
""
],
[
"Ahmar",
"Nourhan El",
""
],
[
"Matar",
"Sayed",
""
],
[
"Haist",
"Maximilian",
""
],
[
"Phillips",
"Darci",
""
],
[
"Tan",
"Yuqi",
""
],
[
"Nolan",
"Garry P.",
""
],
[
"Burack",
"W. Richard",
""
],
[
"Estes",
"Jacob D.",
""
],
[
"Liu",
"Jonathan T. C.",
""
],
[
"Choueiri",
"Toni K",
""
],
[
"Agarwal",
"Neeraj",
""
],
[
"Barry",
"Marc",
""
],
[
"Rodig",
"Scott J.",
""
],
[
"Le",
"Long Phi",
""
],
[
"Gerber",
"Georg",
""
],
[
"Schürch",
"Christian M.",
""
],
[
"Theis",
"Fabian J.",
""
],
[
"Kim",
"Youn H",
""
],
[
"Yeong",
"Joe",
""
],
[
"Signoretti",
"Sabina",
""
],
[
"Howitt",
"Brooke E.",
""
],
[
"Loo",
"Lit-Hsin",
""
],
[
"Ma",
"Qin",
""
],
[
"Jiang",
"Sizun",
""
],
[
"Mahmood",
"Faisal",
""
]
] |
2506.03374 | Haley Dozier | Haley Dozier, Althea Henslee, Ashley Abraham, Andrew Strelzoff, Mark Chappell | Product Quantization for Surface Soil Similarity | To be published in the CSCE 2022 proceedings | null | null | null | cs.LG | http://creativecommons.org/publicdomain/zero/1.0/ | The use of machine learning (ML) techniques has allowed rapid advancements in many scientific and engineering fields. One of these problems is that of surface soil taxonomy, a research area previously hindered by the reliance on human-derived classifications, which are mostly dependent on dividing a dataset based on historical understandings of that data rather than data-driven, statistically observable similarities. Using a ML-based taxonomy allows soil researchers to move beyond the limitations of human visualization and create classifications of high-dimension datasets with a much higher level of specificity than possible with hand-drawn taxonomies. Furthermore, this pipeline allows for the possibility of producing both highly accurate and flexible soil taxonomies with classes built to fit a specific application. The machine learning pipeline outlined in this work combines product quantization with the systematic evaluation of parameters and output to get the best available results, rather than accepting sub-optimal results by using either default settings or best guess settings. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:31:34 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Dozier",
"Haley",
""
],
[
"Henslee",
"Althea",
""
],
[
"Abraham",
"Ashley",
""
],
[
"Strelzoff",
"Andrew",
""
],
[
"Chappell",
"Mark",
""
]
] |
2506.03375 | Alan Frieze | Colin Cooper, Alan Frieze, Wesley Pegden | Cover time of random subgraphs of the hypercube | null | null | null | null | math.CO cs.DS | http://creativecommons.org/licenses/by/4.0/ | $Q_{n,p}$, the random subgraph of the $n$-vertex hypercube $Q_n$, is obtained by independently retaining each edge of $Q_n$ with probability $p$. We give precise values for the cover time of $Q_{n,p}$ above the connectivity threshold. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:32:45 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cooper",
"Colin",
""
],
[
"Frieze",
"Alan",
""
],
[
"Pegden",
"Wesley",
""
]
] |
2506.03376 | Holly B. Parkinson | Holly B. Parkinson (1) for the MicroBooNE Collaboration ((1) University of Edinburgh) | Recent results from MicroBooNE | The 43rd International Symposium on Physics in Collision (PIC2024) | null | null | FERMILAB-CONF-25-0013-V | hep-ex physics.acc-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Modelling and reconstructing neutrino-nucleus scattering is difficult, but it is crucial to do it precisely to enable next-generation oscillation measurements. Liquid argon time projection chambers (LArTPCs), such as MicroBooNE, can be the tool for this job as they are excellent precision neutrino detectors with their ability to produce detailed three-dimensional interaction images and precise energy and spatial resolution. MicroBooNE currently possesses the world's largest neutrino-argon scattering data set collected over five years using the BNB and NuMI neutrino beams at Fermilab. The experiment has performed measurements over a broad range of physics topics including neutrino argon cross sections, searches for BSM physics, and investigations of the MiniBooNE LEE excess. Many of these measurements are essential for improving the modelling of nuclear effects for both MicroBooNE and future LArTPC neutrino experiments, such as DUNE. This talk will give an overview of recent MicroBooNE results, the analysis techniques that enable them, and prospects for future measurements. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:33:50 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Parkinson",
"Holly B.",
"",
"University of Edinburgh"
],
[
"Collaboration",
"for the MicroBooNE",
""
]
] |
2506.03377 | Peio Ardaiz-Gale | Peio Ardaiz Gal\'e and Conchita Mart\'inez P\'erez | The McCullough-Miller complex for right angled Artin groups | 34 pages | null | null | null | math.GR | http://creativecommons.org/licenses/by/4.0/ | McCullough and Miller constructed a contractible complex on which the pure symmetric automorphism group of a free group acts with free abelian stabilizers. This complex has been used for computations such as the cohomological dimension of these groups, their cohomology rings, and results about $\ell^2$-Betti numbers or BNRS-invariants. We generalize this construction to pure symmetric automorphism groups of arbitrary RAAGs and exhibit applications of this generalization. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:36:35 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Galé",
"Peio Ardaiz",
""
],
[
"Pérez",
"Conchita Martínez",
""
]
] |
2506.03378 | Orchid Chetia Phukan | Orchid Chetia Phukan, Mohd Mujtaba Akhtar, Girish, Swarup Ranjan Behera, Abu Osama Siddiqui, Sarthak Jain, Priyabrata Mallick, Jaya Sai Kiran Patibandla, Pailla Balakrishna Reddy, Arun Balaji Buduru, Rajesh Sharma | SNIFR : Boosting Fine-Grained Child Harmful Content Detection Through Audio-Visual Alignment with Cascaded Cross-Transformer | Accepted to INTERSPEECH 2025 | null | null | null | eess.AS cs.CV cs.MM | http://creativecommons.org/licenses/by/4.0/ | As video-sharing platforms have grown over the past decade, child viewership has surged, increasing the need for precise detection of harmful content like violence or explicit scenes. Malicious users exploit moderation systems by embedding unsafe content in minimal frames to evade detection. While prior research has focused on visual cues and advanced such fine-grained detection, audio features remain underexplored. In this study, we embed audio cues with visual for fine-grained child harmful content detection and introduce SNIFR, a novel framework for effective alignment. SNIFR employs a transformer encoder for intra-modality interaction, followed by a cascaded cross-transformer for inter-modality alignment. Our approach achieves superior performance over unimodal and baseline fusion methods, setting a new state-of-the-art. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:37:23 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Phukan",
