Unnamed: 0.1
int64 0
113k
| Unnamed: 0
float64 0
113k
⌀ | title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|---|---|
107,700 | 107,700 |
Open challenges for Machine Learning based Early Decision-Making
research
|
More and more applications require early decisions, i.e. taken as soon as
possible from partially observed data. However, the later a decision is made,
the more its accuracy tends to improve, since the description of the problem to
hand is enriched over time. Such a compromise between the earliness and the
accuracy of decisions has been particularly studied in the field of Early Time
Series Classification. This paper introduces a more general problem, called
Machine Learning based Early Decision Making (ML-EDM), which consists in
optimizing the decision times of models in a wide range of settings where data
is collected over time. After defining the ML-EDM problem, ten challenges are
identified and proposed to the scientific community to further research in this
area. These challenges open important application perspectives, discussed in
this paper.
|
107,701 | 107,701 |
Attention Mechanism in Neural Networks: Where it Comes and Where it Goes
|
A long time ago in the machine learning literature, the idea of incorporating
a mechanism inspired by the human visual system into neural networks was
introduced. This idea is named the attention mechanism, and it has gone through
a long development period. Today, many works have been devoted to this idea in
a variety of tasks. Remarkable performance has recently been demonstrated. The
goal of this paper is to provide an overview from the early work on searching
for ways to implement attention idea with neural networks until the recent
trends. This review emphasizes the important milestones during this progress
regarding different tasks. By this way, this study aims to provide a road map
for researchers to explore the current development and get inspired for novel
approaches beyond the attention.
|
107,702 | 107,702 |
AutoLossGen: Automatic Loss Function Generation for Recommender Systems
|
In recommendation systems, the choice of loss function is critical since a
good loss may significantly improve the model performance. However, manually
designing a good loss is a big challenge due to the complexity of the problem.
A large fraction of previous work focuses on handcrafted loss functions, which
needs significant expertise and human effort. In this paper, inspired by the
recent development of automated machine learning, we propose an automatic loss
function generation framework, AutoLossGen, which is able to generate loss
functions directly constructed from basic mathematical operators without prior
knowledge on loss structure. More specifically, we develop a controller model
driven by reinforcement learning to generate loss functions, and develop
iterative and alternating optimization schedule to update the parameters of
both the controller model and the recommender model. One challenge for
automatic loss generation in recommender systems is the extreme sparsity of
recommendation datasets, which leads to the sparse reward problem for loss
generation and search. To solve the problem, we further develop a reward
filtering mechanism for efficient and effective loss generation. Experimental
results show that our framework manages to create tailored loss functions for
different recommendation models and datasets, and the generated loss gives
better recommendation performance than commonly used baseline losses. Besides,
most of the generated losses are transferable, i.e., the loss generated based
on one model and dataset also works well for another model or dataset. Source
code of the work is available at https://github.com/rutgerswiselab/AutoLossGen.
|
107,703 | 107,703 |
R-MBO: A Multi-surrogate Approach for Preference Incorporation in
Multi-objective Bayesian Optimisation
|
Many real-world multi-objective optimisation problems rely on computationally
expensive function evaluations. Multi-objective Bayesian optimisation (BO) can
be used to alleviate the computation time to find an approximated set of Pareto
optimal solutions. In many real-world problems, a decision-maker has some
preferences on the objective functions. One approach to incorporate the
preferences in multi-objective BO is to use a scalarising function and build a
single surrogate model (mono-surrogate approach) on it. This approach has two
major limitations. Firstly, the fitness landscape of the scalarising function
and the objective functions may not be similar. Secondly, the approach assumes
that the scalarising function distribution is Gaussian, and thus a closed-form
expression of an acquisition function e.g., expected improvement can be used.
We overcome these limitations by building independent surrogate models
(multi-surrogate approach) on each objective function and show that the
distribution of the scalarising function is not Gaussian. We approximate the
distribution using Generalised value distribution. We present an a-priori
multi-surrogate approach to incorporate the desirable objective function values
(or reference point) as the preferences of a decision-maker in multi-objective
BO. The results and comparison with the existing mono-surrogate approach on
benchmark and real-world optimisation problems show the potential of the
proposed approach.
|
107,704 | 107,704 |
Learning Storm Surge with Gradient Boosting
|
Storm surge is a major natural hazard for coastal regions, responsible both
for significant property damage and loss of life. Accurate, efficient models of
storm surge are needed both to assess long-term risk and to guide emergency
management decisions. While high-fidelity ocean circulation models such as the
ADvanced CIRCulation (ADCIRC) model can accurately predict storm surge, they
are very computationally expensive. Consequently, there have been a number of
efforts in recent years to develop data-driven surrogate models for storm
surge. While these models can attain good accuracy and are highly efficient,
they are often limited to a small geographical region and a fixed set of output
locations.
We develop a novel surrogate model for peak storm surge prediction based on
gradient boosting. Unlike most surrogate approaches, our model is not
explicitly constrained to a fixed set of output locations or specific
geographical region. The model is trained with a database of 446 synthetic
storms that make landfall on the Texas coast and obtains a mean absolute error
of 0.25 meters. We additionally present a test of the model on Hurricanes Ike
(2008) and Harvey (2017).
|
107,705 | 107,705 |
FedShuffle: Recipes for Better Use of Local Work in Federated Learning
|
The practice of applying several local updates before aggregation across
clients has been empirically shown to be a successful approach to overcoming
the communication bottleneck in Federated Learning (FL). In this work, we
propose a general recipe, FedShuffle, that better utilizes the local updates in
FL, especially in the heterogeneous regime. Unlike many prior works, FedShuffle
does not assume any uniformity in the number of updates per device. Our
FedShuffle recipe comprises four simple-yet-powerful ingredients: 1) local
shuffling of the data, 2) adjustment of the local learning rates, 3) update
weighting, and 4) momentum variance reduction (Cutkosky and Orabona, 2019). We
present a comprehensive theoretical analysis of FedShuffle and show that both
theoretically and empirically, our approach does not suffer from the objective
function mismatch that is present in FL methods which assume homogeneous
updates in heterogeneous FL setups, e.g., FedAvg (McMahan et al., 2017). In
addition, by combining the ingredients above, FedShuffle improves upon FedNova
(Wang et al., 2020), which was previously proposed to solve this mismatch. We
also show that FedShuffle with momentum variance reduction can improve upon
non-local methods under a Hessian similarity assumption. Finally, through
experiments on synthetic and real-world datasets, we illustrate how each of the
four ingredients used in FedShuffle helps improve the use of local updates in
FL.
