title
stringlengths 7
246
| abstract
stringlengths 3
3.31k
|
|---|---|
Towards Productizing AI/ML Models: An Industry Perspective from Data Scientists | The transition from AI/ML models to production-ready AI-based systems is a challenge for both data scientists and software engineers. In this paper, we report the results of a workshop conducted in a consulting company to understand how this transition is perceived by practitioners. Starting from the need for making AI experiments reproducible, the main themes that emerged are related to the use of the Jupyter Notebook as the primary prototyping tool, and the lack of support for software engineering best practices as well as data science specific functionalities. |
TextMage: The Automated Bangla Caption Generator Based On Deep Learning | Neural Networks and Deep Learning have seen an upsurge of research in the past decade due to the improved results. Generates text from the given image is a crucial task that requires the combination of both sectors which are computer vision and natural language processing in order to understand an image and represent it using a natural language. However existing works have all been done on a particular lingual domain and on the same set of data. This leads to the systems being developed to perform poorly on images that belong to specific locales' geographical context. TextMage is a system that is capable of understanding visual scenes that belong to the Bangladeshi geographical context and use its knowledge to represent what it understands in Bengali. Hence, we have trained a model on our previously developed and published dataset named BanglaLekhaImageCaptions. This dataset contains 9,154 images along with two annotations for each image. In order to access performance, the proposed model has been implemented and evaluated. |
Using Deep Learning and Explainable Artificial Intelligence in Patients' Choices of Hospital Levels | In countries that enabled patients to choose their own providers, a common problem is that the patients did not make rational decisions, and hence, fail to use healthcare resources efficiently. This might cause problems such as overwhelming tertiary facilities with mild condition patients, thus limiting their capacity of treating acute and critical patients. To address such maldistributed patient volume, it is essential to oversee patients choices before further evaluation of a policy or resource allocation. This study used nationwide insurance data, accumulated possible features discussed in existing literature, and used a deep neural network to predict the patients choices of hospital levels. This study also used explainable artificial intelligence methods to interpret the contribution of features for the general public and individuals. In addition, we explored the effectiveness of changing data representations. The results showed that the model was able to predict with high area under the receiver operating characteristics curve (AUC) (0.90), accuracy (0.90), sensitivity (0.94), and specificity (0.97) with highly imbalanced label. Generally, social approval of the provider by the general public (positive or negative) and the number of practicing physicians serving per ten thousand people of the located area are listed as the top effecting features. The changing data representation had a positive effect on the prediction improvement. Deep learning methods can process highly imbalanced data and achieve high accuracy. The effecting features affect the general public and individuals differently. Addressing the sparsity and discrete nature of insurance data leads to better prediction. Applications using deep learning technology are promising in health policy making. More work is required to interpret models and practice implementation. |
Semantic Relatedness and Taxonomic Word Embeddings | This paper connects a series of papers dealing with taxonomic word embeddings. It begins by noting that there are different types of semantic relatedness and that different lexical representations encode different forms of relatedness. A particularly important distinction within semantic relatedness is that of thematic versus taxonomic relatedness. Next, we present a number of experiments that analyse taxonomic embeddings that have been trained on a synthetic corpus that has been generated via a random walk over a taxonomy. These experiments demonstrate how the properties of the synthetic corpus, such as the percentage of rare words, are affected by the shape of the knowledge graph the corpus is generated from. Finally, we explore the interactions between the relative sizes of natural and synthetic corpora on the performance of embeddings when taxonomic and thematic embeddings are combined. |
Device-Free User Authentication, Activity Classification and Tracking using Passive Wi-Fi Sensing: A Deep Learning Based Approach | Privacy issues related to video camera feeds have led to a growing need for suitable alternatives that provide functionalities such as user authentication, activity classification and tracking in a noninvasive manner. Existing infrastructure makes Wi-Fi a possible candidate, yet, utilizing traditional signal processing methods to extract information necessary to fully characterize an event by sensing weak ambient Wi-Fi signals is deemed to be challenging. This paper introduces a novel end to-end deep learning framework that simultaneously predicts the identity, activity and the location of a user to create user profiles similar to the information provided through a video camera. The system is fully autonomous and requires zero user intervention unlike systems that require user-initiated initialization, or a user held transmitting device to facilitate the prediction. The system can also predict the trajectory of the user by predicting the location of a user over consecutive time steps. The performance of the system is evaluated through experiments. |
A Study into patient similarity through representation learning from medical records | Patient similarity assessment, which identifies patients similar to a given patient, can help improve medical care. The assessment can be performed using Electronic Medical Records (EMRs). Patient similarity measurement requires converting heterogeneous EMRs into comparable formats to calculate their distance. While versatile document representation learning methods have been developed in recent years, it is still unclear how complex EMR data should be processed to create the most useful patient representations. This study presents a new data representation method for EMRs that takes the information in clinical narratives into account. To address the limitations of previous approaches in handling complex parts of EMR data, an unsupervised method is proposed for building a patient representation, which integrates unstructured data with structured data extracted from patients' EMRs. In order to model the extracted data, we employed a tree structure that captures the temporal relations of multiple medical events from EMR. We processed clinical notes to extract symptoms, signs, and diseases using different tools such as medspaCy, MetaMap, and scispaCy and mapped entities to the Unified Medical Language System (UMLS). After creating a tree data structure, we utilized two novel relabeling methods for the non-leaf nodes of the tree to capture two temporal aspects of the extracted events. By traversing the tree, we generated a sequence that could create an embedding vector for each patient. The comprehensive evaluation of the proposed method for patient similarity and mortality prediction tasks demonstrated that our proposed model leads to lower mean squared error (MSE), higher precision, and normalized discounted cumulative gain (NDCG) relative to baselines. |
Intra-Instance VICReg: Bag of Self-Supervised Image Patch Embedding | Recently, self-supervised learning (SSL) has achieved tremendous empirical advancements in learning image representation. However, our understanding and knowledge of the representation are still limited. This work shows that the success of the SOTA siamese-network-based SSL approaches is primarily based on learning a representation of image patches. Particularly, we show that when we learn a representation only for fixed-scale image patches and aggregate different patch representations linearly for an image (instance), it can achieve on par or even better results than the baseline methods on several benchmarks. Further, we show that the patch representation aggregation can also improve various SOTA baseline methods by a large margin. We also establish a formal connection between the SSL objective and the image patches co-occurrence statistics modeling, which supplements the prevailing invariance perspective. By visualizing the nearest neighbors of different image patches in the embedding space and projection space, we show that while the projection has more invariance, the embedding space tends to preserve more equivariance and locality. Finally, we propose a hypothesis for the future direction based on the discovery of this work. |
Is AmI (Attacks Meet Interpretability) Robust to Adversarial Examples? | No. |
Balanced background and explanation data are needed in explaining deep learning models with SHAP: An empirical study on clinical decision making | Objective: Shapley additive explanations (SHAP) is a popular post-hoc technique for explaining black box models. While the impact of data imbalance on predictive models has been extensively studied, it remains largely unknown with respect to SHAP-based model explanations. This study sought to investigate the effects of data imbalance on SHAP explanations for deep learning models, and to propose a strategy to mitigate these effects. Materials and Methods: We propose to adjust class distributions in the background and explanation data in SHAP when explaining black box models. Our data balancing strategy is to compose background data and explanation data with an equal distribution of classes. To evaluate the effects of data adjustment on model explanation, we propose to use the beeswarm plot as a qualitative tool to identify "abnormal" explanation artifacts, and quantitatively test the consistency between variable importance and prediction power. We demonstrated our proposed approach in an empirical study that predicted inpatient mortality using the Medical Information Mart for Intensive Care (MIMIC-III) data and a multilayer perceptron. Results: Using the data balancing strategy would allow us to reduce the number of the artifacts in the beeswarm plot, thus mitigating the negative effects of data imbalance. Additionally, with the balancing strategy, the top-ranked variables from the corresponding importance ranking demonstrated improved discrimination power. Discussion and Conclusion: Our findings suggest that balanced background and explanation data could help reduce the noise in explanation results induced by skewed data distribution and improve the reliability of variable importance ranking. Furthermore, these balancing procedures improve the potential of SHAP in identifying patients with abnormal characteristics in clinical applications. |
A Framework of Learning Through Empirical Gain Maximization | We develop in this paper a framework of empirical gain maximization (EGM) to address the robust regression problem where heavy-tailed noise or outliers may present in the response variable. The idea of EGM is to approximate the density function of the noise distribution instead of approximating the truth function directly as usual. Unlike the classical maximum likelihood estimation that encourages equal importance of all observations and could be problematic in the presence of abnormal observations, EGM schemes can be interpreted from a minimum distance estimation viewpoint and allow the ignorance of those observations. Furthermore, it is shown that several well-known robust nonconvex regression paradigms, such as Tukey regression and truncated least square regression, can be reformulated into this new framework. We then develop a learning theory for EGM, by means of which a unified analysis can be conducted for these well-established but not fully-understood regression approaches. Resulting from the new framework, a novel interpretation of existing bounded nonconvex loss functions can be concluded. Within this new framework, the two seemingly irrelevant terminologies, the well-known Tukey's biweight loss for robust regression and the triweight kernel for nonparametric smoothing, are closely related. More precisely, it is shown that the Tukey's biweight loss can be derived from the triweight kernel. Similarly, other frequently employed bounded nonconvex loss functions in machine learning such as the truncated square loss, the Geman-McClure loss, and the exponential squared loss can also be reformulated from certain smoothing kernels in statistics. In addition, the new framework enables us to devise new bounded nonconvex loss functions for robust learning. |
Computationally Efficient Multiscale Neural Networks Applied To Fluid Flow In Complex 3D Porous Media | The permeability of complex porous materials can be obtained via direct flow simulation, which provides the most accurate results, but is very computationally expensive. In particular, the simulation convergence time scales poorly as simulation domains become tighter or more heterogeneous. Semi-analytical models that rely on averaged structural properties (i.e. porosity and tortuosity) have been proposed, but these features only summarize the domain, resulting in limited applicability. On the other hand, data-driven machine learning approaches have shown great promise for building more general models by virtue of accounting for the spatial arrangement of the domains solid boundaries. However, prior approaches building on the Convolutional Neural Network (ConvNet) literature concerning 2D image recognition problems do not scale well to the large 3D domains required to obtain a Representative Elementary Volume (REV). As such, most prior work focused on homogeneous samples, where a small REV entails that that the global nature of fluid flow could be mostly neglected, and accordingly, the memory bottleneck of addressing 3D domains with ConvNets was side-stepped. Therefore, important geometries such as fractures and vuggy domains could not be well-modeled. In this work, we address this limitation with a general multiscale deep learning model that is able to learn from porous media simulation data. By using a coupled set of neural networks that view the domain on different scales, we enable the evaluation of large images in approximately one second on a single Graphics Processing Unit. This model architecture opens up the possibility of modeling domain sizes that would not be feasible using traditional direct simulation tools on a desktop computer. |
Behaviour-Diverse Automatic Penetration Testing: A Curiosity-Driven Multi-Objective Deep Reinforcement Learning Approach | Penetration Testing plays a critical role in evaluating the security of a target network by emulating real active adversaries. Deep Reinforcement Learning (RL) is seen as a promising solution to automating the process of penetration tests by reducing human effort and improving reliability. Existing RL solutions focus on finding a specific attack path to impact the target hosts. However, in reality, a diverse range of attack variations are needed to provide comprehensive assessments of the target network's security level. Hence, the attack agents must consider multiple objectives when penetrating the network. Nevertheless, this challenge is not adequately addressed in the existing literature. To this end, we formulate the automatic penetration testing in the Multi-Objective Reinforcement Learning (MORL) framework and propose a Chebyshev decomposition critic to find diverse adversary strategies that balance different objectives in the penetration test. Additionally, the number of available actions increases with the agent consistently probing the target network, making the training process intractable in many practical situations. Thus, we introduce a coverage-based masking mechanism that reduces attention on previously selected actions to help the agent adapt to future exploration. Experimental evaluation on a range of scenarios demonstrates the superiority of our proposed approach when compared to adapted algorithms in terms of multi-objective learning and performance efficiency. |