"Orchid Chetia",
""
],
[
"Akhtar",
"Mohd Mujtaba",
""
],
[
"Girish",
"",
""
],
[
"Behera",
"Swarup Ranjan",
""
],
[
"Siddiqui",
"Abu Osama",
""
],
[
"Jain",
"Sarthak",
""
],
[
"Mallick",
"Priyabrata",
""
],
[
"Patibandla",
"Jaya Sai Kiran",
""
],
[
"Reddy",
"Pailla Balakrishna",
""
],
[
"Buduru",
"Arun Balaji",
""
],
[
"Sharma",
"Rajesh",
""
]
] |
2506.03379 | Zu-Jian Ying | Zu-Jian Ying | Globalized Nonlinear Critical Quantum Metrology by Two-photon Rabi-Stark model | 10 pages, 8 figures | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Squeezing as a quantum resource for quantum metrology is robust against decoherence and dissipation, while the conventional nonlinear two-photon quantum Rabi model (QRM) provides a squeezing resource immune to the divergence problem of preparation time of probe state (PTPS). However the critical point of the two-photon QRM is locally restricted to one single point, which hinders a wider application. In the present work we propose to combine the Stark coupling with the two-photon QRM to realize a tunable critical point so that the nonlinear critical quantum metrology can be globalized. As demonstrated by the diverging quantum Fisher information (QFI) the protocol enables us to acquire a high measurement precision in a wide range of coupling parameter rather than locally at a single critical point. Moreover, We find that the QFI not only manifests criticality but also exhibits universality. As a particular merit of our protocol, a strong squeezing can be globally retained as the leading quantum resource, while at the same time the PTPS remains in a finite order. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:38:06 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ying",
"Zu-Jian",
""
]
] |
2506.03380 | Zach Patterson | Zach J. Patterson, Emily R. Sologuren, Daniela Rus | Design of Trimmed Helicoid Soft-Rigid Hybrid Robots | 7 pgs. 5 figs. Presented at IEEE Robosoft 2025 | null | null | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | As soft robot design matures, researchers have converged to sophisticated design paradigms to enable the development of more suitable platforms. Two such paradigms are soft-rigid hybrid robots, which utilize rigid structural materials in some aspect of the robot's design, and architectured materials, which deform based on geometric parameters as opposed to purely material ones. In this work, we combine the two design approaches, utilizing trimmed helicoid structures in series with rigid linkages. Additionally, we extend the literature on wave spring-inspired soft structures by deriving a mechanical model of the stiffness for arbitrary geometries. We present a novel manufacturing method for such structures utilizing an injection molding approach and we make available the design tool to generate 3D printed molds for arbitrary designs of this class. Finally, we produce a robot using the above methods and operate it in closed-loop demonstrations. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:39:56 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Patterson",
"Zach J.",
""
],
[
"Sologuren",
"Emily R.",
""
],
[
"Rus",
"Daniela",
""
]
] |
2506.03381 | Artur Grigorev | Artur Grigorev and Khaled Saleh and Jiwon Kim and Adriana-Simona Mihaita | Automated Traffic Incident Response Plans using Generative Artificial Intelligence: Part 1 -- Building the Incident Response Benchmark | null | null | null | null | eess.SY cs.AI cs.LG cs.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Traffic incidents remain a critical public safety concern worldwide, with Australia recording 1,300 road fatalities in 2024, which is the highest toll in 12 years. Similarly, the United States reports approximately 6 million crashes annually, raising significant challenges in terms of a fast reponse time and operational management. Traditional response protocols rely on human decision-making, which introduces potential inconsistencies and delays during critical moments when every minute impacts both safety outcomes and network performance. To address this issue, we propose a novel Incident Response Benchmark that uses generative artificial intelligence to automatically generate response plans for incoming traffic incidents. Our approach aims to significantly reduce incident resolution times by suggesting context-appropriate actions such as variable message sign deployment, lane closures, and emergency resource allocation adapted to specific incident characteristics. First, the proposed methodology uses real-world incident reports from the Performance Measurement System (PeMS) as training and evaluation data. We extract historically implemented actions from these reports and compare them against AI-generated response plans that suggest specific actions, such as lane closures, variable message sign announcements, and/or dispatching appropriate emergency resources. Second, model evaluations reveal that advanced generative AI models like GPT-4o and Grok 2 achieve superior alignment with expert solutions, demonstrated by minimized Hamming distances (averaging 2.96-2.98) and low weighted differences (approximately 0.27-0.28). Conversely, while Gemini 1.5 Pro records the lowest count of missed actions, its extremely high number of unnecessary actions (1547 compared to 225 for GPT-4o) indicates an over-triggering strategy that reduces the overall plan efficiency. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:40:44 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Grigorev",
"Artur",
""
],
[
"Saleh",
"Khaled",
""
],
[
"Kim",
"Jiwon",
""
],
[
"Mihaita",
"Adriana-Simona",
""
]
] |
2506.03382 | Luca Roversi | Matteo Palazzo and Luca Roversi | Towards a Characterization of Two-way Bijections in a Reversible Computational Model | 8 pages, 3 figures, 5 listings. Author's copy of the version which will appear in the Proceedings of the 17th International Conference, RC 2025, Odense, Denmark, July 3-4, 2025 | null | null | null | cs.LO cs.CC cs.PL | http://creativecommons.org/licenses/by/4.0/ | We introduce an imperative, stack-based, and reversible computational model that characterizes Two-way Bijections both implicitly, concerning their computational complexity, and with zero-garbage. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:40:45 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Palazzo",
"Matteo",
""
],
[
"Roversi",
"Luca",
""
]
] |