|
107,706 | 107,706 |
Minimizing Client Drift in Federated Learning via Adaptive Bias
Estimation
|
In Federated Learning (FL), a number of clients or devices collaborate to
train a model without sharing their data. Models are optimized locally at each
client and further communicated to a central hub called the server for
aggregation. While FL serves as an attractive decentralized training paradigm,
heterogeneity amongst clients' data can cause the local optimization to drift
away with respect to the global objective. In order to estimate and therefore
remove this drift, variance reduction techniques have been incorporated into FL
optimization recently. However, these approaches inaccurately estimate the
clients' drift and ultimately fail to remove it properly. In this work, we
address this challenge by introducing an adaptive algorithm that efficiently
reduces clients' drift. In this work, we propose an adaptive algorithm that
accurately estimates drift across clients. Further, unlike previous works, our
approach uses less or the same level of communication bandwidth, compute or
memory. Additionally, our proposed methodology induces stability by
constraining the norm of estimates for client drift, making it more practical
for large scale FL settings. Experimental findings demonstrate that the
proposed algorithm converges significantly faster and achieves higher accuracy
compared to the baselines across various benchmarks for FL.
|
107,707 | 107,707 |
An Adversarial Attack Analysis on Malicious Advertisement URL Detection
Framework
|
Malicious advertisement URLs pose a security risk since they are the source
of cyber-attacks, and the need to address this issue is growing in both
industry and academia. Generally, the attacker delivers an attack vector to the
user by means of an email, an advertisement link or any other means of
communication and directs them to a malicious website to steal sensitive
information and to defraud them. Existing malicious URL detection techniques
are limited and to handle unseen features as well as generalize to test data.
In this study, we extract a novel set of lexical and web-scrapped features and
employ machine learning technique to set up system for fraudulent advertisement
URLs detection. The combination set of six different kinds of features
precisely overcome the obfuscation in fraudulent URL classification. Based on
different statistical properties, we use twelve different formatted datasets
for detection, prediction and classification task. We extend our prediction
analysis for mismatched and unlabelled datasets. For this framework, we analyze
the performance of four machine learning techniques: Random Forest, Gradient
Boost, XGBoost and AdaBoost in the detection part. With our proposed method, we
can achieve a false negative rate as low as 0.0037 while maintaining high
accuracy of 99.63%. Moreover, we devise a novel unsupervised technique for data
clustering using K- Means algorithm for the visual analysis. This paper
analyses the vulnerability of decision tree-based models using the limited
knowledge attack scenario. We considered the exploratory attack and implemented
Zeroth Order Optimization adversarial attack on the detection models.
|
107,708 | 107,708 |
Interpretable Graph Convolutional Network of Multi-Modality Brain
Imaging for Alzheimer's Disease Diagnosis
|
Identification of brain regions related to the specific neurological
disorders are of great importance for biomarker and diagnostic studies. In this
paper, we propose an interpretable Graph Convolutional Network (GCN) framework
for the identification and classification of Alzheimer's disease (AD) using
multi-modality brain imaging data. Specifically, we extended the Gradient Class
Activation Mapping (Grad-CAM) technique to quantify the most discriminative
features identified by GCN from brain connectivity patterns. We then utilized
them to find signature regions of interest (ROIs) by detecting the difference
of features between regions in healthy control (HC), mild cognitive impairment
(MCI), and AD groups. We conducted the experiments on the ADNI database with
imaging data from three modalities, including VBM-MRI, FDG-PET, and AV45-PET,
and showed that the ROI features learned by our method were effective for
enhancing the performances of both clinical score prediction and disease status
identification. It also successfully identified biomarkers associated with AD
and MCI.
|
107,709 | 107,709 |
Counterfactual Explanations for Natural Language Interfaces
|
A key challenge facing natural language interfaces is enabling users to
understand the capabilities of the underlying system. We propose a novel
approach for generating explanations of a natural language interface based on
semantic parsing. We focus on counterfactual explanations, which are post-hoc
explanations that describe to the user how they could have minimally modified
their utterance to achieve their desired goal. In particular, the user provides
an utterance along with a demonstration of their desired goal; then, our
algorithm synthesizes a paraphrase of their utterance that is guaranteed to
achieve their goal. In two user studies, we demonstrate that our approach
substantially improves user performance, and that it generates explanations
that more closely match the user's intent compared to two ablations.
|
107,710 | 107,710 |
Exploring How Anomalous Model Input and Output Alerts Affect
Decision-Making in Healthcare
|
An important goal in the field of human-AI interaction is to help users more
appropriately trust AI systems' decisions. A situation in which the user may
particularly benefit from more appropriate trust is when the AI receives
anomalous input or provides anomalous output. To the best of our knowledge,
this is the first work towards understanding how anomaly alerts may contribute
to appropriate trust of AI. In a formative mixed-methods study with 4
radiologists and 4 other physicians, we explore how AI alerts for anomalous
input, very high and low confidence, and anomalous saliency-map explanations
affect users' experience with mockups of an AI clinical decision support system
(CDSS) for evaluating chest x-rays for pneumonia. We find evidence suggesting
that the four anomaly alerts are desired by non-radiologists, and the
high-confidence alerts are desired by both radiologists and non-radiologists.
In a follow-up user study, we investigate how high- and low-confidence alerts
affect the accuracy and thus appropriate trust of 33 radiologists working with
AI CDSS mockups. We observe that these alerts do not improve users' accuracy or
experience and discuss potential reasons why.
|
107,711 | 107,711 |
Use All The Labels: A Hierarchical Multi-Label Contrastive Learning
Framework
|
Current contrastive learning frameworks focus on leveraging a single
supervisory signal to learn representations, which limits the efficacy on
unseen data and downstream tasks. In this paper, we present a hierarchical
multi-label representation learning framework that can leverage all available
labels and preserve the hierarchical relationship between classes. We introduce
novel hierarchy preserving losses, which jointly apply a hierarchical penalty
to the contrastive loss, and enforce the hierarchy constraint. The loss
function is data driven and automatically adapts to arbitrary multi-label
structures. Experiments on several datasets show that our
relationship-preserving embedding performs well on a variety of tasks and
outperform the baseline supervised and self-supervised approaches. Code is
available at https://github.com/salesforce/hierarchicalContrastiveLearning.
|
107,712 | 107,712 |
ELM: Embedding and Logit Margins for Long-Tail Learning
|
Long-tail learning is the problem of learning under skewed label
distributions, which pose a challenge for standard learners. Several recent
approaches for the problem have proposed enforcing a suitable margin in logit
space. Such techniques are intuitive analogues of the guiding principle behind
SVMs, and are equally applicable to linear models and neural models. However,
when applied to neural models, such techniques do not explicitly control the
geometry of the learned embeddings. This can be potentially sub-optimal, since
embeddings for tail classes may be diffuse, resulting in poor generalization
for these classes. We present Embedding and Logit Margins (ELM), a unified
approach to enforce margins in logit space, and regularize the distribution of
embeddings. This connects losses for long-tail learning to proposals in the
literature on metric embedding, and contrastive learning. We theoretically show
that minimising the proposed ELM objective helps reduce the generalisation gap.