Consensus-Based Optimization on the Sphere: Convergence to Global Minimizers and Machine Learning | We investigate the implementation of a new stochastic Kuramoto-Vicsek-type model for global optimization of nonconvex functions on the sphere. This model belongs to the class of Consensus-Based Optimization. In fact, particles move on the sphere driven by a drift towards an instantaneous consensus point, which is computed as a convex combination of particle locations, weighted by the cost function according to Laplace's principle, and it represents an approximation to a global minimizer. The dynamics is further perturbed by a random vector field to favor exploration, whose variance is a function of the distance of the particles to the consensus point. In particular, as soon as the consensus is reached the stochastic component vanishes. The main results of this paper are about the proof of convergence of the numerical scheme to global minimizers provided conditions of well-preparation of the initial datum. The proof combines previous results of mean-field limit with a novel asymptotic analysis, and classical convergence results of numerical methods for SDE. We present several numerical experiments, which show that the algorithm proposed in the present paper scales well with the dimension and is extremely versatile. To quantify the performances of the new approach, we show that the algorithm is able to perform essentially as good as ad hoc state of the art methods in challenging problems in signal processing and machine learning, namely the phase retrieval problem and the robust subspace detection. |
Rethinking floating point for deep learning | Reducing hardware overhead of neural networks for faster or lower power inference and training is an active area of research. Uniform quantization using integer multiply-add has been thoroughly investigated, which requires learning many quantization parameters, fine-tuning training or other prerequisites. Little effort is made to improve floating point relative to this baseline; it remains energy inefficient, and word size reduction yields drastic loss in needed dynamic range. We improve floating point to be more energy efficient than equivalent bit width integer hardware on a 28 nm ASIC process while retaining accuracy in 8 bits with a novel hybrid log multiply/linear add, Kulisch accumulation and tapered encodings from Gustafson's posit format. With no network retraining, and drop-in replacement of all math and float32 parameters via round-to-nearest-even only, this open-sourced 8-bit log float is within 0.9% top-1 and 0.2% top-5 accuracy of the original float32 ResNet-50 CNN model on ImageNet. Unlike int8 quantization, it is still a general purpose floating point arithmetic, interpretable out-of-the-box. Our 8/38-bit log float multiply-add is synthesized and power profiled at 28 nm at 0.96x the power and 1.12x the area of 8/32-bit integer multiply-add. In 16 bits, our log float multiply-add is 0.59x the power and 0.68x the area of IEEE 754 float16 fused multiply-add, maintaining the same signficand precision and dynamic range, proving useful for training ASICs as well. |
Inverse Rational Control: Inferring What You Think from How You Forage | Complex behaviors are often driven by an internal model, which integrates sensory information over time and facilitates long-term planning. Inferring an agent's internal model is a crucial ingredient in social interactions (theory of mind), for imitation learning, and for interpreting neural activities of behaving agents. Here we describe a generic method to model an agent's behavior under an environment with uncertainty, and infer the agent's internal model, reward function, and dynamic beliefs. We apply our method to a simulated agent performing a naturalistic foraging task. We assume the agent behaves rationally --- that is, they take actions that optimize their subjective utility according to their understanding of the task and its relevant causal variables. We model this rational solution as a Partially Observable Markov Decision Process (POMDP) where the agent may make wrong assumptions about the task parameters. Given the agent's sensory observations and actions, we learn its internal model and reward function by maximum likelihood estimation over a set of task-relevant parameters. The Markov property of the POMDP enables us to characterize the transition probabilities between internal belief states and iteratively estimate the agent's policy using a constrained Expectation-Maximization (EM) algorithm. We validate our method on simulated agents performing suboptimally on a foraging task currently used in many neuroscience experiments, and successfully recover their internal model and reward function. Our work lays a critical foundation to discover how the brain represents and computes with dynamic beliefs. |
Potential Field Guided Actor-Critic Reinforcement Learning | In this paper, we consider the problem of actor-critic reinforcement learning. Firstly, we extend the actor-critic architecture to actor-critic-N architecture by introducing more critics beyond rewards. Secondly, we combine the reward-based critic with a potential-field-based critic to formulate the proposed potential field guided actor-critic reinforcement learning approach (actor-critic-2). This can be seen as a combination of the model-based gradients and the model-free gradients in policy improvement. State with large potential field often contains a strong prior information, such as pointing to the target at a long distance or avoiding collision by the side of an obstacle. In this situation, we should trust potential-field-based critic more as policy evaluation to accelerate policy improvement, where action policy tends to be guided. For example, in practical application, learning to avoid obstacles should be guided rather than learned by trial and error. State with small potential filed is often lack of information, for example, at the local minimum point or around the moving target. At this time, we should trust reward-based critic as policy evaluation more to evaluate the long-term return. In this case, action policy tends to explore. In addition, potential field evaluation can be combined with planning to estimate a better state value function. In this way, reward design can focus more on the final stage of reward, rather than reward shaping or phased reward. Furthermore, potential field evaluation can make up for the lack of communication in multi-agent cooperation problem, i.e., multi-agent each has a reward-based critic and a relative unified potential-field-based critic with prior information. Thirdly, simplified experiments on predator-prey game demonstrate the effectiveness of the proposed approach. |
Heterogeneous Metric Learning with Content-based Regularization for Software Artifact Retrieval | The problem of software artifact retrieval has the goal to effectively locate software artifacts, such as a piece of source code, in a large code repository. This problem has been traditionally addressed through the textual query. In other words, information retrieval techniques will be exploited based on the textual similarity between queries and textual representation of software artifacts, which is generated by collecting words from comments, identifiers, and descriptions of programs. However, in addition to these semantic information, there are rich information embedded in source codes themselves. These source codes, if analyzed properly, can be a rich source for enhancing the efforts of software artifact retrieval. To this end, in this paper, we develop a feature extraction method on source codes. Specifically, this method can capture both the inherent information in the source codes and the semantic information hidden in the comments, descriptions, and identifiers of the source codes. Moreover, we design a heterogeneous metric learning approach, which allows to integrate code features and text features into the same latent semantic space. This, in turn, can help to measure the artifact similarity by exploiting the joint power of both code and text features. Finally, extensive experiments on real-world data show that the proposed method can help to improve the performances of software artifact retrieval with a significant margin. |
Magnetic Manifold Hamiltonian Monte Carlo | Markov chain Monte Carlo (MCMC) algorithms offer various strategies for sampling; the Hamiltonian Monte Carlo (HMC) family of samplers are MCMC algorithms which often exhibit improved mixing properties. The recently introduced magnetic HMC, a generalization of HMC motivated by the physics of particles influenced by magnetic field forces, has been demonstrated to improve the performance of HMC. In many applications, one wishes to sample from a distribution restricted to a constrained set, often manifested as an embedded manifold (for example, the surface of a sphere). We introduce magnetic manifold HMC, an HMC algorithm on embedded manifolds motivated by the physics of particles constrained to a manifold and moving under magnetic field forces. We discuss the theoretical properties of magnetic Hamiltonian dynamics on manifolds, and introduce a reversible and symplectic integrator for the HMC updates. We demonstrate that magnetic manifold HMC produces favorable sampling behaviors relative to the canonical variant of manifold-constrained HMC. |
Guidance on the Assurance of Machine Learning in Autonomous Systems (AMLAS) | Machine Learning (ML) is now used in a range of systems with results that are reported to exceed, under certain conditions, human performance. Many of these systems, in domains such as healthcare , automotive and manufacturing, exhibit high degrees of autonomy and are safety critical. Establishing justified confidence in ML forms a core part of the safety case for these systems. In this document we introduce a methodology for the Assurance of Machine Learning for use in Autonomous Systems (AMLAS). AMLAS comprises a set of safety case patterns and a process for (1) systematically integrating safety assurance into the development of ML components and (2) for generating the evidence base for explicitly justifying the acceptable safety of these components when integrated into autonomous system applications. |
Move Evaluation in Go Using Deep Convolutional Neural Networks | The game of Go is more challenging than other board games, due to the difficulty of constructing a position or move evaluation function. In this paper we investigate whether deep convolutional networks can be used to directly represent and learn this knowledge. We train a large 12-layer convolutional neural network by supervised learning from a database of human professional games. The network correctly predicts the expert move in 55% of positions, equalling the accuracy of a 6 dan human player. When the trained convolutional network was used directly to play games of Go, without any search, it beat the traditional search program GnuGo in 97% of games, and matched the performance of a state-of-the-art Monte-Carlo tree search that simulates a million positions per move. |
Ballooning Multi-Armed Bandits | In this paper, we introduce Ballooning Multi-Armed Bandits (BL-MAB), a novel extension of the classical stochastic MAB model. In the BL-MAB model, the set of available arms grows (or balloons) over time. In contrast to the classical MAB setting where the regret is computed with respect to the best arm overall, the regret in a BL-MAB setting is computed with respect to the best available arm at each time. We first observe that the existing stochastic MAB algorithms result in linear regret for the BL-MAB model. We prove that, if the best arm is equally likely to arrive at any time instant, a sub-linear regret cannot be achieved. Next, we show that if the best arm is more likely to arrive in the early rounds, one can achieve sub-linear regret. Our proposed algorithm determines (1) the fraction of the time horizon for which the newly arriving arms should be explored and (2) the sequence of arm pulls in the exploitation phase from among the explored arms. Making reasonable assumptions on the arrival distribution of the best arm in terms of the thinness of the distribution's tail, we prove that the proposed algorithm achieves sub-linear instance-independent regret. We further quantify explicit dependence of regret on the arrival distribution parameters. We reinforce our theoretical findings with extensive simulation results. We conclude by showing that our algorithm would achieve sub-linear regret even if (a) the distributional parameters are not exactly known, but are obtained using a reasonable learning mechanism or (b) the best arm is not more likely to arrive early, but a large fraction of arms is likely to arrive relatively early. |
Deep Neural Networks for COVID-19 Detection and Diagnosis using Images and Acoustic-based Techniques: A Recent Review | The new coronavirus disease (COVID-19) has been declared a pandemic since March 2020 by the World Health Organization. It consists of an emerging viral infection with respiratory tropism that could develop atypical pneumonia. Experts emphasize the importance of early detection of those who have the COVID-19 virus. In this way, patients will be isolated from other people and the spread of the virus can be prevented. For this reason, it has become an area of interest to develop early diagnosis and detection methods to ensure a rapid treatment process and prevent the virus from spreading. Since the standard testing system is time-consuming and not available for everyone, alternative early-screening techniques have become an urgent need. In this study, the approaches used in the detection of COVID-19 based on deep learning (DL) algorithms, which have been popular in recent years, have been comprehensively discussed. The advantages and disadvantages of different approaches used in literature are examined in detail. The Computed Tomography of the chest and X-ray images give a rich representation of the patient's lung that is less time-consuming and allows an efficient viral pneumonia detection using the DL algorithms. The first step is the pre-processing of these images to remove noise. Next, deep features are extracted using multiple types of deep models (pre-trained models, generative models, generic neural networks, etc.). Finally, the classification is performed using the obtained features to decide whether the patient is infected by coronavirus or it is another lung disease. In this study, we also give a brief review of the latest applications of cough analysis to early screen the COVID-19, and human mobility estimation to limit its spread. |
Multi-channel neural networks for predicting influenza A virus hosts and antigenic types | Influenza occurs every season and occasionally causes pandemics. Despite its low mortality rate, influenza is a major public health concern, as it can be complicated by severe diseases like pneumonia. A fast, accurate and low-cost method to predict the origin host and subtype of influenza viruses could help reduce virus transmission and benefit resource-poor areas. In this work, we propose multi-channel neural networks to predict antigenic types and hosts of influenza A viruses with hemagglutinin and neuraminidase protein sequences. An integrated data set containing complete protein sequences were used to produce a pre-trained model, and two other data sets were used for testing the model's performance. One test set contained complete protein sequences, and another test set contained incomplete protein sequences. The results suggest that multi-channel neural networks are applicable and promising for predicting influenza A virus hosts and antigenic subtypes with complete and partial protein sequences. |