2506.03383 | Giorgio Parisi | Giorgio Parisi | From Bjorken Scaling to Scaling Violations | 14 pages, 1 figure | null | null | null | physics.hist-ph hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | This paper traces the historical and conceptual journey from Bjorken scaling to the discovery of scaling violations in deep inelastic scattering, culminating in the development of Quantum Chromodynamics (QCD). Beginning with the challenges faced by early strong interaction theories in the 1950s, we explore the emergence of agnostic approaches such as the bootstrap philosophy and current algebra, which sought to describe hadronic phenomena without relying on specific field theories. The pivotal role of experimental results from SLAC in the late 1960s is highlighted, leading to Bjorken's proposal of scaling in deep inelastic scattering and Feynman's parton model. We then delve into the theoretical breakthroughs of the 1970s, including Wilson's operator product expansion and the renormalization group, which provided the framework for understanding scaling violations. The discovery of asymptotic freedom in non-Abelian gauge theories by Gross, Wilczek, and Politzer marked a turning point, establishing QCD as the theory of strong interactions. Finally, we discuss the formulation of the Altarelli-Parisi equations, which elegantly describe the evolution of parton distribution functions and scaling violations, and their profound impact on the study of hard processes in particle physics. This paper not only recounts the key developments but also reflects on the interplay between theory and experiment that drove the field forward. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:42:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Parisi",
"Giorgio",
""
]
] |
2506.03384 | Bryce Morsky | Bryce Morsky | How urban scaling and resource distribution shape social welfare and migration dynamics | null | null | null | null | physics.soc-ph econ.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many outputs of cities scale in universal ways, including infrastructure, crime, and economic activity. Through a mathematical model, this study investigates the interplay between such scaling laws in human organization and governmental allocations of resources, focusing on impacts to migration patterns and social welfare. We find that if superlinear scaling resources of cities -- such as economic and social activity -- are the primary drivers of city dwellers' utility, then cities tend to converge to similar sizes and social welfare through migration. In contrast, if sublinear scaling resources, such as infrastructure, primarily impact utility, then migration tends to lead to megacities and inequity between large and small cities. These findings have implications for policymakers, economists, and political scientists addressing the challenges of equitable and efficient resource allocation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:43:02 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Morsky",
"Bryce",
""
]
] |
2506.03385 | Oliver Huang | Oliver Huang and Patrick Lee and Carolina Nobre | From Reality to Recognition: Evaluating Visualization Analogies for Novice Chart Comprehension | null | null | 10.2312/eved.20251004 | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | Novice learners often have difficulty learning new visualization types because they tend to interpret novel visualizations through the mental models of simpler charts they have previously encountered. Traditional visualization teaching methods, which usually rely on directly translating conceptual aspects of data into concrete data visualizations, often fail to attend to the needs of novice learners navigating this tension. To address this, we conducted an empirical exploration of how analogies can be used to help novices with chart comprehension. We introduced visualization analogies: visualizations that map data structures to real-world contexts to facilitate an intuitive understanding of novel chart types. We evaluated this pedagogical technique using a within-subject study (N=128) where we taught 8 chart types using visualization analogies. Our findings show that visualization analogies improve visual analysis skills and help learners transfer their understanding to actual charts. They effectively introduce visual embellishments, cater to diverse learning preferences, and are preferred by novice learners over traditional chart visualizations. This study offers empirical insights and open-source tools to advance visualization education through analogical reasoning. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:44:31 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Huang",
"Oliver",
""
],
[
"Lee",
"Patrick",
""
],
[
"Nobre",
"Carolina",
""
]
] |
2506.03386 | Yinon Ashkenazy | Rann Shikler, Noya Dimanstein Firman and Yinon Ashkenazy | Model for the Thermodynamics of Iron at High Pressures Near Melting | null | null | null | null | physics.geo-ph cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-sa/4.0/ | The Fe pressure-temperature phase diagram and its melting line have a wide range of applications, including providing constraints for iron-core planetary models. We propose an equation of state (EOS) model based on the interstitial theory of simple condensed matter (ITCM), as suggested by A.V. Granato. When applied to Fe, this model enables the extrapolation of measured melting lines to the conditions of the Earth's inner core boundary (ICB). The ITCM describes the solid-liquid phase transition in metals as resulting from a strong structural perturbation due to a high concentration of interstitial-like defects. The strong nonlinearity of their self-interaction causes the stabilization of this interstitial-rich phase. The original model is expanded to describe melting over a wide range of pressures and temperatures rather than focusing on a specific isobaric transition. Using this model, we fit the measured melting data, extrapolate it to cover ICB conditions, and develop a multiphase equation of state that encompasses this regime. The model is used to explain contradictory data regarding the location of the melting line, resulting from a novel phase transition between two separate liquid phases, specifically between FCC-based and HCP-based liquids. This additional liquid phase offers a new interpretation of the previously suggested near-melting high-pressure phase and may also provide a solution to the inner core nucleation paradox. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:48:21 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Shikler",
"Rann",
""
],
[
"Firman",
"Noya Dimanstein",
""
],
[
"Ashkenazy",
"Yinon",
""
]
] |