The ELM method is shown to perform well empirically, and results in tighter
tail class embeddings.
|
107,713 | 107,713 |
Neural network controllers for uncertain linear systems
|
We consider the design of reliable neural network (NN)-based approximations
of traditional stabilizing controllers for linear systems affected by polytopic
uncertainty, including controllers with variable structure and those based on a
minimal selection policy. We develop a systematic procedure to certify the
closed-loop stability and performance of a polytopic system when a rectified
linear unit (ReLU)-based approximation replaces such traditional controllers.
We provide sufficient conditions to ensure stability involving the worst-case
approximation error and the Lipschitz constant characterizing the error
function between ReLU-based and traditional controller-based state-to-input
mappings, and further provide offline, mixed-integer optimization-based methods
that allow us to compute those quantities exactly.
|
107,714 | 107,714 |
TransHER: Translating Knowledge Graph Embedding with Hyper-Ellipsoidal
Restriction
|
Knowledge graph embedding methods are important for knowledge graph
completion (link prediction) due to their robust performance and efficiency on
large-magnitude datasets. One state-of-the-art method, PairRE, leverages two
separate vectors for relations to model complex relations (i.e., 1-to-N,
N-to-1, and N-to-N) in knowledge graphs. However, such a method strictly
restricts entities on the hyper-ellipsoid surface and thus limits the
optimization of entity distribution, which largely hinders the performance of
knowledge graph completion. To address this problem, we propose a novel score
function TransHER, which leverages relation-specific translations between head
and tail entities restricted on separate hyper-ellipsoids. Specifically, given
a triplet, our model first maps entities onto two separate hyper-ellipsoids and
then conducts a relation-specific translation on one of them. The
relation-specific translation provides TransHER with more direct guidance in
optimization and the ability to learn semantic characteristics of entities with
complex relations. Experimental results show that TransHER can achieve
state-of-the-art performance and generalize to datasets in different domains
and scales. All our code will be publicly available.
|
107,715 | 107,715 |
Offline Visual Representation Learning for Embodied Navigation
|
How should we learn visual representations for embodied agents that must see
and move? The status quo is tabula rasa in vivo, i.e. learning visual
representations from scratch while also learning to move, potentially augmented
with auxiliary tasks (e.g. predicting the action taken between two successive
observations). In this paper, we show that an alternative 2-stage strategy is
far more effective: (1) offline pretraining of visual representations with
self-supervised learning (SSL) using large-scale pre-rendered images of indoor
environments (Omnidata), and (2) online finetuning of visuomotor
representations on specific tasks with image augmentations under long learning
schedules. We call this method Offline Visual Representation Learning (OVRL).
We conduct large-scale experiments - on 3 different 3D datasets (Gibson, HM3D,
MP3D), 2 tasks (ImageNav, ObjectNav), and 2 policy learning algorithms (RL, IL)
- and find that the OVRL representations lead to significant across-the-board
improvements in state of art, on ImageNav from 29.2% to 54.2% (+25% absolute,
86% relative) and on ObjectNav from 18.1% to 23.2% (+5.1% absolute, 28%
relative). Importantly, both results were achieved by the same visual encoder
generalizing to datasets that were not seen during pretraining. While the
benefits of pretraining sometimes diminish (or entirely disappear) with long
finetuning schedules, we find that OVRL's performance gains continue to
increase (not decrease) as the agent is trained for 2 billion frames of
experience.
|
107,716 | 107,716 |
Adversarial Fine-tune with Dynamically Regulated Adversary
|
Adversarial training is an effective method to boost model robustness to
malicious, adversarial attacks. However, such improvement in model robustness
often leads to a significant sacrifice of standard performance on clean images.
In many real-world applications such as health diagnosis and autonomous
surgical robotics, the standard performance is more valued over model
robustness against such extremely malicious attacks. This leads to the
question: To what extent we can boost model robustness without sacrificing
standard performance? This work tackles this problem and proposes a simple yet
effective transfer learning-based adversarial training strategy that
disentangles the negative effects of adversarial samples on model's standard
performance. In addition, we introduce a training-friendly adversarial attack
algorithm, which facilitates the boost of adversarial robustness without
introducing significant training complexity. Extensive experimentation
indicates that the proposed method outperforms previous adversarial training
algorithms towards the target: to improve model robustness while preserving
model's standard performance on clean data.
|
107,717 | 107,717 |
BI-GreenNet: Learning Green's functions by boundary integral network
|
Green's function plays a significant role in both theoretical analysis and
numerical computing of partial differential equations (PDEs). However, in most
cases, Green's function is difficult to compute. The troubles arise in the
following three folds. Firstly, compared with the original PDE, the dimension
of Green's function is doubled, making it impossible to be handled by
traditional mesh-based methods. Secondly, Green's function usually contains
singularities which increase the difficulty to get a good approximation.