ClosNets: a Priori Sparse Topologies for Faster DNN Training | Fully-connected layers in deep neural networks (DNN) are often the throughput and power bottleneck during training. This is due to their large size and low data reuse. Pruning dense layers can significantly reduce the size of these networks, but this approach can only be applied after training. In this work we propose a novel fully-connected layer that reduces the memory requirements of DNNs without sacrificing accuracy. We replace a dense matrix with products of sparse matrices whose topologies we pick in advance. This allows us to: (1) train significantly smaller networks without a loss in accuracy, and (2) store the network weights without having to store connection indices. We therefore achieve significant training speedups due to the smaller network size, and a reduced amount of computation per epoch. We tested several sparse layer topologies and found that Clos networks perform well due to their high path diversity, shallowness, and high model accuracy. With the ClosNets, we are able to reduce dense layer sizes by as much as an order of magnitude without hurting model accuracy. |
A Decentralized Policy Gradient Approach to Multi-task Reinforcement Learning | We develop a mathematical framework for solving multi-task reinforcement learning (MTRL) problems based on a type of policy gradient method. The goal in MTRL is to learn a common policy that operates effectively in different environments; these environments have similar (or overlapping) state spaces, but have different rewards and dynamics. We highlight two fundamental challenges in MTRL that are not present in its single task counterpart, and illustrate them with simple examples. We then develop a decentralized entropy-regularized policy gradient method for solving the MTRL problem, and study its finite-time convergence rate. We demonstrate the effectiveness of the proposed method using a series of numerical experiments. These experiments range from small-scale "GridWorld" problems that readily demonstrate the trade-offs involved in multi-task learning to large-scale problems, where common policies are learned to navigate an airborne drone in multiple (simulated) environments. |
KNAS: Green Neural Architecture Search | Many existing neural architecture search (NAS) solutions rely on downstream training for architecture evaluation, which takes enormous computations. Considering that these computations bring a large carbon footprint, this paper aims to explore a green (namely environmental-friendly) NAS solution that evaluates architectures without training. Intuitively, gradients, induced by the architecture itself, directly decide the convergence and generalization results. It motivates us to propose the gradient kernel hypothesis: Gradients can be used as a coarse-grained proxy of downstream training to evaluate random-initialized networks. To support the hypothesis, we conduct a theoretical analysis and find a practical gradient kernel that has good correlations with training loss and validation performance. According to this hypothesis, we propose a new kernel based architecture search approach KNAS. Experiments show that KNAS achieves competitive results with orders of magnitude faster than "train-then-test" paradigms on image classification tasks. Furthermore, the extremely low search cost enables its wide applications. The searched network also outperforms strong baseline RoBERTA-large on two text classification tasks. Codes are available at \url{https://github.com/Jingjing-NLP/KNAS} . |
Subgoal Discovery for Hierarchical Dialogue Policy Learning | Developing agents to engage in complex goal-oriented dialogues is challenging partly because the main learning signals are very sparse in long conversations. In this paper, we propose a divide-and-conquer approach that discovers and exploits the hidden structure of the task to enable efficient policy learning. First, given successful example dialogues, we propose the Subgoal Discovery Network (SDN) to divide a complex goal-oriented task into a set of simpler subgoals in an unsupervised fashion. We then use these subgoals to learn a multi-level policy by hierarchical reinforcement learning. We demonstrate our method by building a dialogue agent for the composite task of travel planning. Experiments with simulated and real users show that our approach performs competitively against a state-of-the-art method that requires human-defined subgoals. Moreover, we show that the learned subgoals are often human comprehensible. |
RLScheduler: An Automated HPC Batch Job Scheduler Using Reinforcement Learning | Today high-performance computing (HPC) platforms are still dominated by batch jobs. Accordingly, effective batch job scheduling is crucial to obtain high system efficiency. Existing HPC batch job schedulers typically leverage heuristic priority functions to prioritize and schedule jobs. But, once configured and deployed by the experts, such priority functions can hardly adapt to the changes of job loads, optimization goals, or system settings, potentially leading to degraded system efficiency when changes occur. To address this fundamental issue, we present RLScheduler, an automated HPC batch job scheduler built on reinforcement learning. RLScheduler relies on minimal manual interventions or expert knowledge, but can learn high-quality scheduling policies via its own continuous 'trial and error'. We introduce a new kernel-based neural network structure and trajectory filtering mechanism in RLScheduler to improve and stabilize the learning process. Through extensive evaluations, we confirm that RLScheduler can learn high-quality scheduling policies towards various workloads and various optimization goals with relatively low computation cost. Moreover, we show that the learned models perform stably even when applied to unseen workloads, making them practical for production use. |
Unsupervised Learning of Unbiased Visual Representations | Deep neural networks are known for their inability to learn robust representations when biases exist in the dataset. This results in a poor generalization to unbiased datasets, as the predictions strongly rely on peripheral and confounding factors, which are erroneously learned by the network. Many existing works deal with this issue by either employing an explicit supervision on the bias attributes, or assuming prior knowledge about the bias. In this work we study this problem in a more difficult scenario, in which no explicit annotation about the bias is available, and without any prior knowledge about its nature. We propose a fully unsupervised debiasing framework, consisting of three steps: first, we exploit the natural preference for learning malignant biases, obtaining a bias-capturing model; then, we perform a pseudo-labelling step to obtain bias labels; finally we employ state-of-the-art supervised debiasing techniques to obtain an unbiased model. We also propose a theoretical framework to assess the biasness of a model, and provide a detailed analysis on how biases affect the training of neural networks. We perform experiments on synthetic and real-world datasets, showing that our method achieves state-of-the-art performance in a variety of settings, sometimes even higher than fully supervised debiasing approaches. |
Deep Marching Tetrahedra: a Hybrid Representation for High-Resolution 3D Shape Synthesis | We introduce DMTet, a deep 3D conditional generative model that can synthesize high-resolution 3D shapes using simple user guides such as coarse voxels. It marries the merits of implicit and explicit 3D representations by leveraging a novel hybrid 3D representation. Compared to the current implicit approaches, which are trained to regress the signed distance values, DMTet directly optimizes for the reconstructed surface, which enables us to synthesize finer geometric details with fewer artifacts. Unlike deep 3D generative models that directly generate explicit representations such as meshes, our model can synthesize shapes with arbitrary topology. The core of DMTet includes a deformable tetrahedral grid that encodes a discretized signed distance function and a differentiable marching tetrahedra layer that converts the implicit signed distance representation to the explicit surface mesh representation. This combination allows joint optimization of the surface geometry and topology as well as generation of the hierarchy of subdivisions using reconstruction and adversarial losses defined explicitly on the surface mesh. Our approach significantly outperforms existing work on conditional shape synthesis from coarse voxel inputs, trained on a dataset of complex 3D animal shapes. Project page: https://nv-tlabs.github.io/DMTet/. |
Real-Time Detection of Dictionary DGA Network Traffic using Deep Learning | Botnets and malware continue to avoid detection by static rules engines when using domain generation algorithms (DGAs) for callouts to unique, dynamically generated web addresses. Common DGA detection techniques fail to reliably detect DGA variants that combine random dictionary words to create domain names that closely mirror legitimate domains. To combat this, we created a novel hybrid neural network, Bilbo the `bagging` model, that analyses domains and scores the likelihood they are generated by such algorithms and therefore are potentially malicious. Bilbo is the first parallel usage of a convolutional neural network (CNN) and a long short-term memory (LSTM) network for DGA detection. Our unique architecture is found to be the most consistent in performance in terms of AUC, F1 score, and accuracy when generalising across different dictionary DGA classification tasks compared to current state-of-the-art deep learning architectures. We validate using reverse-engineered dictionary DGA domains and detail our real-time implementation strategy for scoring real-world network logs within a large financial enterprise. In four hours of actual network traffic, the model discovered at least five potential command-and-control networks that commercial vendor tools did not flag. |
Fast Adjustable Threshold For Uniform Neural Network Quantization (Winning solution of LPIRC-II) | Neural network quantization procedure is the necessary step for porting of neural networks to mobile devices. Quantization allows accelerating the inference, reducing memory consumption and model size. It can be performed without fine-tuning using calibration procedure (calculation of parameters necessary for quantization), or it is possible to train the network with quantization from scratch. Training with quantization from scratch on the labeled data is rather long and resource-consuming procedure. Quantization of network without fine-tuning leads to accuracy drop because of outliers which appear during the calibration. In this article we suggest to simplify the quantization procedure significantly by introducing the trained scale factors for quantization thresholds. It allows speeding up the process of quantization with fine-tuning up to 8 epochs as well as reducing the requirements to the set of train images. By our knowledge, the proposed method allowed us to get the first public available quantized version of MNAS without significant accuracy reduction - 74.8% vs 75.3% for original full-precision network. Model and code are ready for use and available at: https://github.com/agoncharenko1992/FAT-fast_adjustable_threshold. |
Score Function Features for Discriminative Learning | Feature learning forms the cornerstone for tackling challenging learning problems in domains such as speech, computer vision and natural language processing. In this paper, we consider a novel class of matrix and tensor-valued features, which can be pre-trained using unlabeled samples. We present efficient algorithms for extracting discriminative information, given these pre-trained features and labeled samples for any related task. Our class of features are based on higher-order score functions, which capture local variations in the probability density function of the input. We establish a theoretical framework to characterize the nature of discriminative information that can be extracted from score-function features, when used in conjunction with labeled samples. We employ efficient spectral decomposition algorithms (on matrices and tensors) for extracting discriminative components. The advantage of employing tensor-valued features is that we can extract richer discriminative information in the form of an overcomplete representations. Thus, we present a novel framework for employing generative models of the input for discriminative learning. |
Scalable Learning With a Structural Recurrent Neural Network for Short-Term Traffic Prediction | This paper presents a scalable deep learning approach for short-term traffic prediction based on historical traffic data in a vehicular road network. Capturing the spatio-temporal relationship of the big data often requires a significant amount of computational burden or an ad-hoc design aiming for a specific type of road network. To tackle the problem, we combine a road network graph with recurrent neural networks (RNNs) to construct a structural RNN (SRNN). The SRNN employs a spatio-temporal graph to infer the interaction between adjacent road segments as well as the temporal dynamics of the time series data. The model is scalable thanks to two key aspects. First, the proposed SRNN architecture is built by using the semantic similarity of the spatio-temporal dynamic interactions of all segments. Second, we design the architecture to deal with fixed-length tensors regardless of the graph topology. With the real traffic speed data measured in the city of Santander, we demonstrate the proposed SRNN outperforms the image-based approaches using the capsule network (CapsNet) by 14.1% and the convolutional neural network (CNN) by 5.87%, respectively, in terms of root mean squared error (RMSE). Moreover, we show that the proposed model is scalable. The SRNN model trained with data of a road network is able to predict traffic speed of different road networks, with the fixed number of parameters to train. |
Hopular: Modern Hopfield Networks for Tabular Data | While Deep Learning excels in structured data as encountered in vision and natural language processing, it failed to meet its expectations on tabular data. For tabular data, Support Vector Machines (SVMs), Random Forests, and Gradient Boosting are the best performing techniques with Gradient Boosting in the lead. Recently, we saw a surge of Deep Learning methods that were tailored to tabular data but still underperform compared to Gradient Boosting on small-sized datasets. We suggest "Hopular", a novel Deep Learning architecture for medium- and small-sized datasets, where each layer is equipped with continuous modern Hopfield networks. The modern Hopfield networks use stored data to identify feature-feature, feature-target, and sample-sample dependencies. Hopular's novelty is that every layer can directly access the original input as well as the whole training set via stored data in the Hopfield networks. Therefore, Hopular can step-wise update its current model and the resulting prediction at every layer like standard iterative learning algorithms. In experiments on small-sized tabular datasets with less than 1,000 samples, Hopular surpasses Gradient Boosting, Random Forests, SVMs, and in particular several Deep Learning methods. In experiments on medium-sized tabular data with about 10,000 samples, Hopular outperforms XGBoost, CatBoost, LightGBM and a state-of-the art Deep Learning method designed for tabular data. Thus, Hopular is a strong alternative to these methods on tabular data. |