2506.03387 | Reika Katsumata | Tara (Tera) Huang, Evon S. Petek, and Reika Katsumata | Nanoconfinement Effects on Intermolecular Forces Observed via Dewetting | null | null | null | null | cond-mat.soft | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Although wettability is a macroscopic manifestation of molecular-level forces, such as van der Waals (vdW) forces, the impact of nanoconfinement on material properties in reduced film thickness remains unexplored in predicting film stability. In this work, we investigate how nanoconfinement influences intermolecular interactions using a model trilayer system composed of a thick polystyrene (PS) base, a poly(methyl methacrylate) (PMMA) middle layer with tunable thickness (15-95 nm), and a 10 nm top PS film. We find that the dewetting behavior of the top PS layer is highly sensitive to middle PMMA thickness, deviating from classical vdW-based predictions that assume bulk material properties. By incorporating nanoconfinement-induced changes in PMMA refractive index into the calculation of the Hamaker constant, we present a modified theoretical framework that successfully captures the observed behavior. This study links dewetting behavior and material property change as a function of underlayer thickness, providing direct evidence that nanoconfinement in soft matter systems significantly influences long-range intermolecular interactions. We show that film stability can be tuned solely by adjusting underlying layer thickness, while preserving both chemistry and thickness of top functional film. This finding carries broad implications for thin-film technologies across scientific and engineering disciplines by enabling performance-targeted interface design. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:55:40 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Tara",
"",
"",
"Tera"
],
[
"Huang",
"",
""
],
[
"Petek",
"Evon S.",
""
],
[
"Katsumata",
"Reika",
""
]
] |
2506.03388 | Pengyu Chen | Pengyu Chen, Xiao Huang, Teng Fei, Sicheng Wang | Cross-Modal Urban Sensing: Evaluating Sound-Vision Alignment Across Street-Level and Aerial Imagery | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Environmental soundscapes convey substantial ecological and social information regarding urban environments; however, their potential remains largely untapped in large-scale geographic analysis. In this study, we investigate the extent to which urban sounds correspond with visual scenes by comparing various visual representation strategies in capturing acoustic semantics. We employ a multimodal approach that integrates geo-referenced sound recordings with both street-level and remote sensing imagery across three major global cities: London, New York, and Tokyo. Utilizing the AST model for audio, along with CLIP and RemoteCLIP for imagery, as well as CLIPSeg and Seg-Earth OV for semantic segmentation, we extract embeddings and class-level features to evaluate cross-modal similarity. The results indicate that street view embeddings demonstrate stronger alignment with environmental sounds compared to segmentation outputs, whereas remote sensing segmentation is more effective in interpreting ecological categories through a Biophony--Geophony--Anthrophony (BGA) framework. These findings imply that embedding-based models offer superior semantic alignment, while segmentation-based methods provide interpretable links between visual structure and acoustic ecology. This work advances the burgeoning field of multimodal urban sensing by offering novel perspectives for incorporating sound into geospatial analysis. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 20:56:37 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Pengyu",
""
],
[
"Huang",
"Xiao",
""
],
[
"Fei",
"Teng",
""
],
[
"Wang",
"Sicheng",
""
]
] |
2506.03389 | Tommaso Lorenzi | Tommaso Lorenzi, Kevin J. Painter | Pattern formation within phenotype-structured chemotactic populations | null | null | null | null | q-bio.PE | http://creativecommons.org/licenses/by/4.0/ | Populations can become spatially organised through chemotaxis autoattraction, wherein population members release their own chemoattractant. Standard models of this process usually assume phenotypic homogeneity, but recent studies have shed illumination on the inherent heterogeneity within populations: in terms of chemotactic behaviour, trait heterogeneity can range from the sensitivity to attractant gradients to the rate at which attractants are produced. We propose a framework that accounts for this heterogeneity, extending the standard Keller-Segel model to a non-local formulation in which the population is continuously structured across some phenotype state space. Focussing on autoattraction, we allow both the chemotactic sensitivity and the rate of attractant secretion to vary across the population and suppose members can switch between different phenotype states. We extend classical Turing-type linear stability analyses to determine the impact of phenotypic structuring on pattern formation, showing that the rate of switching influences both the critical condition for self-organisation and subsequent pattern dynamics. Scenarios in which the chemotactic sensitivity and attractant secretion are positively or negatively correlated are used to highlight the significance of these results. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:00:13 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Lorenzi",
"Tommaso",
""
],
[
"Painter",
"Kevin J.",
""
]
] |
2506.03390 | Bryce Wedig | Bryce Wedig, Tansu Daylan, Simon Birrer, Francis-Yan Cyr-Racine, Cora Dvorkin, Douglas P. Finkbeiner, Alan Huang, Xiaosheng Huang, Rahul Karthik, Narayan Khadka, Priyamvada Natarajan, Anna M. Nierenberg, Annika H. G. Peter, Justin D. R. Pierel, Xianzhe TZ Tang, Risa H. Wechsler | The Roman View of Strong Gravitational Lenses | 23 pages, 12 figures, accepted for publication in The Astrophysical Journal, code at https://github.com/AstroMusers/mejiro, data at https://doi.org/10.5281/zenodo.14216839 | null | null | null | astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Galaxy-galaxy strong gravitational lenses can constrain dark matter models and the Lambda Cold Dark Matter cosmological paradigm at sub-galactic scales. Currently, there is a dearth of images of these rare systems with high signal-to-noise and angular resolution. The Nancy Grace Roman Space Telescope (hereafter, Roman), scheduled for launch in late 2026, will play a transformative role in strong lensing science with its planned wide-field surveys. With its remarkable 0.281 square degree field of view and diffraction-limited angular resolution of ~0.1 arcsec, Roman is uniquely suited to characterizing dark matter substructure from a robust population of strong lenses. We present a yield simulation of detectable strong lenses in Roman's planned High Latitude Wide Area Survey (HLWAS). We simulate a population of galaxy-galaxy strong lenses across cosmic time with Cold Dark Matter subhalo populations, select those detectable in the HLWAS, and generate simulated images accounting for realistic Wide Field Instrument detector effects. For a fiducial case of single 146-second exposures, we predict around 160,000 detectable strong lenses in the HLWAS, of which about 500 will have sufficient signal-to-noise to be amenable to detailed substructure characterization. We investigate the effect of the variation of the point-spread function across Roman's field of view on detecting individual subhalos and the suppression of the subhalo mass function at low masses. Our simulation products are available to support strong lens science with Roman, such as training neural networks and validating dark matter substructure analysis pipelines. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:00:31 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wedig",