Lastly, the computational domain may be very complex or even unbounded. To
override these problems, we leverage the fundamental solution, boundary
integral method and neural networks to develop a new method for computing
Green's function with high accuracy in this paper. We focus on Green's function
of Poisson and Helmholtz equations in bounded domains, unbounded domains. We
also consider Poisson equation and Helmholtz domains with interfaces. Extensive
numerical experiments illustrate the efficiency and the accuracy of our method
for solving Green's function. In addition, we also use the Green's function
calculated by our method to solve a class of PDE, and also obtain
high-precision solutions, which shows the good generalization ability of our
method on solving PDEs.
|
107,718 | 107,718 |
Watts: Infrastructure for Open-Ended Learning
|
This paper proposes a framework called Watts for implementing, comparing, and
recombining open-ended learning (OEL) algorithms. Motivated by modularity and
algorithmic flexibility, Watts atomizes the components of OEL systems to
promote the study of and direct comparisons between approaches. Examining
implementations of three OEL algorithms, the paper introduces the modules of
the framework. The hope is for Watts to enable benchmarking and to explore new
types of OEL algorithms. The repo is available at
\url{https://github.com/aadharna/watts}
|
107,719 | 107,719 |
Covariance-aware Feature Alignment with Pre-computed Source Statistics
for Test-time Adaptation
|
The accuracy of deep neural networks is degraded when the distribution of
features in the test environment (target domain) differs from that of the
training (source) environment. To mitigate the degradation, test-time
adaptation (TTA), where a model adapts to the target domain without access to
the source dataset, can be used in the test environment. However, the existing
TTA methods lack feature distribution alignment between the source and target
domains, which unsupervised domain adaptation mainly addresses, because
accessing the source dataset is prohibited in the TTA setting. In this paper,
we propose a novel TTA method, named Covariance-Aware Feature alignment (CAFe),
which explicitly aligns the source and target feature distributions at test
time. To perform alignment without accessing the source data, CAFe uses
auxiliary feature statistics (mean and covariance) pre-computed on the source
domain, which are lightweight and easily prepared. Further, to improve
efficiency and stability, we propose feature grouping, which splits the feature
dimensions into groups according to their correlations by using spectral
clustering to avoid degeneration of the covariance matrix. We empirically show
that CAFe outperforms prior TTA methods on a variety of distribution shifts.
|
107,720 | 107,720 |
Music Enhancement via Image Translation and Vocoding
|
Consumer-grade music recordings such as those captured by mobile devices
typically contain distortions in the form of background noise, reverb, and
microphone-induced EQ. This paper presents a deep learning approach to enhance
low-quality music recordings by combining (i) an image-to-image translation
model for manipulating audio in its mel-spectrogram representation and (ii) a
music vocoding model for mapping synthetically generated mel-spectrograms to
perceptually realistic waveforms. We find that this approach to music
enhancement outperforms baselines which use classical methods for
mel-spectrogram inversion and an end-to-end approach directly mapping noisy
waveforms to clean waveforms. Additionally, in evaluating the proposed method
with a listening test, we analyze the reliability of common audio enhancement
evaluation metrics when used in the music domain.
|
107,721 | 107,721 |
On the Normalizing Constant of the Continuous Categorical Distribution
|
Probability distributions supported on the simplex enjoy a wide range of
applications across statistics and machine learning. Recently, a novel family
of such distributions has been discovered: the continuous categorical. This
family enjoys remarkable mathematical simplicity; its density function
resembles that of the Dirichlet distribution, but with a normalizing constant
that can be written in closed form using elementary functions only. In spite of
this mathematical simplicity, our understanding of the normalizing constant
remains far from complete. In this work, we characterize the numerical behavior
of the normalizing constant and we present theoretical and methodological
advances that can, in turn, help to enable broader applications of the
continuous categorical distribution. Our code is available at
https://github.com/cunningham-lab/cb_and_cc/.
|
107,722 | 107,722 |
A Decision Model for Federated Learning Architecture Pattern Selection
|
Federated learning is growing fast in both academia and industry to resolve
data hungriness and privacy issues in machine learning. A federated learning
system being widely distributed with different components and stakeholders
requires software system design thinking. For instance, multiple patterns and
tactics have been summarised by researchers that cover various aspects, from
client management, training configuration, model deployment, etc. However, the
multitude of patterns leaves the designers confused about when and which
pattern to adopt or adapt. Therefore, in this paper, we present a set of
decision models to assist designers and architects who have limited knowledge
in federated learning, in selecting architectural patterns for federated
learning architecture design. Each decision model maps functional and
non-functional requirements of federated learning systems to a set of patterns.
we also clarify the trade-offs that may be implicit in the patterns. We
evaluated the decision model through a set of interviews with practitioners to
assess the correctness and usefulness in guiding the architecture design
process through various design decision options.
|
107,723 | 107,723 |
Model Selection, Adaptation, and Combination for Deep Transfer Learning
through Neural Networks in Renewable Energies
|
There is recent interest in using model hubs, a collection of pre-trained
models, in computer vision tasks. To utilize the model hub, we first select a
source model and then adapt the model for the target to compensate for
differences. While there is yet limited research on a model selection and
adaption for computer vision tasks, this holds even more for the field of
renewable power. At the same time, it is a crucial challenge to provide
forecasts for the increasing demand for power forecasts based on weather
features from a numerical weather prediction. We close these gaps by conducting
the first thorough experiment for model selection and adaptation for transfer
learning in renewable power forecast, adopting recent results from the field of
computer vision on six datasets. We adopt models based on data from different
seasons and limit the amount of training data. As an extension of the current
state of the art, we utilize a Bayesian linear regression for forecasting the
response based on features extracted from a neural network. This approach
outperforms the baseline with only seven days of training data. We further show
how combining multiple models through ensembles can significantly improve the
model selection and adaptation approach. In fact, with more than 30 days of
training data, both proposed model combination techniques achieve similar
results to those models trained with a full year of training data.
|
107,724 | 107,724 |
On the Convergence of Momentum-Based Algorithms for Federated Stochastic
Bilevel Optimization Problems
|
In this paper, we studied the federated stochastic bilevel optimization
problem. In particular, we developed two momentum-based algorithms for
optimizing this kind of problem. In addition, we established the convergence
rate of these two algorithms, providing their sample and communication
complexities. To the best of our knowledge, this is the first work achieving
such favorable theoretical results.
|
107,725 | 107,725 |
AlphaZero-Inspired General Board Game Learning and Playing
|
Recently, the seminal algorithms AlphaGo and AlphaZero have started a new era
in game learning and deep reinforcement learning. While the achievements of
AlphaGo and AlphaZero - playing Go and other complex games at super human level
- are truly impressive, these architectures have the drawback that they are
very complex and require high computational resources. Many researchers are
looking for methods that are similar to AlphaZero, but have lower computational
demands and are thus more easily reproducible. In this paper, we pick an
important element of AlphaZero - the Monte Carlo Tree Search (MCTS) planning
stage - and combine it with reinforcement learning (RL) agents. We wrap MCTS
for the first time around RL n-tuple networks to create versatile agents that
keep at the same time the computational demands low. We apply this new
architecture to several complex games (Othello, ConnectFour, Rubik's Cube) and
show the advantages achieved with this AlphaZero-inspired MCTS wrapper. In
particular, we present results that this AlphaZero-inspired agent is the first
one trained on standard hardware (no GPU or TPU) to beat the very strong
Othello program Edax up to and including level 7 (where most other algorithms
could only defeat Edax up to level 2).
|
107,726 | 107,726 |
Control-Aware Prediction Objectives for Autonomous Driving
|
Autonomous vehicle software is typically structured as a modular pipeline of
individual components (e.g., perception, prediction, and planning) to help
separate concerns into interpretable sub-tasks. Even when end-to-end training
is possible, each module has its own set of objectives used for safety
assurance, sample efficiency, regularization, or interpretability. However,
intermediate objectives do not always align with overall system performance.