Forecasting adverse surgical events using self-supervised transfer learning for physiological signals | Hundreds of millions of surgical procedures take place annually across the world, which generate a prevalent type of electronic health record (EHR) data comprising time series physiological signals. Here, we present a transferable embedding method (i.e., a method to transform time series signals into input features for predictive machine learning models) named PHASE (PHysiologicAl Signal Embeddings) that enables us to more accurately forecast adverse surgical outcomes based on physiological signals. We evaluate PHASE on minute-by-minute EHR data of more than 50,000 surgeries from two operating room (OR) datasets and patient stays in an intensive care unit (ICU) dataset. PHASE outperforms other state-of-the-art approaches, such as long-short term memory networks trained on raw data and gradient boosted trees trained on handcrafted features, in predicting five distinct outcomes: hypoxemia, hypocapnia, hypotension, hypertension, and phenylephrine administration. In a transfer learning setting where we train embedding models in one dataset then embed signals and predict adverse events in unseen data, PHASE achieves significantly higher prediction accuracy at lower computational cost compared to conventional approaches. Finally, given the importance of understanding models in clinical applications we demonstrate that PHASE is explainable and validate our predictive models using local feature attribution methods. |
DyNet: Dynamic Convolution for Accelerating Convolutional Neural Networks | Convolution operator is the core of convolutional neural networks (CNNs) and occupies the most computation cost. To make CNNs more efficient, many methods have been proposed to either design lightweight networks or compress models. Although some efficient network structures have been proposed, such as MobileNet or ShuffleNet, we find that there still exists redundant information between convolution kernels. To address this issue, we propose a novel dynamic convolution method to adaptively generate convolution kernels based on image contents. To demonstrate the effectiveness, we apply dynamic convolution on multiple state-of-the-art CNNs. On one hand, we can reduce the computation cost remarkably while maintaining the performance. For ShuffleNetV2/MobileNetV2/ResNet18/ResNet50, DyNet can reduce 37.0/54.7/67.2/71.3% FLOPs without loss of accuracy. On the other hand, the performance can be largely boosted if the computation cost is maintained. Based on the architecture MobileNetV3-Small/Large, DyNet achieves 70.3/77.1% Top-1 accuracy on ImageNet with an improvement of 2.9/1.9%. To verify the scalability, we also apply DyNet on segmentation task, the results show that DyNet can reduce 69.3% FLOPs while maintaining Mean IoU on segmentation task. |
Neuro-Inspired Deep Neural Networks with Sparse, Strong Activations | While end-to-end training of Deep Neural Networks (DNNs) yields state of the art performance in an increasing array of applications, it does not provide insight into, or control over, the features being extracted. We report here on a promising neuro-inspired approach to DNNs with sparser and stronger activations. We use standard stochastic gradient training, supplementing the end-to-end discriminative cost function with layer-wise costs promoting Hebbian ("fire together," "wire together") updates for highly active neurons, and anti-Hebbian updates for the remaining neurons. Instead of batch norm, we use divisive normalization of activations (suppressing weak outputs using strong outputs), along with implicit $\ell_2$ normalization of neuronal weights. Experiments with standard image classification tasks on CIFAR-10 demonstrate that, relative to baseline end-to-end trained architectures, our proposed architecture (a) leads to sparser activations (with only a slight compromise on accuracy), (b) exhibits more robustness to noise (without being trained on noisy data), (c) exhibits more robustness to adversarial perturbations (without adversarial training). |
SubTab: Subsetting Features of Tabular Data for Self-Supervised Representation Learning | Self-supervised learning has been shown to be very effective in learning useful representations, and yet much of the success is achieved in data types such as images, audio, and text. The success is mainly enabled by taking advantage of spatial, temporal, or semantic structure in the data through augmentation. However, such structure may not exist in tabular datasets commonly used in fields such as healthcare, making it difficult to design an effective augmentation method, and hindering a similar progress in tabular data setting. In this paper, we introduce a new framework, Subsetting features of Tabular data (SubTab), that turns the task of learning from tabular data into a multi-view representation learning problem by dividing the input features to multiple subsets. We argue that reconstructing the data from the subset of its features rather than its corrupted version in an autoencoder setting can better capture its underlying latent representation. In this framework, the joint representation can be expressed as the aggregate of latent variables of the subsets at test time, which we refer to as collaborative inference. Our experiments show that the SubTab achieves the state of the art (SOTA) performance of 98.31% on MNIST in tabular setting, on par with CNN-based SOTA models, and surpasses existing baselines on three other real-world datasets by a significant margin. |
Stochastic Neighbor Embedding with Gaussian and Student-t Distributions: Tutorial and Survey | Stochastic Neighbor Embedding (SNE) is a manifold learning and dimensionality reduction method with a probabilistic approach. In SNE, every point is consider to be the neighbor of all other points with some probability and this probability is tried to be preserved in the embedding space. SNE considers Gaussian distribution for the probability in both the input and embedding spaces. However, t-SNE uses the Student-t and Gaussian distributions in these spaces, respectively. In this tutorial and survey paper, we explain SNE, symmetric SNE, t-SNE (or Cauchy-SNE), and t-SNE with general degrees of freedom. We also cover the out-of-sample extension and acceleration for these methods. Some simulations to visualize the embeddings are also provided. |
PIGNet: A physics-informed deep learning model toward generalized drug-target interaction predictions | Recently, deep neural network (DNN)-based drug-target interaction (DTI) models were highlighted for their high accuracy with affordable computational costs. Yet, the models' insufficient generalization remains a challenging problem in the practice of in-silico drug discovery. We propose two key strategies to enhance generalization in the DTI model. The first is to predict the atom-atom pairwise interactions via physics-informed equations parameterized with neural networks and provides the total binding affinity of a protein-ligand complex as their sum. We further improved the model generalization by augmenting a broader range of binding poses and ligands to training data. We validated our model, PIGNet, in the comparative assessment of scoring functions (CASF) 2016, demonstrating the outperforming docking and screening powers than previous methods. Our physics-informing strategy also enables the interpretation of predicted affinities by visualizing the contribution of ligand substructures, providing insights for further ligand optimization. |
Advocating for Multiple Defense Strategies against Adversarial Examples | It has been empirically observed that defense mechanisms designed to protect neural networks against $\ell_\infty$ adversarial examples offer poor performance against $\ell_2$ adversarial examples and vice versa. In this paper we conduct a geometrical analysis that validates this observation. Then, we provide a number of empirical insights to illustrate the effect of this phenomenon in practice. Then, we review some of the existing defense mechanism that attempts to defend against multiple attacks by mixing defense strategies. Thanks to our numerical experiments, we discuss the relevance of this method and state open questions for the adversarial examples community. |
Big Data for Traffic Estimation and Prediction: A Survey of Data and Tools | Big data has been used widely in many areas including the transportation industry. Using various data sources, traffic states can be well estimated and further predicted for improving the overall operation efficiency. Combined with this trend, this study presents an up-to-date survey of open data and big data tools used for traffic estimation and prediction. Different data types are categorized and the off-the-shelf tools are introduced. To further promote the use of big data for traffic estimation and prediction tasks, challenges and future directions are given for future studies. |
Normalization Before Shaking Toward Learning Symmetrically Distributed Representation Without Margin in Speech Emotion Recognition | Regularization is crucial to the success of many practical deep learning models, in particular in a more often than not scenario where there are only a few to a moderate number of accessible training samples. In addition to weight decay, data augmentation and dropout, regularization based on multi-branch architectures, such as Shake-Shake regularization, has been proven successful in many applications and attracted more and more attention. However, beyond model-based representation augmentation, it is unclear how Shake-Shake regularization helps to provide further improvement on classification tasks, let alone the baffling interaction between batch normalization and shaking. In this work, we present our investigation on Shake-Shake regularization, drawing connections to the vicinal risk minimization principle and discriminative feature learning in verification tasks. Furthermore, we identify a strong resemblance between batch normalized residual blocks and batch normalized recurrent neural networks, where both of them share a similar convergence behavior, which could be mitigated by a proper initialization of batch normalization. Based on the findings, our experiments on speech emotion recognition demonstrate simultaneously an improvement on the classification accuracy and a reduction on the generalization gap both with statistical significance. |
Generation of Policy-Level Explanations for Reinforcement Learning | Though reinforcement learning has greatly benefited from the incorporation of neural networks, the inability to verify the correctness of such systems limits their use. Current work in explainable deep learning focuses on explaining only a single decision in terms of input features, making it unsuitable for explaining a sequence of decisions. To address this need, we introduce Abstracted Policy Graphs, which are Markov chains of abstract states. This representation concisely summarizes a policy so that individual decisions can be explained in the context of expected future transitions. Additionally, we propose a method to generate these Abstracted Policy Graphs for deterministic policies given a learned value function and a set of observed transitions, potentially off-policy transitions used during training. Since no restrictions are placed on how the value function is generated, our method is compatible with many existing reinforcement learning methods. We prove that the worst-case time complexity of our method is quadratic in the number of features and linear in the number of provided transitions, $O(|F|^2 |tr\_samples|)$. By applying our method to a family of domains, we show that our method scales well in practice and produces Abstracted Policy Graphs which reliably capture relationships within these domains. |
Can Active Learning Preemptively Mitigate Fairness Issues? | Dataset bias is one of the prevailing causes of unfairness in machine learning. Addressing fairness at the data collection and dataset preparation stages therefore becomes an essential part of training fairer algorithms. In particular, active learning (AL) algorithms show promise for the task by drawing importance to the most informative training samples. However, the effect and interaction between existing AL algorithms and algorithmic fairness remain under-explored. In this paper, we study whether models trained with uncertainty-based AL heuristics such as BALD are fairer in their decisions with respect to a protected class than those trained with identically independently distributed (i.i.d.) sampling. We found a significant improvement on predictive parity when using BALD, while also improving accuracy compared to i.i.d. sampling. We also explore the interaction of algorithmic fairness methods such as gradient reversal (GRAD) and BALD. We found that, while addressing different fairness issues, their interaction further improves the results on most benchmarks and metrics we explored. |
A Conceptual Framework for Establishing Trust in Real World Intelligent Systems | Intelligent information systems that contain emergent elements often encounter trust problems because results do not get sufficiently explained and the procedure itself can not be fully retraced. This is caused by a control flow depending either on stochastic elements or on the structure and relevance of the input data. Trust in such algorithms can be established by letting users interact with the system so that they can explore results and find patterns that can be compared with their expected solution. Reflecting features and patterns of human understanding of a domain against algorithmic results can create awareness of such patterns and may increase the trust that a user has in the solution. If expectations are not met, close inspection can be used to decide whether a solution conforms to the expectations or whether it goes beyond the expected. By either accepting or rejecting a solution, the user's set of expectations evolves and a learning process for the users is established. In this paper we present a conceptual framework that reflects and supports this process. The framework is the result of an analysis of two exemplary case studies from two different disciplines with information systems that assist experts in their complex tasks. |
BliStr: The Blind Strategymaker | BliStr is a system that automatically develops strategies for E prover on a large set of problems. The main idea is to interleave (i) iterated low-timelimit local search for new strategies on small sets of similar easy problems with (ii) higher-timelimit evaluation of the new strategies on all problems. The accumulated results of the global higher-timelimit runs are used to define and evolve the notion of "similar easy problems", and to control the selection of the next strategy to be improved. The technique was used to significantly strengthen the set of E strategies used by the MaLARea, PS-E, E-MaLeS, and E systems in the CASC@Turing 2012 competition, particularly in the Mizar division. Similar improvement was obtained on the problems created from the Flyspeck corpus. |
A Dual Process Model for Optimizing Cross Entropy in Neural Networks | Minimizing cross-entropy is a widely used method for training artificial neural networks. Many training procedures based on backpropagation use cross-entropy directly as their loss function. Instead, this theoretical essay investigates a dual process model with two processes, in which one process minimizes the Kullback-Leibler divergence while its dual counterpart minimizes the Shannon entropy. Postulating that learning consists of two dual processes complementing each other, the model defines an equilibrium state for both processes in which the loss function assumes its minimum. An advantage of the proposed model is that it allows deriving the optimal learning rate and momentum weight to update network weights for backpropagation. Furthermore, the model introduces the golden ratio and complex numbers as important new concepts in machine learning. |