"Bryce",
""
],
[
"Daylan",
"Tansu",
""
],
[
"Birrer",
"Simon",
""
],
[
"Cyr-Racine",
"Francis-Yan",
""
],
[
"Dvorkin",
"Cora",
""
],
[
"Finkbeiner",
"Douglas P.",
""
],
[
"Huang",
"Alan",
""
],
[
"Huang",
"Xiaosheng",
""
],
[
"Karthik",
"Rahul",
""
],
[
"Khadka",
"Narayan",
""
],
[
"Natarajan",
"Priyamvada",
""
],
[
"Nierenberg",
"Anna M.",
""
],
[
"Peter",
"Annika H. G.",
""
],
[
"Pierel",
"Justin D. R.",
""
],
[
"Tang",
"Xianzhe TZ",
""
],
[
"Wechsler",
"Risa H.",
""
]
] |
2506.03391 | Tri Kurniawan Wijaya | Tri Kurniawan Wijaya, Xinyang Shao, Gonzalo Fiz Pontiveros, Edoardo D'Amico | Universal Reusability in Recommender Systems: The Case for Dataset- and Task-Independent Frameworks | null | null | null | null | cs.IR cs.AI cs.DB cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recommender systems are pivotal in delivering personalized experiences across industries, yet their adoption and scalability remain hindered by the need for extensive dataset- and task-specific configurations. Existing systems often require significant manual intervention, domain expertise, and engineering effort to adapt to new datasets or tasks, creating barriers to entry and limiting reusability. In contrast, recent advancements in large language models (LLMs) have demonstrated the transformative potential of reusable systems, where a single model can handle diverse tasks without significant reconfiguration. Inspired by this paradigm, we propose the Dataset- and Task-Independent Recommender System (DTIRS), a framework aimed at maximizing the reusability of recommender systems while minimizing barriers to entry. Unlike LLMs, which achieve task generalization directly, DTIRS focuses on eliminating the need to rebuild or reconfigure recommendation pipelines for every new dataset or task, even though models may still need retraining on new data. By leveraging the novel Dataset Description Language (DsDL), DTIRS enables standardized dataset descriptions and explicit task definitions, allowing autonomous feature engineering, model selection, and optimization. This paper introduces the concept of DTIRS and establishes a roadmap for transitioning from Level-1 automation (dataset-agnostic but task-specific systems) to Level-2 automation (fully dataset- and task-independent systems). Achieving this paradigm would maximize code reusability and lower barriers to adoption. We discuss key challenges, including the trade-offs between generalization and specialization, computational overhead, and scalability, while presenting DsDL as a foundational tool for this vision. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:00:34 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wijaya",
"Tri Kurniawan",
""
],
[
"Shao",
"Xinyang",
""
],
[
"Pontiveros",
"Gonzalo Fiz",
""
],
[
"D'Amico",
"Edoardo",
""
]
] |
2506.03392 | Aref Ghoreishee | Aref Ghoreishee, Abhishek Mishra, John Walsh, Anup Das, Nagarajan Kandasamy | Improving Performance of Spike-based Deep Q-Learning using Ternary Neurons | null | null | null | null | cs.LG cs.NE cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | We propose a new ternary spiking neuron model to improve the representation capacity of binary spiking neurons in deep Q-learning. Although a ternary neuron model has recently been introduced to overcome the limited representation capacity offered by the binary spiking neurons, we show that its performance is worse than that of binary models in deep Q-learning tasks. We hypothesize gradient estimation bias during the training process as the underlying potential cause through mathematical and empirical analysis. We propose a novel ternary spiking neuron model to mitigate this issue by reducing the estimation bias. We use the proposed ternary spiking neuron as the fundamental computing unit in a deep spiking Q-learning network (DSQN) and evaluate the network's performance in seven Atari games from the Gym environment. Results show that the proposed ternary spiking neuron mitigates the drastic performance degradation of ternary neurons in Q-learning tasks and improves the network performance compared to the existing binary neurons, making DSQN a more practical solution for on-board autonomous decision-making tasks. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:06:13 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ghoreishee",
"Aref",
""
],
[
"Mishra",
"Abhishek",
""
],
[
"Walsh",
"John",
""
],
[
"Das",
"Anup",
""
],
[
"Kandasamy",
"Nagarajan",
""
]
] |
2506.03393 | Jack Wolf | Jack M. Wolf, Joseph S. Koopmeiners, David M. Vock | Jointly modeling multiple endpoints for efficient treatment effect estimation in randomized controlled trials | 17 pages, 6 figures | null | null | null | stat.ME | http://creativecommons.org/licenses/by-sa/4.0/ | Randomized controlled trials are the gold standard for evaluating the efficacy of an intervention. However, there is often a trade-off between selecting the most scientifically relevant primary endpoint versus a less relevant, but more powerful, endpoint. For example, in the context of tobacco regulatory science many trials evaluate cigarettes per day as the primary endpoint instead of abstinence from smoking due to limited power. Additionally, it is often of interest to consider subgroup analyses to answer additional questions; such analyses are rarely adequately powered. In practice, trials often collect multiple endpoints. Heuristically, if multiple endpoints demonstrate a similar treatment effect we would be more confident in the results of this trial. However, there is limited research on leveraging information from secondary endpoints besides using composite endpoints which can be difficult to interpret. In this paper, we develop an estimator for the treatment effect on the primary endpoint based on a joint model for primary and secondary efficacy endpoints. This estimator gains efficiency over the standard treatment effect estimator when the model is correctly specified but is robust to model misspecification via model averaging. We illustrate our approach by estimating the effect of very low nicotine content cigarettes on the proportion of Black people who smoke who achieve abstinence and find our approach reduces the standard error by 27%. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:06:13 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wolf",