For example, optimizing the likelihood of a trajectory prediction module might
focus more on easy-to-predict agents than safety-critical or rare behaviors
(e.g., jaywalking). In this paper, we present control-aware prediction
objectives (CAPOs), to evaluate the downstream effect of predictions on control
without requiring the planner be differentiable. We propose two types of
importance weights that weight the predictive likelihood: one using an
attention model between agents, and another based on control variation when
exchanging predicted trajectories for ground truth trajectories.
Experimentally, we show our objectives improve overall system performance in
suburban driving scenarios using the CARLA simulator.
|
107,727 | 107,727 |
Towards Flexible Inference in Sequential Decision Problems via
Bidirectional Transformers
|
Randomly masking and predicting word tokens has been a successful approach in
pre-training language models for a variety of downstream tasks. In this work,
we observe that the same idea also applies naturally to sequential decision
making, where many well-studied tasks like behavior cloning, offline RL,
inverse dynamics, and waypoint conditioning correspond to different sequence
maskings over a sequence of states, actions, and returns. We introduce the
FlexiBiT framework, which provides a unified way to specify models which can be
trained on many different sequential decision making tasks. We show that a
single FlexiBiT model is simultaneously capable of carrying out many tasks with
performance similar to or better than specialized models. Additionally, we show
that performance can be further improved by fine-tuning our general model on
specific tasks of interest.
|
107,728 | 107,728 |
Anomaly Detection by Leveraging Incomplete Anomalous Knowledge with
Anomaly-Aware Bidirectional GANs
|
The goal of anomaly detection is to identify anomalous samples from normal
ones. In this paper, a small number of anomalies are assumed to be available at
the training stage, but they are assumed to be collected only from several
anomaly types, leaving the majority of anomaly types not represented in the
collected anomaly dataset at all. To effectively leverage this kind of
incomplete anomalous knowledge represented by the collected anomalies, we
propose to learn a probability distribution that can not only model the normal
samples, but also guarantee to assign low density values for the collected
anomalies. To this end, an anomaly-aware generative adversarial network (GAN)
is developed, which, in addition to modeling the normal samples as most GANs
do, is able to explicitly avoid assigning probabilities for collected anomalous
samples. Moreover, to facilitate the computation of anomaly detection criteria
like reconstruction error, the proposed anomaly-aware GAN is designed to be
bidirectional, attaching an encoder for the generator. Extensive experimental
results demonstrate that our proposed method is able to effectively make use of
the incomplete anomalous information, leading to significant performance gains
compared to existing methods.
|
107,729 | 107,729 |
Policy Gradient Stock GAN for Realistic Discrete Order Data Generation
in Financial Markets
|
This study proposes a new generative adversarial network (GAN) for generating
realistic orders in financial markets. In some previous works, GANs for
financial markets generated fake orders in continuous spaces because of GAN
architectures' learning limitations. However, in reality, the orders are
discrete, such as order prices, which has minimum order price unit, or order
types. Thus, we change the generation method to place the generated fake orders
into discrete spaces in this study. Because this change disabled the ordinary
GAN learning algorithm, this study employed a policy gradient, frequently used
in reinforcement learning, for the learning algorithm. Through our experiments,
we show that our proposed model outperforms previous models in generated order
distribution. As an additional benefit of introducing the policy gradient, the
entropy of the generated policy can be used to check GAN's learning status. In
the future, higher performance GANs, better evaluation methods, or the
applications of our GANs can be addressed.
|
107,730 | 107,730 |
BAGNet: Bidirectional Aware Guidance Network for Malignant Breast
lesions Segmentation
|
Breast lesions segmentation is an important step of computer-aided diagnosis
system, and it has attracted much attention. However, accurate segmentation of
malignant breast lesions is a challenging task due to the effects of
heterogeneous structure and similar intensity distributions. In this paper, a
novel bidirectional aware guidance network (BAGNet) is proposed to segment the
malignant lesion from breast ultrasound images. Specifically, the bidirectional
aware guidance network is used to capture the context between global
(low-level) and local (high-level) features from the input coarse saliency map.
The introduction of the global feature map can reduce the interference of
surrounding tissue (background) on the lesion regions. To evaluate the
segmentation performance of the network, we compared with several
state-of-the-art medical image segmentation methods on the public breast
ultrasound dataset using six commonly used evaluation metrics. Extensive
experimental results indicate that our method achieves the most competitive
segmentation results on malignant breast ultrasound images.
|
107,731 | 107,731 |
Actor-Critic Scheduling for Path-Aware Air-to-Ground Multipath
Multimedia Delivery
|
Reinforcement Learning (RL) has recently found wide applications in network
traffic management and control because some of its variants do not require
prior knowledge of network models. In this paper, we present a novel scheduler
for real-time multimedia delivery in multipath systems based on an Actor-Critic
(AC) RL algorithm. We focus on a challenging scenario of real-time video
streaming from an Unmanned Aerial Vehicle (UAV) using multiple wireless paths.
The scheduler acting as an RL agent learns in real-time the optimal policy for
path selection, path rate allocation and redundancy estimation for flow
protection. The scheduler, implemented as a module of the GStreamer framework,
can be used in real or simulated settings. The simulation results show that our
scheduler can target a very low loss rate at the receiver by dynamically
adapting in real-time the scheduling policy to the path conditions without
performing training or relying on prior knowledge of network channel models.
|
107,732 | 107,732 |
Continual Learning with Bayesian Model based on a Fixed Pre-trained
Feature Extractor
|
Deep learning has shown its human-level performance in various applications.