HyBNN and FedHyBNN: (Federated) Hybrid Binary Neural Networks | Binary Neural Networks (BNNs), neural networks with weights and activations constrained to -1(0) and +1, are an alternative to deep neural networks which offer faster training, lower memory consumption and lightweight models, ideal for use in resource constrained devices while being able to utilize the architecture of their deep neural network counterpart. However, the input binarization step used in BNNs causes a severe accuracy loss. In this paper, we introduce a novel hybrid neural network architecture, Hybrid Binary Neural Network (HyBNN), consisting of a task-independent, general, full-precision variational autoencoder with a binary latent space and a task specific binary neural network that is able to greatly limit the accuracy loss due to input binarization by using the full precision variational autoencoder as a feature extractor. We use it to combine the state-of-the-art accuracy of deep neural networks with the much faster training time, quicker test-time inference and power efficiency of binary neural networks. We show that our proposed system is able to very significantly outperform a vanilla binary neural network with input binarization. We also introduce FedHyBNN, a highly communication efficient federated counterpart to HyBNN and demonstrate that it is able to reach the same accuracy as its non-federated equivalent. We make our source code, experimental parameters and models available at: https://anonymous.4open.science/r/HyBNN. |
Why resampling outperforms reweighting for correcting sampling bias with stochastic gradients | A data set sampled from a certain population is biased if the subgroups of the population are sampled at proportions that are significantly different from their underlying proportions. Training machine learning models on biased data sets requires correction techniques to compensate for the bias. We consider two commonly-used techniques, resampling and reweighting, that rebalance the proportions of the subgroups to maintain the desired objective function. Though statistically equivalent, it has been observed that resampling outperforms reweighting when combined with stochastic gradient algorithms. By analyzing illustrative examples, we explain the reason behind this phenomenon using tools from dynamical stability and stochastic asymptotics. We also present experiments from regression, classification, and off-policy prediction to demonstrate that this is a general phenomenon. We argue that it is imperative to consider the objective function design and the optimization algorithm together while addressing the sampling bias. |
Memristive Stochastic Computing for Deep Learning Parameter Optimization | Stochastic Computing (SC) is a computing paradigm that allows for the low-cost and low-power computation of various arithmetic operations using stochastic bit streams and digital logic. In contrast to conventional representation schemes used within the binary domain, the sequence of bit streams in the stochastic domain is inconsequential, and computation is usually non-deterministic. In this brief, we exploit the stochasticity during switching of probabilistic Conductive Bridging RAM (CBRAM) devices to efficiently generate stochastic bit streams in order to perform Deep Learning (DL) parameter optimization, reducing the size of Multiply and Accumulate (MAC) units by 5 orders of magnitude. We demonstrate that in using a 40-nm Complementary Metal Oxide Semiconductor (CMOS) process our scalable architecture occupies 1.55mm$^2$ and consumes approximately 167$\mu$W when optimizing parameters of a Convolutional Neural Network (CNN) while it is being trained for a character recognition task, observing no notable reduction in accuracy post-training. |
Sparse Object-level Supervision for Instance Segmentation with Pixel Embeddings | Most state-of-the-art instance segmentation methods have to be trained on densely annotated images. While difficult in general, this requirement is especially daunting for biomedical images, where domain expertise is often required for annotation and no large public data collections are available for pre-training. We propose to address the dense annotation bottleneck by introducing a proposal-free segmentation approach based on non-spatial embeddings, which exploits the structure of the learned embedding space to extract individual instances in a differentiable way. The segmentation loss can then be applied directly to instances and the overall pipeline can be trained in a fully- or weakly supervised manner. We consider the challenging case of positive-unlabeled supervision, where a novel self-supervised consistency loss is introduced for the unlabeled parts of the training data. We evaluate the proposed method on 2D and 3D segmentation problems in different microscopy modalities as well as on the Cityscapes and CVPPP instance segmentation benchmarks, achieving state-of-the-art results on the latter. The code is available at: https://github.com/kreshuklab/spoco |
A Pre-processing Method for Fairness in Ranking | Fair ranking problems arise in many decision-making processes that often necessitate a trade-off between accuracy and fairness. Many existing studies have proposed correction methods such as adding fairness constraints to a ranking model's loss. However, the challenge of correcting the data bias for fair ranking remains, and the trade-off of the ranking models leaves room for improvement. In this paper, we propose a fair ranking framework that evaluates the order of training data in a pairwise manner as well as various fairness measurements in ranking. This study is the first proposal of a pre-processing method that solves fair ranking problems using the pairwise ordering method with our best knowledge. The fair pairwise ordering method is prominent in training the fair ranking models because it ensures that the resulting ranking likely becomes parity across groups. As far as the fairness measurements in ranking are represented as a linear constraint of the ranking models, we proved that the minimization of loss function subject to the constraints is reduced to the closed solution of the minimization problem augmented by weights to training data. This closed solution inspires us to present a practical and stable algorithm that iterates the optimization of weights and model parameters. The empirical results over real-world datasets demonstrated that our method outperforms the existing methods in the trade-off between accuracy and fairness over real-world datasets and various fairness measurements. |
On Learning to Rank Long Sequences with Contextual Bandits | Motivated by problems of learning to rank long item sequences, we introduce a variant of the cascading bandit model that considers flexible length sequences with varying rewards and losses. We formulate two generative models for this problem within the generalized linear setting, and design and analyze upper confidence algorithms for it. Our analysis delivers tight regret bounds which, when specialized to vanilla cascading bandits, results in sharper guarantees than previously available in the literature. We evaluate our algorithms on a number of real-world datasets, and show significantly improved empirical performance as compared to known cascading bandit baselines. |
Self-Assembling Modular Networks for Interpretable Multi-Hop Reasoning | Multi-hop QA requires a model to connect multiple pieces of evidence scattered in a long context to answer the question. The recently proposed HotpotQA (Yang et al., 2018) dataset is comprised of questions embodying four different multi-hop reasoning paradigms (two bridge entity setups, checking multiple properties, and comparing two entities), making it challenging for a single neural network to handle all four. In this work, we present an interpretable, controller-based Self-Assembling Neural Modular Network (Hu et al., 2017, 2018) for multi-hop reasoning, where we design four novel modules (Find, Relocate, Compare, NoOp) to perform unique types of language reasoning. Based on a question, our layout controller RNN dynamically infers a series of reasoning modules to construct the entire network. Empirically, we show that our dynamic, multi-hop modular network achieves significant improvements over the static, single-hop baseline (on both regular and adversarial evaluation). We further demonstrate the interpretability of our model via three analyses. First, the controller can softly decompose the multi-hop question into multiple single-hop sub-questions to promote compositional reasoning behavior of the main network. Second, the controller can predict layouts that conform to the layouts designed by human experts. Finally, the intermediate module can infer the entity that connects two distantly-located supporting facts by addressing the sub-question from the controller. |
Data Assimilation in the Latent Space of a Neural Network | There is an urgent need to build models to tackle Indoor Air Quality issue. Since the model should be accurate and fast, Reduced Order Modelling technique is used to reduce the dimensionality of the problem. The accuracy of the model, that represent a dynamic system, is improved integrating real data coming from sensors using Data Assimilation techniques. In this paper, we formulate a new methodology called Latent Assimilation that combines Data Assimilation and Machine Learning. We use a Convolutional neural network to reduce the dimensionality of the problem, a Long-Short-Term-Memory to build a surrogate model of the dynamic system and an Optimal Interpolated Kalman Filter to incorporate real data. Experimental results are provided for CO2 concentration within an indoor space. This methodology can be used for example to predict in real-time the load of virus, such as the SARS-COV-2, in the air by linking it to the concentration of CO2. |
TricycleGAN: Unsupervised Image Synthesis and Segmentation Based on Shape Priors | Medical image segmentation is routinely performed to isolate regions of interest, such as organs and lesions. Currently, deep learning is the state of the art for automatic segmentation, but is usually limited by the need for supervised training with large datasets that have been manually segmented by trained clinicians. The goal of semi-superised and unsupervised image segmentation is to greatly reduce, or even eliminate, the need for training data and therefore to minimze the burden on clinicians when training segmentation models. To this end we introduce a novel network architecture for capable of unsupervised and semi-supervised image segmentation called TricycleGAN. This approach uses three generative models to learn translations between medical images and segmentation maps using edge maps as an intermediate step. Distinct from other approaches based on generative networks, TricycleGAN relies on shape priors rather than colour and texture priors. As such, it is particularly well-suited for several domains of medical imaging, such as ultrasound imaging, where commonly used visual cues may be absent. We present experiments with TricycleGAN on a clinical dataset of kidney ultrasound images and the benchmark ISIC 2018 skin lesion dataset. |
Predicting Medical Interventions from Vital Parameters: Towards a Decision Support System for Remote Patient Monitoring | Cardiovascular diseases and heart failures in particular are the main cause of non-communicable disease mortality in the world. Constant patient monitoring enables better medical treatment as it allows practitioners to react on time and provide the appropriate treatment. Telemedicine can provide constant remote monitoring so patients can stay in their homes, only requiring medical sensing equipment and network connections. A limiting factor for telemedical centers is the amount of patients that can be monitored simultaneously. We aim to increase this amount by implementing a decision support system. This paper investigates a machine learning model to estimate a risk score based on patient vital parameters that allows sorting all cases every day to help practitioners focus their limited capacities on the most severe cases. The model we propose reaches an AUCROC of 0.84, whereas the baseline rule-based model reaches an AUCROC of 0.73. Our results indicate that the usage of deep learning to improve the efficiency of telemedical centers is feasible. This way more patients could benefit from better health-care through remote monitoring. |
Fast Decomposable Submodular Function Minimization using Constrained Total Variation | We consider the problem of minimizing the sum of submodular set functions assuming minimization oracles of each summand function. Most existing approaches reformulate the problem as the convex minimization of the sum of the corresponding Lov\'asz extensions and the squared Euclidean norm, leading to algorithms requiring total variation oracles of the summand functions; without further assumptions, these more complex oracles require many calls to the simpler minimization oracles often available in practice. In this paper, we consider a modified convex problem requiring constrained version of the total variation oracles that can be solved with significantly fewer calls to the simple minimization oracles. We support our claims by showing results on graph cuts for 2D and 3D graphs |
A Deep Value-network Based Approach for Multi-Driver Order Dispatching | Recent works on ride-sharing order dispatching have highlighted the importance of taking into account both the spatial and temporal dynamics in the dispatching process for improving the transportation system efficiency. At the same time, deep reinforcement learning has advanced to the point where it achieves superhuman performance in a number of fields. In this work, we propose a deep reinforcement learning based solution for order dispatching and we conduct large scale online A/B tests on DiDi's ride-dispatching platform to show that the proposed method achieves significant improvement on both total driver income and user experience related metrics. In particular, we model the ride dispatching problem as a Semi Markov Decision Process to account for the temporal aspect of the dispatching actions. To improve the stability of the value iteration with nonlinear function approximators like neural networks, we propose Cerebellar Value Networks (CVNet) with a novel distributed state representation layer. We further derive a regularized policy evaluation scheme for CVNet that penalizes large Lipschitz constant of the value network for additional robustness against adversarial perturbation and noises. Finally, we adapt various transfer learning methods to CVNet for increased learning adaptability and efficiency across multiple cities. We conduct extensive offline simulations based on real dispatching data as well as online AB tests through the DiDi's platform. Results show that CVNet consistently outperforms other recently proposed dispatching methods. We finally show that the performance can be further improved through the efficient use of transfer learning. |