"Jack M.",
""
],
[
"Koopmeiners",
"Joseph S.",
""
],
[
"Vock",
"David M.",
""
]
] |
2506.03394 | Shafqaat Ahmad Mr | Shafqaat Ahmad | Temporal Vegetation Index-Based Unsupervised Crop Stress Detection via Eigenvector-Guided Contrastive Learning | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Early detection of crop stress is vital for minimizing yield loss and enabling timely intervention in precision agriculture. Traditional approaches using NDRE often detect stress only after visible symptoms appear or require labeled datasets, limiting scalability. This study introduces EigenCL, a novel unsupervised contrastive learning framework guided by temporal NDRE dynamics and biologically grounded eigen decomposition. Using over 10,000 Sentinel-2 NDRE image patches from drought-affected Iowa cornfields, we constructed five-point NDRE time series per patch and derived an RBF similarity matrix. The principal eigenvector explaining 76% of the variance and strongly correlated (r = 0.95) with raw NDRE values was used to define stress-aware similarity for contrastive embedding learning. Unlike existing methods that rely on visual augmentations, EigenCL pulls embeddings together based on biologically similar stress trajectories and pushes apart divergent ones. The learned embeddings formed physiologically meaningful clusters, achieving superior clustering metrics (Silhouette: 0.748, DBI: 0.35) and enabling 76% early stress detection up to 12 days before conventional NDRE thresholds. Downstream classification yielded 95% k-NN and 91% logistic regression accuracy. Validation on an independent 2023 Nebraska dataset confirmed generalizability without retraining. EigenCL offers a label-free, scalable approach for early stress detection that aligns with underlying plant physiology and is suitable for real-world deployment in data-scarce agricultural environments. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:06:26 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ahmad",
"Shafqaat",
""
]
] |
2506.03395 | William Daniels | William S. Daniels, Douglas W. Nychka, Dorit M. Hammerling | A Bayesian hierarchical model for methane emission source apportionment | null | null | null | null | stat.AP | http://creativecommons.org/licenses/by/4.0/ | Reducing methane emissions from the oil and gas sector is a key component of short-term climate action. Emission reduction efforts are often conducted at the individual site-level, where being able to apportion emissions between a finite number of potentially emitting equipment is necessary for leak detection and repair as well as regulatory reporting of annualized emissions. We present a hierarchical Bayesian model, referred to as the multisource detection, localization, and quantification (MDLQ) model, for performing source apportionment on oil and gas sites using methane measurements from point sensor networks. The MDLQ model accounts for autocorrelation in the sensor data and enforces sparsity in the emission rate estimates via a spike-and-slab prior, as oil and gas equipment often emit intermittently. We use the MDLQ model to apportion methane emissions on an experimental oil and gas site designed to release methane in known quantities, providing a means of model evaluation. Data from this experiment are unique in their size (i.e., the number of controlled releases) and in their close approximation of emission characteristics on real oil and gas sites. As such, this study provides a baseline level of apportionment accuracy that can be expected when using point sensor networks on operational sites. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:07:14 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Daniels",
"William S.",
""
],
[
"Nychka",
"Douglas W.",
""
],
[
"Hammerling",
"Dorit M.",
""
]
] |
2506.03396 | Jinhan Kim | Jinhan Kim, Nargiz Humbatova, Gunel Jahangirova, Shin Yoo, Paolo Tonella | Fault Localisation and Repair for DL Systems: An Empirical Study with LLMs | arXiv admin note: text overlap with arXiv:2301.11568 | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | Numerous Fault Localisation (FL) and repair techniques have been proposed to address faults in Deep Learning (DL) models. However, their effectiveness in practical applications remains uncertain due to the reliance on pre-defined rules. This paper presents a comprehensive evaluation of state-of-the-art FL and repair techniques, examining their advantages and limitations. Moreover, we introduce a novel approach that harnesses the power of Large Language Models (LLMs) in localising and repairing DL faults. Our evaluation, conducted on a carefully designed benchmark, reveals the strengths and weaknesses of current FL and repair techniques. We emphasise the importance of enhanced accuracy and the need for more rigorous assessment methods that employ multiple ground truth patches. Notably, LLMs exhibit remarkable performance in both FL and repair tasks. For instance, the GPT-4 model achieves 44% and 82% improvements in FL and repair tasks respectively, compared to the second-best tool, demonstrating the potential of LLMs in this domain. Our study sheds light on the current state of FL and repair techniques and suggests that LLMs could be a promising avenue for future advancements. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:08:33 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Kim",
"Jinhan",
""
],
[
"Humbatova",
"Nargiz",
""
],
[
"Jahangirova",
"Gunel",
""
],
[
"Yoo",
"Shin",
""
],
[
"Tonella",
"Paolo",
""
]
] |
2506.03397 | Vahid Nourozi | Vahid Nourozi | Reinforcement Learning Enhanced Greedy Decoding for Quantum Stabilizer Codes over $\mathbb{F}_q$ | null | null | null | null | quant-ph math.QA | http://creativecommons.org/licenses/by/4.0/ | We construct new classical Goppa codes and corresponding quantum stabilizer codes from plane curves defined by separated polynomials. In particular, over $\mathbb{F}_3$ with the Hermitian curve $y^3 + y = x^4$, we obtain a ternary code of length 27, dimension 13, distance 4, which yields a [[27, 13, 4]]$_3$ quantum code. To decode, we introduce an RL-on-Greedy algorithm: first apply a standard greedy syndrome decoder, then use a trained Deep Q-Network to correct any residual syndrome. Simulation under a depolarizing noise model shows that RL-on-Greedy dramatically reduces logical failure compared to greedy alone. Our work thus broadens the class of Goppa- and quantum-stabilizer codes from separated-polynomial curves and delivers a learned decoder with near-optimal performance. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:08:36 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Nourozi",