However, current deep learning models are characterised by catastrophic
forgetting of old knowledge when learning new classes. This poses a challenge
particularly in intelligent diagnosis systems where initially only training
data of a limited number of diseases are available. In this case, updating the
intelligent system with data of new diseases would inevitably downgrade its
performance on previously learned diseases. Inspired by the process of learning
new knowledge in human brains, we propose a Bayesian generative model for
continual learning built on a fixed pre-trained feature extractor. In this
model, knowledge of each old class can be compactly represented by a collection
of statistical distributions, e.g. with Gaussian mixture models, and naturally
kept from forgetting in continual learning over time. Unlike existing
class-incremental learning methods, the proposed approach is not sensitive to
the continual learning process and can be additionally well applied to the
data-incremental learning scenario. Experiments on multiple medical and natural
image classification tasks showed that the proposed approach outperforms
state-of-the-art approaches which even keep some images of old classes during
continual learning of new classes.
|
107,733 | 107,733 |
It's DONE: Direct ONE-shot learning with Hebbian weight imprinting
|
Learning a new concept from one example is a superior function of human brain
and it is drawing attention in the field of machine learning as one-shot
learning task. In this paper, we propose the simplest method for this task with
a nonparametric weight imprinting, named Direct ONE-shot learning (DONE). DONE
adds new classes to a pretrained deep neural network (DNN) classifier with
neither training optimization nor pretrained-DNN modification. DONE is inspired
by Hebbian theory and directly uses the neural activity input of the final
dense layer obtained from a data that belongs to the new additional class as
the connectivity weight (synaptic strength) with a newly-provided-output neuron
for the new class, by transforming all statistical properties of the neural
activity into those of synaptic strength. DONE requires just one inference for
learning a new concept and its procedure is simple, deterministic, not
requiring parameter tuning and hyperparameters. The performance of DONE depends
entirely on the pretrained DNN model used as a backbone model, and we confirmed
that DONE with a well-trained backbone model performs a practical-level
accuracy. DONE has some advantages including a DNN's practical use that is
difficult to spend high cost for a training, an evaluation of existing DNN
models, and the understanding of the brain. DONE might be telling us one-shot
learning is an easy task that can be achieved by a simple principle not only
for humans but also for current well-trained DNN models.
|
107,734 | 107,734 |
Semantic Communication: An Information Bottleneck View
|
Motivated by recent success of machine learning tools at the PHY layer and
driven by high bandwidth demands of the next wireless communication standard
6G, the old idea of semantic communication by Weaver from 1949 has received
considerable attention. It breaks with the classic design paradigm according to
Shannon by aiming to transmit the meaning of a message rather than its exact
copy and thus potentially allows for savings in bandwidth.
In this work, inspired by Weaver, we propose an information-theoretic
framework where the semantic context is explicitly introduced into
probabilistic models. In particular, for bandwidth efficient transmission, we
define semantic communication system design as an Information Bottleneck
optimization problem and consider important implementation aspects. Further, we
uncover the restrictions of the classic 5G communication system design w.r.t.
semantic context. Notably, based on the example of distributed image
classification, we reveal the huge potential of a semantic communication system
design. Numerical results show a tremendous saving in bandwidth of 20 dB with
our proposed approach ISCNet compared to a classic PHY layer design.
|
107,735 | 107,735 |
Phase Shift Design in RIS Empowered Wireless Networks: From Optimization
to AI-Based Methods
|
Reconfigurable intelligent surfaces (RISs) have a revolutionary capability to
customize the radio propagation environment for wireless networks. To fully
exploit the advantages of RISs in wireless systems, the phases of the
reflecting elements must be jointly designed with conventional communication
resources, such as beamformers, transmit power, and computation time. However,
due to the unique constraints on the phase shift, and massive numbers of
reflecting units and users in large-scale networks, the resulting optimization
problems are challenging to solve. This paper provides a review of current
optimization methods and artificial intelligence-based methods for handling the
constraints imposed by RIS and compares them in terms of solution quality and
computational complexity. Future challenges in phase shift optimization
involving RISs are also described and potential solutions are discussed.
|
107,736 | 107,736 |
Machine learning for knowledge acquisition and accelerated
inverse-design for non-Hermitian systems
|
Non-Hermitian systems offer new platforms for unusual physical properties
that can be flexibly manipulated by redistribution of the real and imaginary
parts of refractive indices, whose presence breaks conventional wave
propagation symmetries, leading to asymmetric reflection and symmetric
transmission with respect to the wave propagation direction. Here, we use
supervised and unsupervised learning techniques for knowledge acquisition in
non-Hermitian systems which accelerate the inverse design process. In
particular, we construct a deep learning model that relates the transmission
and asymmetric reflection in non-conservative settings and proposes
sub-manifold learning to recognize non-Hermitian features from transmission
spectra. The developed deep learning framework determines the feasibility of a
desired spectral response for a given structure and uncovers the role of
effective gain-loss parameters to tailor the spectral response. These findings
pave the way for intelligent inverse design and shape our understanding of the
physical mechanism in general non-Hermitian systems.
|
107,737 | 107,737 |
Learning General Inventory Management Policy for Large Supply Chain
Network
|
Inventory management in warehouses directly affects profits made by
manufacturers. Particularly, large manufacturers produce a very large variety
of products that are handled by a significantly large number of retailers. In
such a case, the computational complexity of classical inventory management
algorithms is inordinately large. In recent years, learning-based approaches
have become popular for addressing such problems. However, previous studies
have not been managed systems where both the number of products and retailers
are large. This study proposes a reinforcement learning-based warehouse
inventory management algorithm that can be used for supply chain systems where
both the number of products and retailers are large. To solve the computational
problem of handling large systems, we provide a means of approximate simulation
of the system in the training phase. Our experiments on both real and
artificial data demonstrate that our algorithm with approximated simulation can
successfully handle large supply chain networks.
|
107,738 | 107,738 |
Federated Learning on Heterogeneous and Long-Tailed Data via Classifier
Re-Training with Federated Features
|
Federated learning (FL) provides a privacy-preserving solution for
distributed machine learning tasks. One challenging problem that severely
damages the performance of FL models is the co-occurrence of data heterogeneity
and long-tail distribution, which frequently appears in real FL applications.