Self-Imitation Learning via Generalized Lower Bound Q-learning | Self-imitation learning motivated by lower-bound Q-learning is a novel and effective approach for off-policy learning. In this work, we propose a n-step lower bound which generalizes the original return-based lower-bound Q-learning, and introduce a new family of self-imitation learning algorithms. To provide a formal motivation for the potential performance gains provided by self-imitation learning, we show that n-step lower bound Q-learning achieves a trade-off between fixed point bias and contraction rate, drawing close connections to the popular uncorrected n-step Q-learning. We finally show that n-step lower bound Q-learning is a more robust alternative to return-based self-imitation learning and uncorrected n-step, over a wide range of continuous control benchmark tasks. |
Transient-evoked otoacoustic emission signals predicting outcomes of acute sensorineural hearing loss in patients with Meniere's Disease | Background: Fluctuating hearing loss is characteristic of Meniere's Disease (MD) during acute episodes. However, no reliable audiometric hallmarks are available for counselling the hearing recovery possibility. Aims/Objectives: To find parameters for predicting MD hearing outcomes. Material and Methods: We applied machine learning techniques to analyse transient-evoked otoacoustic emission (TEOAE) signals recorded from patients with MD. Thirty unilateral MD patients were recruited prospectively after onset of acute cochleo-vestibular symptoms. Serial TEOAE and pure-tone audiogram (PTA) data were recorded longitudinally. Denoised TEOAE signals were projected onto the three most prominent principal directions through a linear transformation. Binary classification was performed using a support vector machine (SVM). TEOAE signal parameters, including signal energy and group delay, were compared between improved and nonimproved groups using Welchs t-test. Results: Signal energy did not differ (p = 0.64) but a significant difference in 1-kHz (p = 0.045) group delay was recorded between improved and nonimproved groups. The SVM achieved a cross-validated accuracy of >80% in predicting hearing outcomes. Conclusions and Significance: This study revealed that baseline TEOAE parameters obtained during acute MD episodes, when processed through machine learning technology, may provide information on outer hair cell function to predict hearing recovery. |
Causal Modeling for Fairness in Dynamical Systems | In many application areas---lending, education, and online recommenders, for example---fairness and equity concerns emerge when a machine learning system interacts with a dynamically changing environment to produce both immediate and long-term effects for individuals and demographic groups. We discuss causal directed acyclic graphs (DAGs) as a unifying framework for the recent literature on fairness in such dynamical systems. We show that this formulation affords several new directions of inquiry to the modeler, where causal assumptions can be expressed and manipulated. We emphasize the importance of computing interventional quantities in the dynamical fairness setting, and show how causal assumptions enable simulation (when environment dynamics are known) and off-policy estimation (when dynamics are unknown) of intervention on short- and long-term outcomes, at both the group and individual levels. |
Mention Memory: incorporating textual knowledge into Transformers through entity mention attention | Natural language understanding tasks such as open-domain question answering often require retrieving and assimilating factual information from multiple sources. We propose to address this problem by integrating a semi-parametric representation of a large text corpus into a Transformer model as a source of factual knowledge. Specifically, our method represents knowledge with `mention memory', a table of dense vector representations of every entity mention in a corpus. The proposed model - TOME - is a Transformer that accesses the information through internal memory layers in which each entity mention in the input passage attends to the mention memory. This approach enables synthesis of and reasoning over many disparate sources of information within a single Transformer model. In experiments using a memory of 150 million Wikipedia mentions, TOME achieves strong performance on several open-domain knowledge-intensive tasks, including the claim verification benchmarks HoVer and FEVER and several entity-based QA benchmarks. We also show that the model learns to attend to informative mentions without any direct supervision. Finally we demonstrate that the model can generalize to new unseen entities by updating the memory without retraining. |
Leveraging Acoustic Cues and Paralinguistic Embeddings to Detect Expression from Voice | Millions of people reach out to digital assistants such as Siri every day, asking for information, making phone calls, seeking assistance, and much more. The expectation is that such assistants should understand the intent of the users query. Detecting the intent of a query from a short, isolated utterance is a difficult task. Intent cannot always be obtained from speech-recognized transcriptions. A transcription driven approach can interpret what has been said but fails to acknowledge how it has been said, and as a consequence, may ignore the expression present in the voice. Our work investigates whether a system can reliably detect vocal expression in queries using acoustic and paralinguistic embedding. Results show that the proposed method offers a relative equal error rate (EER) decrease of 60% compared to a bag-of-word based system, corroborating that expression is significantly represented by vocal attributes, rather than being purely lexical. Addition of emotion embedding helped to reduce the EER by 30% relative to the acoustic embedding, demonstrating the relevance of emotion in expressive voice. |
Survey and Systematization of 3D Object Detection Models and Methods | This paper offers a comprehensive survey of recent developments in 3D object detection covering the full pipeline from input data, over data representation and feature extraction to the actual detection modules. We include basic concepts, focus our survey on a broad spectrum of different approaches arising in the last ten years and propose a systematization which offers a practical framework to compare those approaches on the methods level. |
ROI Constrained Bidding via Curriculum-Guided Bayesian Reinforcement Learning | Real-Time Bidding (RTB) is an important mechanism in modern online advertising systems. Advertisers employ bidding strategies in RTB to optimize their advertising effects subject to various financial requirements, among which a widely adopted one is the return-on-investment (ROI) constraint. ROIs change non-monotonically during the sequential bidding process, usually presenting a see-saw effect between constraint satisfaction and objective optimization. Existing solutions to the constraint-objective trade-off are typically established in static or mildly changing markets. However, these methods fail significantly in non-stationary advertising markets due to their inability to adapt to varying dynamics and partial observability. In this work, we specialize in ROI-Constrained Bidding in non-stationary markets. Based on a Partially Observable Constrained Markov Decision Process, we propose the first hard barrier solution to accommodate non-monotonic constraints. Our method exploits a parameter-free indicator-augmented reward function and develops a Curriculum-Guided Bayesian Reinforcement Learning (CBRL) framework to adaptively control the constraint-objective trade-off in non-stationary advertising markets. Extensive experiments on a large-scale industrial dataset with two problem settings reveal that CBRL generalizes well in both in-distribution and out-of-distribution data regimes, and enjoys outstanding stability. |
Determining Optimal Number of k-Clusters based on Predefined Level-of-Similarity | This paper proposes a centroid-based clustering algorithm which is capable of clustering data-points with n-features, without having to specify the number of clusters to be formed. The core logic behind the algorithm is a similarity measure, which collectively decides whether to assign an incoming data-point to a pre-existing cluster, or create a new cluster and assign the data-point to it. The proposed clustering algorithm is application-specific and is applicable when the need is to perform clustering analysis of a stream of data-points, where the similarity measure between an incoming data-point and the cluster to which the data-point is to be associated with, is greater than the predefined Level-of-Similarity. |
Learning Optimal Antenna Tilt Control Policies: A Contextual Linear Bandit Approach | Controlling antenna tilts in cellular networks is imperative to reach an efficient trade-off between network coverage and capacity. In this paper, we devise algorithms learning optimal tilt control policies from existing data (in the so-called passive learning setting) or from data actively generated by the algorithms (the active learning setting). We formalize the design of such algorithms as a Best Policy Identification (BPI) problem in Contextual Linear Multi-Arm Bandits (CL-MAB). An arm represents an antenna tilt update; the context captures current network conditions; the reward corresponds to an improvement of performance, mixing coverage and capacity; and the objective is to identify, with a given level of confidence, an approximately optimal policy (a function mapping the context to an arm with maximal reward). For CL-MAB in both active and passive learning settings, we derive information-theoretical lower bounds on the number of samples required by any algorithm returning an approximately optimal policy with a given level of certainty, and devise algorithms achieving these fundamental limits. We apply our algorithms to the Remote Electrical Tilt (RET) optimization problem in cellular networks, and show that they can produce optimal tilt update policy using much fewer data samples than naive or existing rule-based learning algorithms. |
Specialized Decision Surface and Disentangled Feature for Weakly-Supervised Polyphonic Sound Event Detection | In this paper, a special decision surface for the weakly-supervised sound event detection (SED) and a disentangled feature (DF) for the multi-label problem in polyphonic SED are proposed. We approach SED as a multiple instance learning (MIL) problem and utilize a neural network framework with a pooling module to solve it. General MIL approaches include two kinds: the instance-level approaches and embedding-level approaches. We present a method of generating instance-level probabilities for the embedding level approaches which tend to perform better than the instance-level approaches in terms of bag-level classification but can not provide instance-level probabilities in current approaches. Moreover, we further propose a specialized decision surface (SDS) for the embedding-level attention pooling. We analyze and explained why an embedding-level attention module with SDS is better than other typical pooling modules from the perspective of the high-level feature space. As for the problem of the unbalanced dataset and the co-occurrence of multiple categories in the polyphonic event detection task, we propose a DF to reduce interference among categories, which optimizes the high-level feature space by disentangling it based on class-wise identifiable information and obtaining multiple different subspaces. Experiments on the dataset of DCASE 2018 Task 4 show that the proposed SDS and DF significantly improve the detection performance of the embedding-level MIL approach with an attention pooling module and outperform the first place system in the challenge by 6.6 percentage points. |
SVM-based Multiview Face Recognition by Generalization of Discriminant Analysis | Identity verification of authentic persons by their multiview faces is a real valued problem in machine vision. Multiview faces are having difficulties due to non-linear representation in the feature space. This paper illustrates the usability of the generalization of LDA in the form of canonical covariate for face recognition to multiview faces. In the proposed work, the Gabor filter bank is used to extract facial features that characterized by spatial frequency, spatial locality and orientation. Gabor face representation captures substantial amount of variations of the face instances that often occurs due to illumination, pose and facial expression changes. Convolution of Gabor filter bank to face images of rotated profile views produce Gabor faces with high dimensional features vectors. Canonical covariate is then used to Gabor faces to reduce the high dimensional feature spaces into low dimensional subspaces. Finally, support vector machines are trained with canonical sub-spaces that contain reduced set of features and perform recognition task. The proposed system is evaluated with UMIST face database. The experiment results demonstrate the efficiency and robustness of the proposed system with high recognition rates. |
Shift Invariance Can Reduce Adversarial Robustness | Shift invariance is a critical property of CNNs that improves performance on classification. However, we show that invariance to circular shifts can also lead to greater sensitivity to adversarial attacks. We first characterize the margin between classes when a shift-invariant linear classifier is used. We show that the margin can only depend on the DC component of the signals. Then, using results about infinitely wide networks, we show that in some simple cases, fully connected and shift-invariant neural networks produce linear decision boundaries. Using this, we prove that shift invariance in neural networks produces adversarial examples for the simple case of two classes, each consisting of a single image with a black or white dot on a gray background. This is more than a curiosity; we show empirically that with real datasets and realistic architectures, shift invariance reduces adversarial robustness. Finally, we describe initial experiments using synthetic data to probe the source of this connection. |
Author's Sentiment Prediction | We introduce PerSenT, a dataset of crowd-sourced annotations of the sentiment expressed by the authors towards the main entities in news articles. The dataset also includes paragraph-level sentiment annotations to provide more fine-grained supervision for the task. Our benchmarks of multiple strong baselines show that this is a difficult classification task. The results also suggest that simply fine-tuning document-level representations from BERT isn't adequate for this task. Making paragraph-level decisions and aggregating them over the entire document is also ineffective. We present empirical and qualitative analyses that illustrate the specific challenges posed by this dataset. We release this dataset with 5.3k documents and 38k paragraphs covering 3.2k unique entities as a challenge in entity sentiment analysis. |
DeepCorn: A Semi-Supervised Deep Learning Method for High-Throughput Image-Based Corn Kernel Counting and Yield Estimation | The success of modern farming and plant breeding relies on accurate and efficient collection of data. For a commercial organization that manages large amounts of crops, collecting accurate and consistent data is a bottleneck. Due to limited time and labor, accurately phenotyping crops to record color, head count, height, weight, etc. is severely limited. However, this information, combined with other genetic and environmental factors, is vital for developing new superior crop species that help feed the world's growing population. Recent advances in machine learning, in particular deep learning, have shown promise in mitigating this bottleneck. In this paper, we propose a novel deep learning method for counting on-ear corn kernels in-field to aid in the gathering of real-time data and, ultimately, to improve decision making to maximize yield. We name this approach DeepCorn, and show that this framework is robust under various conditions. DeepCorn estimates the density of corn kernels in an image of corn ears and predicts the number of kernels based on the estimated density map. DeepCorn uses a truncated VGG-16 as a backbone for feature extraction and merges feature maps from multiple scales of the network to make it robust against image scale variations. We also adopt a semi-supervised learning approach to further improve the performance of our proposed method. Our proposed method achieves the MAE and RMSE of 41.36 and 60.27 in the corn kernel counting task, respectively. Our experimental results demonstrate the superiority and effectiveness of our proposed method compared to other state-of-the-art methods. |