"Vahid",
""
]
] |
2506.03398 | Giuseppe Bevilacqua | Alessandro Fregosi and Carmela Marinelli and Carlo Gabbanini and Valerio Biancalana and Maria Allegrini and Ennio Arimondo and Francesco Petiziol and Sandro Wimberger and Andrea Fioretti and Giuseppe Bevilacqua | Multipassage Landau-Zener tunneling oscillations in transverse/longitudinal dual dressing of atomic qubits | 15 pages, 6 figures | null | null | null | quant-ph physics.atom-ph | http://creativecommons.org/licenses/by/4.0/ | We investigate the time evolution of a non-resonant dressed-atom
qubit in an XZ original configuration. It is composed of two
electromagnetic fields, one oscillating parallel and the other
orthogonal to the quantisation magnetic static field. The
experiments are performed in rubidium and caesium atomic
magnetometers, confined in a magneto-optical trap and in a vapour
cell, respectively. Static fields in the $\mu$T range and kHz
oscillating fields with large Rabi frequencies are applied. This
dual-dressing configuration is an extension of the Landau-Zener
multipassage interferometry in the presence of an additional
dressing field controlling the tunneling process by its amplitude
and phase. Our measurement of the qubit coherence introduces
additional features to the transition probability readout of
standard interferometry. The coherence time evolution is
characterized by oscillations at several frequencies, each of them
produced by a different quantum contribution. Such frequency
description introduces a new picture of the qubit multipassage
evolution. Because the present low-frequency dressing operation
does not fall within the standard Floquet engineering paradigm based
on the high-frequency expansion, we develop an ad-hoc dressing
perturbation treatment. Numerical simulations support the adiabatic
and non-adiabatic qubit evolution. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:10:00 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Fregosi",
"Alessandro",
""
],
[
"Marinelli",
"Carmela",
""
],
[
"Gabbanini",
"Carlo",
""
],
[
"Biancalana",
"Valerio",
""
],
[
"Allegrini",
"Maria",
""
],
[
"Arimondo",
"Ennio",
""
],
[
"Petiziol",
"Francesco",
""
],
[
"Wimberger",
"Sandro",
""
],
[
"Fioretti",
"Andrea",
""
],
[
"Bevilacqua",
"Giuseppe",
""
]
] |
2506.03399 | Sean Steinle | Sean Steinle | Sampling Preferences Yields Simple Trustworthiness Scores | null | null | null | null | cs.HC cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the onset of large language models (LLMs), the performance of artificial intelligence (AI) models is becoming increasingly multi-dimensional. Accordingly, there have been several large, multi-dimensional evaluation frameworks put forward to evaluate LLMs. Though these frameworks are much more realistic than previous attempts which only used a single score like accuracy, multi-dimensional evaluations can complicate decision-making since there is no obvious way to select an optimal model. This work introduces preference sampling, a method to extract a scalar trustworthiness score from multi-dimensional evaluation results by considering the many characteristics of model performance which users value. We show that preference sampling improves upon alternate aggregation methods by using multi-dimensional trustworthiness evaluations of LLMs from TrustLLM and DecodingTrust. We find that preference sampling is consistently reductive, fully reducing the set of candidate models 100% of the time whereas Pareto optimality never reduces the set by more than 50%. Likewise, preference sampling is consistently sensitive to user priors-allowing users to specify the relative weighting and confidence of their preferences-whereas averaging scores is intransigent to the users' prior knowledge. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:14:35 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Steinle",
"Sean",
""
]
] |
2506.03400 | Collin Hague | Collin Hague, Artur Wolek | Occlusion-Aware Ground Target Tracking by a Dubins Vehicle Using Visibility Volumes | 56 pages, 22 figures, 1 table | null | null | null | eess.SY cs.RO cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper considers the problem of tracking a point of interest (POI) moving along a known trajectory on the ground with an uncrewed aerial vehicle (UAV) modeled as a Dubins vehicle using a line-of-sight (LOS) sensor through an urban environment that may occlude the POI. A visibility volume (VV) encodes a time-varying, three-dimensional representation of the sensing constraints for a particular POI position. A constant-altitude, translating, and radially time-varying circular standoff orbit is then inscribed within the dynamically changing VV centered at the POI position. The time-varying VV is approximated by placing static VVs along the POI's trajectory using an adaptive metric that restricts the volume change of consecutive visibility volumes to below a specified rate. The time-varying circular standoff orbit is proven to be feasible for a Dubins vehicle and is approximated with a piecewise set of linearly interpolated circular orbits inside the static VVs. A steering controller is derived that drives the UAV to converge to the time-varying standoff orbit. Numerical simulations and a flight test illustrate the proposed approach. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:16:53 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Hague",
"Collin",
""
],
[
"Wolek",
"Artur",
""
]
] |
2506.03401 | Xiwei Xu | Xiwei Xu and Hans Weytjens and Dawen Zhang and Qinghua Lu and Ingo Weber and Liming Zhu | RAGOps: Operating and Managing Retrieval-Augmented Generation Pipelines | null | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent studies show that 60% of LLM-based compound systems in enterprise environments leverage some form of retrieval-augmented generation (RAG), which enhances the relevance and accuracy of LLM (or other genAI) outputs by retrieving relevant information from external data sources. LLMOps involves the practices and techniques for managing the lifecycle and operations of LLM compound systems in production environments. It supports enhancing LLM systems through continuous operations and feedback evaluation. RAGOps extends LLMOps by incorporating a strong focus on data management to address the continuous changes in external data sources. This necessitates automated methods for evaluating and testing data operations, enhancing retrieval relevance and generation quality. In this paper, we (1) characterize the generic architecture of RAG applications based on the 4+1 model view for describing software architectures, (2) outline the lifecycle of RAG systems, which integrates the management lifecycles of both the LLM and the data, (3) define the key design considerations of RAGOps across different stages of the RAG lifecycle and quality trade-off analyses, (4) highlight the overarching research challenges around RAGOps, and (5) present two use cases of RAG applications and the corresponding RAGOps considerations. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:22:23 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xu",