In this paper, we reveal an intriguing fact that the biased classifier is the
primary factor leading to the poor performance of the global model. Motivated
by the above finding, we propose a novel and privacy-preserving FL method for
heterogeneous and long-tailed data via Classifier Re-training with Federated
Features (CReFF). The classifier re-trained on federated features can produce
comparable performance as the one re-trained on real data in a
privacy-preserving manner without information leakage of local data or class
distribution. Experiments on several benchmark datasets show that the proposed
CReFF is an effective solution to obtain a promising FL model under
heterogeneous and long-tailed data. Comparative results with the
state-of-the-art FL methods also validate the superiority of CReFF. Our code is
available at https://github.com/shangxinyi/CReFF-FL.
|
107,739 | 107,739 |
List-Mode PET Image Reconstruction Using Deep Image Prior
|
List-mode positron emission tomography (PET) image reconstruction is an
important tool for PET scanners with many lines-of-response (LORs) and
additional information such as time-of-flight and depth-of-interaction. Deep
learning is one possible solution to enhance the quality of PET image
reconstruction. However, the application of deep learning techniques to
list-mode PET image reconstruction have not been progressed because list data
is a sequence of bit codes and unsuitable for processing by convolutional
neural networks (CNN). In this study, we propose a novel list-mode PET image
reconstruction method using an unsupervised CNN called deep image prior (DIP)
and a framework of alternating direction method of multipliers. The proposed
list-mode DIP reconstruction (LM-DIPRecon) method alternatively iterates
regularized list-mode dynamic row action maximum likelihood algorithm
(LM-DRAMA) and magnetic resonance imaging conditioned DIP (MR-DIP). We
evaluated LM-DIPRecon using both simulation and clinical data, and it achieved
sharper images and better tradeoff curves between contrast and noise than the
LM-DRAMA and MR-DIP. These results indicated that the LM-DIPRecon is useful for
quantitative PET imaging with limited events. In addition, as list data has
finer temporal information than dynamic sinograms, list-mode deep image prior
reconstruction is expected to be useful for 4D PET imaging and motion
correction.
|
107,740 | 107,740 |
WeaNF: Weak Supervision with Normalizing Flows
|
A popular approach to decrease the need for costly manual annotation of large
data sets is weak supervision, which introduces problems of noisy labels,
coverage and bias. Methods for overcoming these problems have either relied on
discriminative models, trained with cost functions specific to weak
supervision, and more recently, generative models, trying to model the output
of the automatic annotation process. In this work, we explore a novel direction
of generative modeling for weak supervision: Instead of modeling the output of
the annotation process (the labeling function matches), we generatively model
the input-side data distributions (the feature space) covered by labeling
functions. Specifically, we estimate a density for each weak labeling source,
or labeling function, by using normalizing flows. An integral part of our
method is the flow-based modeling of multiple simultaneously matching labeling
functions, and therefore phenomena such as labeling function overlap and
correlations are captured. We analyze the effectiveness and modeling
capabilities on various commonly used weak supervision data sets, and show that
weakly supervised normalizing flows compare favorably to standard weak
supervision baselines.
|
107,741 | 107,741 |
Improving the Robustness of Federated Learning for Severely Imbalanced
Datasets
|
With the ever increasing data deluge and the success of deep neural networks,
the research of distributed deep learning has become pronounced. Two common
approaches to achieve this distributed learning is synchronous and asynchronous
weight update. In this manuscript, we have explored very simplistic synchronous
weight update mechanisms. It has been seen that with an increasing number of
worker nodes, the performance degrades drastically. This effect has been
studied in the context of extreme imbalanced classification (e.g. outlier
detection). In practical cases, the assumed conditions of i.i.d. may not be
fulfilled. There may also arise global class imbalance situations like that of
outlier detection where the local servers receive severely imbalanced data and
may not get any samples from the minority class. In that case, the DNNs in the
local servers will get completely biased towards the majority class that they
receive. This would highly impact the learning at the parameter server (which
practically does not see any data). It has been observed that in a parallel
setting if one uses the existing federated weight update mechanisms at the
parameter server, the performance degrades drastically with the increasing
number of worker nodes. This is mainly because, with the increasing number of
nodes, there is a high chance that one worker node gets a very small portion of
the data, either not enough to train the model without overfitting or having a
highly imbalanced class distribution. The chapter, hence, proposes a workaround
to this problem by introducing the concept of adaptive cost-sensitive momentum
averaging. It is seen that for the proposed system, there was no to minimal
degradation in performance while most of the other methods hit their bottom
performance before that.
|
107,742 | 107,742 |
Autoencoder based Hybrid Multi-Task Predictor Network for Daily
Open-High-Low-Close Prices Prediction of Indian Stocks
|
Stock prices are highly volatile and sudden changes in trends are often very
problematic for traditional forecasting models to handle. The standard Long
Short Term Memory (LSTM) networks are regarded as the state-of-the-art models
for such predictions. But, these models fail to handle sudden and drastic
changes in the price trend. Moreover, there are some inherent constraints with
the open, high, low and close (OHLC) prices of the stocks. Literature lacks the
study on the inherent property of OHLC prices. We argue that predicting the
OHLC prices for the next day is much more informative than predicting the
trends of the stocks as the trend is mostly calculated using these OHLC prices
only. The problem mainly is focused on Buy-Today Sell-Tomorrow (BTST) trading.
In this regard, AEs when pre-trained with the stock prices, may be beneficial.
A novel framework is proposed where a pre-trained encoder is cascaded in front
of the multi-task predictor network. This hybrid network can leverage the power
of a combination of networks and can both handle the OHLC constraints as well
as capture any sudden drastic changes in the prices. It is seen that such a
network is much more efficient at predicting stock prices. The experiments have
been extended to recommend the most profitable and most overbought stocks on
the next day. The model has been tested for multiple Indian companies and it is
found that the recommendations from the proposed model have not resulted in a
single loss for a test period of 300 days.
|
107,743 | 107,743 |
DOTIN: Dropping Task-Irrelevant Nodes for GNNs
|
Scalability is an important consideration for deep graph neural networks.
Inspired by the conventional pooling layers in CNNs, many recent graph learning
approaches have introduced the pooling strategy to reduce the size of graphs
for learning, such that the scalability and efficiency can be improved.