Sliced-Wasserstein Gradient Flows | Minimizing functionals in the space of probability distributions can be done with Wasserstein gradient flows. To solve them numerically, a possible approach is to rely on the Jordan-Kinderlehrer-Otto (JKO) scheme which is analogous to the proximal scheme in Euclidean spaces. However, it requires solving a nested optimization problem at each iteration, and is known for its computational challenges, especially in high dimension. To alleviate it, very recent works propose to approximate the JKO scheme leveraging Brenier's theorem, and using gradients of Input Convex Neural Networks to parameterize the density (JKO-ICNN). However, this method comes with a high computational cost and stability issues. Instead, this work proposes to use gradient flows in the space of probability measures endowed with the sliced-Wasserstein (SW) distance. We argue that this method is more flexible than JKO-ICNN, since SW enjoys a closed-form differentiable approximation. Thus, the density at each step can be parameterized by any generative model which alleviates the computational burden and makes it tractable in higher dimensions. |
Soft-CP: A Credible and Effective Data Augmentation for Semantic Segmentation of Medical Lesions | The medical datasets are usually faced with the problem of scarcity and data imbalance. Moreover, annotating large datasets for semantic segmentation of medical lesions is domain-knowledge and time-consuming. In this paper, we propose a new object-blend method(short in soft-CP) that combines the Copy-Paste augmentation method for semantic segmentation of medical lesions offline, ensuring the correct edge information around the lession to solve the issue above-mentioned. We proved the method's validity with several datasets in different imaging modalities. In our experiments on the KiTS19[2] dataset, Soft-CP outperforms existing medical lesions synthesis approaches. The Soft-CP augementation provides gains of +26.5% DSC in the low data regime(10% of data) and +10.2% DSC in the high data regime(all of data), In offline training data, the ratio of real images to synthetic images is 3:1. |
Local Unsupervised Learning for Image Analysis | Local Hebbian learning is believed to be inferior in performance to end-to-end training using a backpropagation algorithm. We question this popular belief by designing a local algorithm that can learn convolutional filters at scale on large image datasets. These filters combined with patch normalization and very steep non-linearities result in a good classification accuracy for shallow networks trained locally, as opposed to end-to-end. The filters learned by our algorithm contain both orientation selective units and unoriented color units, resembling the responses of pyramidal neurons located in the cytochrome oxidase 'interblob' and 'blob' regions in the primary visual cortex of primates. It is shown that convolutional networks with patch normalization significantly outperform standard convolutional networks on the task of recovering the original classes when shadows are superimposed on top of standard CIFAR-10 images. Patch normalization approximates the retinal adaptation to the mean light intensity, important for human vision. We also demonstrate a successful transfer of learned representations between CIFAR-10 and ImageNet 32x32 datasets. All these results taken together hint at the possibility that local unsupervised training might be a powerful tool for learning general representations (without specifying the task) directly from unlabeled data. |
Propagation Networks for Model-Based Control Under Partial Observation | There has been an increasing interest in learning dynamics simulators for model-based control. Compared with off-the-shelf physics engines, a learnable simulator can quickly adapt to unseen objects, scenes, and tasks. However, existing models like interaction networks only work for fully observable systems; they also only consider pairwise interactions within a single time step, both restricting their use in practical systems. We introduce Propagation Networks (PropNet), a differentiable, learnable dynamics model that handles partially observable scenarios and enables instantaneous propagation of signals beyond pairwise interactions. Experiments show that our propagation networks not only outperform current learnable physics engines in forward simulation, but also achieve superior performance on various control tasks. Compared with existing model-free deep reinforcement learning algorithms, model-based control with propagation networks is more accurate, efficient, and generalizable to new, partially observable scenes and tasks. |
Neural Operator with Regularity Structure for Modeling Dynamics Driven by SPDEs | Stochastic partial differential equations (SPDEs) are significant tools for modeling dynamics in many areas including atmospheric sciences and physics. Neural Operators, generations of neural networks with capability of learning maps between infinite-dimensional spaces, are strong tools for solving parametric PDEs. However, they lack the ability to modeling SPDEs which usually have poor regularity due to the driving noise. As the theory of regularity structure has achieved great successes in analyzing SPDEs and provides the concept model feature vectors that well-approximate SPDEs' solutions, we propose the Neural Operator with Regularity Structure (NORS) which incorporates the feature vectors for modeling dynamics driven by SPDEs. We conduct experiments on various of SPDEs including the dynamic Phi41 model and the 2d stochastic Navier-Stokes equation, and the results demonstrate that the NORS is resolution-invariant, efficient, and achieves one order of magnitude lower error with a modest amount of data. |
A Predictive On-Demand Placement of UAV Base Stations Using Echo State Network | The unmanned aerial vehicles base stations (UAV-BSs) have great potential in being widely used in many dynamic application scenarios. In those scenarios, the movements of served user equipments (UEs) are inevitable, so the UAV-BSs needs to be re-positioned dynamically for providing seamless services. In this paper, we propose a system framework consisting of UEs clustering, UAV-BS placement, UEs trajectories prediction, and UAV-BS reposition matching scheme, to serve the UEs seamlessly as well as minimize the energy cost of UAV-BSs' reposition trajectories. An Echo State Network (ESN) based algorithm for predicting the future trajectories of UEs and a Kuhn-Munkres-based algorithm for finding the energy-efficient reposition trajectories of UAV-BSs is designed, respectively. We conduct a simulation using a real open dataset for performance validation. The simulation results indicate that the proposed framework achieves high prediction accuracy and provides the energy-efficient matching scheme. |
Skyline Identification in Multi-Armed Bandits | We introduce a variant of the classical PAC multi-armed bandit problem. There is an ordered set of $n$ arms $A[1],\dots,A[n]$, each with some stochastic reward drawn from some unknown bounded distribution. The goal is to identify the $skyline$ of the set $A$, consisting of all arms $A[i]$ such that $A[i]$ has larger expected reward than all lower-numbered arms $A[1],\dots,A[i-1]$. We define a natural notion of an $\varepsilon$-approximate skyline and prove matching upper and lower bounds for identifying an $\varepsilon$-skyline. Specifically, we show that in order to identify an $\varepsilon$-skyline from among $n$ arms with probability $1-\delta$, $$ \Theta\bigg(\frac{n}{\varepsilon^2} \cdot \min\bigg\{ \log\bigg(\frac{1}{\varepsilon \delta}\bigg), \log\bigg(\frac{n}{\delta}\bigg) \bigg\} \bigg) $$ samples are necessary and sufficient. When $\varepsilon \gg 1/n$, our results improve over the naive algorithm, which draws enough samples to approximate the expected reward of every arm; the algorithm of (Auer et al., AISTATS'16) for Pareto-optimal arm identification is likewise superseded. Our results show that the sample complexity of the skyline problem lies strictly in between that of best arm identification (Even-Dar et al., COLT'02) and that of approximating the expected reward of every arm. |
Learning to Reason with Neural Networks: Generalization, Unseen Data and Boolean Measures | This paper considers the Pointer Value Retrieval (PVR) benchmark introduced in [ZRKB21], where a 'reasoning' function acts on a string of digits to produce the label. More generally, the paper considers the learning of logical functions with gradient descent (GD) on neural networks. It is first shown that in order to learn logical functions with gradient descent on symmetric neural networks, the generalization error can be lower-bounded in terms of the noise-stability of the target function, supporting a conjecture made in [ZRKB21]. It is then shown that in the distribution shift setting, when the data withholding corresponds to freezing a single feature (referred to as canonical holdout), the generalization error of gradient descent admits a tight characterization in terms of the Boolean influence for several relevant architectures. This is shown on linear models and supported experimentally on other models such as MLPs and Transformers. In particular, this puts forward the hypothesis that for such architectures and for learning logical functions such as PVR functions, GD tends to have an implicit bias towards low-degree representations, which in turn gives the Boolean influence for the generalization error under quadratic loss. |
AReS and MaRS - Adversarial and MMD-Minimizing Regression for SDEs | Stochastic differential equations are an important modeling class in many disciplines. Consequently, there exist many methods relying on various discretization and numerical integration schemes. In this paper, we propose a novel, probabilistic model for estimating the drift and diffusion given noisy observations of the underlying stochastic system. Using state-of-the-art adversarial and moment matching inference techniques, we avoid the discretization schemes of classical approaches. This leads to significant improvements in parameter accuracy and robustness given random initial guesses. On four established benchmark systems, we compare the performance of our algorithms to state-of-the-art solutions based on extended Kalman filtering and Gaussian processes. |
PipeSim: Trace-driven Simulation of Large-Scale AI Operations Platforms | Operationalizing AI has become a major endeavor in both research and industry. Automated, operationalized pipelines that manage the AI application lifecycle will form a significant part of tomorrow's infrastructure workloads. To optimize operations of production-grade AI workflow platforms we can leverage existing scheduling approaches, yet it is challenging to fine-tune operational strategies that achieve application-specific cost-benefit tradeoffs while catering to the specific domain characteristics of machine learning (ML) models, such as accuracy, robustness, or fairness. We present a trace-driven simulation-based experimentation and analytics environment that allows researchers and engineers to devise and evaluate such operational strategies for large-scale AI workflow systems. Analytics data from a production-grade AI platform developed at IBM are used to build a comprehensive simulation model. Our simulation model describes the interaction between pipelines and system infrastructure, and how pipeline tasks affect different ML model metrics. We implement the model in a standalone, stochastic, discrete event simulator, and provide a toolkit for running experiments. Synthetic traces are made available for ad-hoc exploration as well as statistical analysis of experiments to test and examine pipeline scheduling, cluster resource allocation, and similar operational mechanisms. |
Predicting invasive ductal carcinoma using a Reinforcement Sample Learning Strategy using Deep Learning | Invasive ductal carcinoma is a prevalent, potentially deadly disease associated with a high rate of morbidity and mortality. Its malignancy is the second leading cause of death from cancer in women. The mammogram is an extremely useful resource for mass detection and invasive ductal carcinoma diagnosis. We are proposing a method for Invasive ductal carcinoma that will use convolutional neural networks (CNN) on mammograms to assist radiologists in diagnosing the disease. Due to the varying image clarity and structure of certain mammograms, it is difficult to observe major cancer characteristics such as microcalcification and mass, and it is often difficult to interpret and diagnose these attributes. The aim of this study is to establish a novel method for fully automated feature extraction and classification in invasive ductal carcinoma computer-aided diagnosis (CAD) systems. This article presents a tumor classification algorithm that makes novel use of convolutional neural networks on breast mammogram images to increase feature extraction and training speed. The algorithm makes two contributions. |
Using Small Proxy Datasets to Accelerate Hyperparameter Search | One of the biggest bottlenecks in a machine learning workflow is waiting for models to train. Depending on the available computing resources, it can take days to weeks to train a neural network on a large dataset with many classes such as ImageNet. For researchers experimenting with new algorithmic approaches, this is impractically time consuming and costly. We aim to generate smaller "proxy datasets" where experiments are cheaper to run but results are highly correlated with experimental results on the full dataset. We generate these proxy datasets using by randomly sampling from examples or classes, training on only the easiest or hardest examples and training on synthetic examples generated by "data distillation". We compare these techniques to the more widely used baseline of training on the full dataset for fewer epochs. For each proxying strategy, we estimate three measures of "proxy quality": how much of the variance in experimental results on the full dataset can be explained by experimental results on the proxy dataset. Experiments on Imagenette and Imagewoof (Howard, 2019) show that running hyperparameter search on the easiest 10% of examples explains 81% of the variance in experiment results on the target task, and using the easiest 50% of examples can explain 95% of the variance, significantly more than training on all the data for fewer epochs, a more widely used baseline. These "easy" proxies are higher quality than training on the full dataset for a reduced number of epochs (but equivalent computational cost), and, unexpectedly, higher quality than proxies constructed from the hardest examples. Without access to a trained model, researchers can improve proxy quality by restricting the subset to fewer classes; proxies built on half the classes are higher quality than those with an equivalent number of examples spread across all classes. |