"Xiwei",
""
],
[
"Weytjens",
"Hans",
""
],
[
"Zhang",
"Dawen",
""
],
[
"Lu",
"Qinghua",
""
],
[
"Weber",
"Ingo",
""
],
[
"Zhu",
"Liming",
""
]
] |
2506.03402 | Xiaotian Su | Xiaotian Su, April Wang | The Stress of Improvisation: Instructors' Perspectives on Live Coding in Programming Classes | 6 pages | In *Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems* (CHI EA '25), Association for Computing Machinery, New York, NY, USA, 2025 | 10.1145/3706599.3719993 | Article 525 | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Live coding is a pedagogical technique in which an instructor writes and executes code in front of students to impart skills like incremental development and debugging. Although live coding offers many benefits, instructors face many challenges in the classroom, like cognitive challenges and psychological stress, most of which have yet to be formally studied. To understand the obstacles faced by instructors in CS classes, we conducted (1) a formative interview with five teaching assistants in exercise sessions and (2) a contextual inquiry study with four lecturers for large-scale classes. We found that the improvisational and unpredictable nature of live coding makes it difficult for instructors to manage their time and keep students engaged, resulting in more mental stress than presenting static slides. We discussed opportunities for augmenting existing IDEs and presentation setups to help enhance live coding experience. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:23:14 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Su",
"Xiaotian",
""
],
[
"Wang",
"April",
""
]
] |
2506.03403 | Orchid Chetia Phukan | Orchid Chetia Phukan, Girish, Mohd Mujtaba Akhtar, Swarup Ranjan Behera, Pailla Balakrishna Reddy, Arun Balaji Buduru, Rajesh Sharma | HYFuse: Aligning Heterogeneous Speech Pre-Trained Representations in Hyperbolic Space for Speech Emotion Recognition | Accepted to INTERSPEECH 2025 | null | null | null | eess.AS | http://creativecommons.org/licenses/by/4.0/ | Compression-based representations (CBRs) from neural audio codecs such as EnCodec capture intricate acoustic features like pitch and timbre, while representation-learning-based representations (RLRs) from pre-trained models trained for speech representation learning such as WavLM encode high-level semantic and prosodic information. Previous research on Speech Emotion Recognition (SER) has explored both, however, fusion of CBRs and RLRs haven't been explored yet. In this study, we solve this gap and investigate the fusion of RLRs and CBRs and hypothesize they will be more effective by providing complementary information. To this end, we propose, HYFuse, a novel framework that fuses the representations by transforming them to hyperbolic space. With HYFuse, through fusion of x-vector (RLR) and Soundstream (CBR), we achieve the top performance in comparison to individual representations as well as the homogeneous fusion of RLRs and CBRs and report SOTA. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:26:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Phukan",
"Orchid Chetia",
""
],
[
"Girish",
"",
""
],
[
"Akhtar",
"Mohd Mujtaba",
""
],
[
"Behera",
"Swarup Ranjan",
""
],
[
"Reddy",
"Pailla Balakrishna",
""
],
[
"Buduru",
"Arun Balaji",
""
],
[
"Sharma",
"Rajesh",
""
]
] |
2506.03404 | Johan Obando-Ceron | Walter Mayor, Johan Obando-Ceron, Aaron Courville, Pablo Samuel Castro | The Impact of On-Policy Parallelized Data Collection on Deep Reinforcement Learning Networks | Proceedings of the 42nd International Conference on Machine Learning (ICML 2025) | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | The use of parallel actors for data collection has been an effective technique used in reinforcement learning (RL) algorithms. The manner in which data is collected in these algorithms, controlled via the number of parallel environments and the rollout length, induces a form of bias-variance trade-off; the number of training passes over the collected data, on the other hand, must strike a balance between sample efficiency and overfitting. We conduct an empirical analysis of these trade-offs on PPO, one of the most popular RL algorithms that uses parallel actors, and establish connections to network plasticity and, more generally, optimization stability. We examine its impact on network architectures, as well as the hyper-parameter sensitivity when scaling data. Our analyses indicate that larger dataset sizes can increase final performance across a variety of settings, and that scaling parallel environments is more effective than increasing rollout lengths. These findings highlight the critical role of data collection strategies in improving agent performance. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:27:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mayor",
"Walter",
""
],
[
"Obando-Ceron",
"Johan",
""
],
[
"Courville",
"Aaron",
""
],
[
"Castro",
"Pablo Samuel",
""
]
] |
2506.03406 | Runqiu Wang | Runqiu Wang, Ran Dai | Online multi-layer FDR control | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When hypotheses are tested in a stream and real-time decision-making is needed, online sequential hypothesis testing procedures are needed. Furthermore, these hypotheses are commonly partitioned into groups by their nature. For example, the RNA nanocapsules can be partitioned based on therapeutic nucleic acids (siRNAs) being used, as well as the delivery nanocapsules. When selecting effective RNA nanocapsules, simultaneous false discovery rate control at multiple partition levels is needed. In this paper, we develop hypothesis testing procedures which controls false discovery rate (FDR) simultaneously for multiple partitions of hypotheses in an online fashion. We provide rigorous proofs on their FDR or modified FDR (mFDR) control properties and use extensive simulations to demonstrate their performance. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:29:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"Runqiu",
""
],
[
"Dai",
"Ran",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.