The Potential Benefits of Filtering Versus Hyper-Parameter Optimization | The quality of an induced model by a learning algorithm is dependent on the quality of the training data and the hyper-parameters supplied to the learning algorithm. Prior work has shown that improving the quality of the training data (i.e., by removing low quality instances) or tuning the learning algorithm hyper-parameters can significantly improve the quality of an induced model. A comparison of the two methods is lacking though. In this paper, we estimate and compare the potential benefits of filtering and hyper-parameter optimization. Estimating the potential benefit gives an overly optimistic estimate but also empirically demonstrates an approximation of the maximum potential benefit of each method. We find that, while both significantly improve the induced model, improving the quality of the training set has a greater potential effect than hyper-parameter optimization. |
Pretraining Text Encoders with Adversarial Mixture of Training Signal Generators | We present a new framework AMOS that pretrains text encoders with an Adversarial learning curriculum via a Mixture Of Signals from multiple auxiliary generators. Following ELECTRA-style pretraining, the main encoder is trained as a discriminator to detect replaced tokens generated by auxiliary masked language models (MLMs). Different from ELECTRA which trains one MLM as the generator, we jointly train multiple MLMs of different sizes to provide training signals at various levels of difficulty. To push the discriminator to learn better with challenging replaced tokens, we learn mixture weights over the auxiliary MLMs' outputs to maximize the discriminator loss by backpropagating the gradient from the discriminator via Gumbel-Softmax. For better pretraining efficiency, we propose a way to assemble multiple MLMs into one unified auxiliary model. AMOS outperforms ELECTRA and recent state-of-the-art pretrained models by about 1 point on the GLUE benchmark for BERT base-sized models. |
Sensing and Mapping for Better Roads: Initial Plan for Using Federated Learning and Implementing a Digital Twin to Identify the Road Conditions in a Developing Country -- Sri Lanka | We propose how a developing country like Sri Lanka can benefit from privacy-enabled machine learning techniques such as Federated Learning to detect road conditions using crowd-sourced data collection and proposed the idea of implementing a Digital Twin for the national road system in Sri Lanka. Developing countries such as Sri Lanka are far behind in implementing smart road systems and smart cities compared to the developed countries. The proposed work discussed in this paper matches the UN Sustainable Development Goal (SDG) 9: "Build Resilient Infrastructure, Promote Inclusive and Sustainable Industrialization and Foster Innovation". Our proposed work discusses how the government and private sector vehicles that conduct routine trips to collect crowd-sourced data using smartphone devices to identify the road conditions and detect where the potholes, surface unevenness (roughness), and other major distresses are located on the roads. We explore Mobile Edge Computing (MEC) techniques that can bring machine learning intelligence closer to the edge devices where produced data is stored and show how the applications of Federated Learning can be made to detect and improve road conditions. During the second phase of this study, we plan to implement a Digital Twin for the road system in Sri Lanka. We intend to use data provided by both Dedicated and Non-Dedicated systems in the proposed Digital Twin for the road system. As of writing this paper, and best to our knowledge, there is no Digital Twin system implemented for roads and other infrastructure systems in Sri Lanka. The proposed Digital Twin will be one of the first implementations of such systems in Sri Lanka. Lessons learned from this pilot project will benefit other developing countries who wish to follow the same path and make data-driven decisions. |
Multi-Plane Program Induction with 3D Box Priors | We consider two important aspects in understanding and editing images: modeling regular, program-like texture or patterns in 2D planes, and 3D posing of these planes in the scene. Unlike prior work on image-based program synthesis, which assumes the image contains a single visible 2D plane, we present Box Program Induction (BPI), which infers a program-like scene representation that simultaneously models repeated structure on multiple 2D planes, the 3D position and orientation of the planes, and camera parameters, all from a single image. Our model assumes a box prior, i.e., that the image captures either an inner view or an outer view of a box in 3D. It uses neural networks to infer visual cues such as vanishing points, wireframe lines to guide a search-based algorithm to find the program that best explains the image. Such a holistic, structured scene representation enables 3D-aware interactive image editing operations such as inpainting missing pixels, changing camera parameters, and extrapolate the image contents. |
PrivFairFL: Privacy-Preserving Group Fairness in Federated Learning | Group fairness ensures that the outcome of machine learning (ML) based decision making systems are not biased towards a certain group of people defined by a sensitive attribute such as gender or ethnicity. Achieving group fairness in Federated Learning (FL) is challenging because mitigating bias inherently requires using the sensitive attribute values of all clients, while FL is aimed precisely at protecting privacy by not giving access to the clients' data. As we show in this paper, this conflict between fairness and privacy in FL can be resolved by combining FL with Secure Multiparty Computation (MPC) and Differential Privacy (DP). In doing so, we propose a method for training group-fair ML models in cross-device FL under complete and formal privacy guarantees, without requiring the clients to disclose their sensitive attribute values. |
Beneficial Perturbations Network for Defending Adversarial Examples | Deep neural networks can be fooled by adversarial attacks: adding carefully computed small adversarial perturbations to clean inputs can cause misclassification on state-of-the-art machine learning models. The reason is that neural networks fail to accommodate the distribution drift of the input data caused by adversarial perturbations. Here, we present a new solution - Beneficial Perturbation Network (BPN) - to defend against adversarial attacks by fixing the distribution drift. During training, BPN generates and leverages beneficial perturbations (somewhat opposite to well-known adversarial perturbations) by adding new, out-of-network biasing units. Biasing units influence the parameter space of the network, to preempt and neutralize future adversarial perturbations on input data samples. To achieve this, BPN creates reverse adversarial attacks during training, with very little cost, by recycling the training gradients already computed. Reverse attacks are captured by the biasing units, and the biases can in turn effectively defend against future adversarial examples. Reverse attacks are a shortcut, i.e., they affect the network's parameters without requiring instantiation of adversarial examples that could assist training. We provide comprehensive empirical evidence showing that 1) BPN is robust to adversarial examples and is much more running memory and computationally efficient compared to classical adversarial training. 2) BPN can defend against adversarial examples with negligible additional computation and parameter costs compared to training only on clean examples; 3) BPN hurts the accuracy on clean examples much less than classic adversarial training; 4) BPN can improve the generalization of the network 5) BPN trained only with Fast Gradient Sign Attack can generalize to defend PGD attacks. |
Verifying Inverse Model Neural Networks | Inverse problems exist in a wide variety of physical domains from aerospace engineering to medical imaging. The goal is to infer the underlying state from a set of observations. When the forward model that produced the observations is nonlinear and stochastic, solving the inverse problem is very challenging. Neural networks are an appealing solution for solving inverse problems as they can be trained from noisy data and once trained are computationally efficient to run. However, inverse model neural networks do not have guarantees of correctness built-in, which makes them unreliable for use in safety and accuracy-critical contexts. In this work we introduce a method for verifying the correctness of inverse model neural networks. Our approach is to overapproximate a nonlinear, stochastic forward model with piecewise linear constraints and encode both the overapproximate forward model and the neural network inverse model as a mixed-integer program. We demonstrate this verification procedure on a real-world airplane fuel gauge case study. The ability to verify and consequently trust inverse model neural networks allows their use in a wide variety of contexts, from aerospace to medicine. |
UFO-BLO: Unbiased First-Order Bilevel Optimization | Bilevel optimization (BLO) is a popular approach with many applications including hyperparameter optimization, neural architecture search, adversarial robustness and model-agnostic meta-learning. However, the approach suffers from time and memory complexity proportional to the length $r$ of its inner optimization loop, which has led to several modifications being proposed. One such modification is \textit{first-order} BLO (FO-BLO) which approximates outer-level gradients by zeroing out second derivative terms, yielding significant speed gains and requiring only constant memory as $r$ varies. Despite FO-BLO's popularity, there is a lack of theoretical understanding of its convergence properties. We make progress by demonstrating a rich family of examples where FO-BLO-based stochastic optimization does not converge to a stationary point of the BLO objective. We address this concern by proposing a new FO-BLO-based unbiased estimate of outer-level gradients, enabling us to theoretically guarantee this convergence, with no harm to memory and expected time complexity. Our findings are supported by experimental results on Omniglot and Mini-ImageNet, popular few-shot meta-learning benchmarks. |
Multimodal Co-learning: Challenges, Applications with Datasets, Recent Advances and Future Directions | Multimodal deep learning systems which employ multiple modalities like text, image, audio, video, etc., are showing better performance in comparison with individual modalities (i.e., unimodal) systems. Multimodal machine learning involves multiple aspects: representation, translation, alignment, fusion, and co-learning. In the current state of multimodal machine learning, the assumptions are that all modalities are present, aligned, and noiseless during training and testing time. However, in real-world tasks, typically, it is observed that one or more modalities are missing, noisy, lacking annotated data, have unreliable labels, and are scarce in training or testing and or both. This challenge is addressed by a learning paradigm called multimodal co-learning. The modeling of a (resource-poor) modality is aided by exploiting knowledge from another (resource-rich) modality using transfer of knowledge between modalities, including their representations and predictive models. Co-learning being an emerging area, there are no dedicated reviews explicitly focusing on all challenges addressed by co-learning. To that end, in this work, we provide a comprehensive survey on the emerging area of multimodal co-learning that has not been explored in its entirety yet. We review implementations that overcome one or more co-learning challenges without explicitly considering them as co-learning challenges. We present the comprehensive taxonomy of multimodal co-learning based on the challenges addressed by co-learning and associated implementations. The various techniques employed to include the latest ones are reviewed along with some of the applications and datasets. Our final goal is to discuss challenges and perspectives along with the important ideas and directions for future work that we hope to be beneficial for the entire research community focusing on this exciting domain. |
Training dataset generation for bridge game registration | This paper presents a method for automatic generation of a training dataset for a deep convolutional neural network used for playing card detection. The solution allows to skip the time-consuming processes of manual image collecting and labelling recognised objects. The YOLOv4 network trained on the generated dataset achieved an efficiency of 99.8% in the cards detection task. The proposed method is a part of a project that aims to automate the process of broadcasting duplicate bridge competitions using a vision system and neural networks. |
Electro-Magnetic Side-Channel Attack Through Learned Denoising and Classification | This paper proposes an upgraded electro-magnetic side-channel attack that automatically reconstructs the intercepted data. A novel system is introduced, running in parallel with leakage signal interception and catching compromising data in real-time. Based on deep learning and character recognition the proposed system retrieves more than 57% of characters present in intercepted signals regardless of signal type: analog or digital. The approach is also extended to a protection system that triggers an alarm if the system is compromised, demonstrating a success rate over 95%. Based on software-defined radio and graphics processing unit architectures, this solution can be easily deployed onto existing information systems where information shall be kept secret. |
Deep Multi Label Classification in Affine Subspaces | Multi-label classification (MLC) problems are becoming increasingly popular in the context of medical imaging. This has in part been driven by the fact that acquiring annotations for MLC is far less burdensome than for semantic segmentation and yet provides more expressiveness than multi-class classification. However, to train MLCs, most methods have resorted to similar objective functions as with traditional multi-class classification settings. We show in this work that such approaches are not optimal and instead propose a novel deep MLC classification method in affine subspace. At its core, the method attempts to pull features of class-labels towards different affine subspaces while maximizing the distance between them. We evaluate the method using two MLC medical imaging datasets and show a large performance increase compared to previous multi-label frameworks. This method can be seen as a plug-in replacement loss function and is trainable in an end-to-end fashion. |
Self-Trained One-class Classification for Unsupervised Anomaly Detection | Anomaly detection (AD), separating anomalies from normal data, has various applications across domains, from manufacturing to healthcare. While most previous works have shown to be effective for cases with fully or partially labeled data, they are less practical for AD applications due to tedious data labeling processes. In this work, we focus on unsupervised AD problems whose entire training data are unlabeled and may contain both normal and anomalous samples. To tackle this problem, we build a robust one-class classification framework via data refinement. To refine the data accurately, we propose an ensemble of one-class classifiers, each of which is trained on a disjoint subset of training data. Moreover, we propose a self-training of deep representation one-class classifiers (STOC) that iteratively refines the data and deep representations. In experiments, we show the efficacy of our method for unsupervised anomaly detection on benchmarks from image and tabular data domains. For example, with a 10% anomaly ratio on CIFAR-10 data, the proposed method outperforms state-of-the-art one-class classification method by 6.3 AUC and 12.5 average precision. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.