However, these pooling-based methods are mainly tailored to a single
graph-level task and pay more attention to local information, limiting their
performance in multi-task settings which often require task-specific global
information. In this paper, departure from these pooling-based efforts, we
design a new approach called DOTIN (\underline{D}r\underline{o}pping
\underline{T}ask-\underline{I}rrelevant \underline{N}odes) to reduce the size
of graphs. Specifically, by introducing $K$ learnable virtual nodes to
represent the graph embeddings targeted to $K$ different graph-level tasks,
respectively, up to 90\% raw nodes with low attentiveness with an attention
model -- a transformer in this paper, can be adaptively dropped without notable
performance decreasing. Achieving almost the same accuracy, our method speeds
up GAT by about 50\% on graph-level tasks including graph classification and
graph edit distance (GED) with about 60\% less memory, on D\&D dataset. Code
will be made publicly available in https://github.com/Sherrylone/DOTIN.
|
107,744 | 107,744 |
Regotron: Regularizing the Tacotron2 architecture via monotonic
alignment loss
|
Recent deep learning Text-to-Speech (TTS) systems have achieved impressive
performance by generating speech close to human parity. However, they suffer
from training stability issues as well as incorrect alignment of the
intermediate acoustic representation with the input text sequence. In this
work, we introduce Regotron, a regularized version of Tacotron2 which aims to
alleviate the training issues and at the same time produce monotonic
alignments. Our method augments the vanilla Tacotron2 objective function with
an additional term, which penalizes non-monotonic alignments in the
location-sensitive attention mechanism. By properly adjusting this
regularization term we show that the loss curves become smoother, and at the
same time Regotron consistently produces monotonic alignments in unseen
examples even at an early stage (13\% of the total number of epochs) of its
training process, whereas the fully converged Tacotron2 fails to do so.
Moreover, our proposed regularization method has no additional computational
overhead, while reducing common TTS mistakes and achieving slighlty improved
speech naturalness according to subjective mean opinion scores (MOS) collected
from 50 evaluators.
|
107,745 | 107,745 |
Cumulative Stay-time Representation for Electronic Health Records in
Medical Event Time Prediction
|
We address the problem of predicting when a disease will develop, i.e.,
medical event time (MET), from a patient's electronic health record (EHR). The
MET of non-communicable diseases like diabetes is highly correlated to
cumulative health conditions, more specifically, how much time the patient
spent with specific health conditions in the past. The common time-series
representation is indirect in extracting such information from EHR because it
focuses on detailed dependencies between values in successive observations, not
cumulative information. We propose a novel data representation for EHR called
cumulative stay-time representation (CTR), which directly models such
cumulative health conditions. We derive a trainable construction of CTR based
on neural networks that has the flexibility to fit the target data and
scalability to handle high-dimensional EHR. Numerical experiments using
synthetic and real-world datasets demonstrate that CTR alone achieves a high
prediction performance, and it enhances the performance of existing models when
combined with them.
|
107,746 | 107,746 |
Fuzzy Cognitive Maps and Hidden Markov Models: Comparative Analysis of
Efficiency within the Confines of the Time Series Classification Task
|
Time series classification is one of the very popular machine learning tasks.
In this paper, we explore the application of Hidden Markov Model (HMM) for time
series classification. We distinguish between two modes of HMM application. The
first, in which a single model is built for each class. The second, in which
one HMM is built for each time series. We then transfer both approaches for
classifier construction to the domain of Fuzzy Cognitive Maps. The identified
four models, HMM NN (HMM, one per series), HMM 1C (HMM, one per class), FCM NN,
and FCM 1C are then studied in a series of experiments. We compare the
performance of different models and investigate the impact of their
hyperparameters on the time series classification accuracy. The empirical
evaluation shows a clear advantage of the one-model-per-series approach. The
results show that the choice between HMM and FCM should be dataset-dependent.
|
107,747 | 107,747 |
COSTI: a New Classifier for Sequences of Temporal Intervals
|
Classification of sequences of temporal intervals is a part of time series
analysis which concerns series of events. We propose a new method of
transforming the problem to a task of multivariate series classification. We
use one of the state-of-the-art algorithms from the latter domain on the new
representation to obtain significantly better accuracy than the
state-of-the-art methods from the former field. We discuss limitations of this
workflow and address them by developing a novel method for classification
termed COSTI (short for Classification of Sequences of Temporal Intervals)
operating directly on sequences of temporal intervals. The proposed method
remains at a high level of accuracy and obtains better performance while
avoiding shortcomings connected to operating on transformed data. We propose a
generalized version of the problem of classification of temporal intervals,
where each event is supplemented with information about its intensity. We also
provide two new data sets where this information is of substantial value.
|
107,748 | 107,748 |
EVI: Multilingual Spoken Dialogue Tasks and Dataset for Knowledge-Based
Enrolment, Verification, and Identification
|
Knowledge-based authentication is crucial for task-oriented spoken dialogue
systems that offer personalised and privacy-focused services. Such systems
should be able to enrol (E), verify (V), and identify (I) new and recurring
users based on their personal information, e.g. postcode, name, and date of
birth. In this work, we formalise the three authentication tasks and their
evaluation protocols, and we present EVI, a challenging spoken multilingual
dataset with 5,506 dialogues in English, Polish, and French. Our proposed
models set the first competitive benchmarks, explore the challenges of
multilingual natural language processing of spoken dialogue, and set directions
for future research.
|
107,749 | 107,749 |
Unsupervised Spatial-spectral Hyperspectral Image Reconstruction and
Clustering with Diffusion Geometry
|
Hyperspectral images, which store a hundred or more spectral bands of
reflectance, have become an important data source in natural and social
sciences. Hyperspectral images are often generated in large quantities at a
relatively coarse spatial resolution. As such, unsupervised machine learning
algorithms incorporating known structure in hyperspectral imagery are needed to
analyze these images automatically. This work introduces the Spatial-Spectral
Image Reconstruction and Clustering with Diffusion Geometry (DSIRC) algorithm
for partitioning highly mixed hyperspectral images. DSIRC reduces measurement
noise through a shape-adaptive reconstruction procedure. In particular, for
each pixel, DSIRC locates spectrally correlated pixels within a data-adaptive
spatial neighborhood and reconstructs that pixel's spectral signature using
those of its neighbors. DSIRC then locates high-density, high-purity pixels far
in diffusion distance (a data-dependent distance metric) from other
high-density, high-purity pixels and treats these as cluster exemplars, giving
each a unique label. Non-modal pixels are assigned the label of their diffusion
distance-nearest neighbor of higher density and purity that is already labeled.
Strong numerical results indicate that incorporating spatial information
through image reconstruction substantially improves the performance of
pixel-wise clustering.
|
107,750 | 107,750 |
Multi-Player Multi-Armed Bandits with Finite Shareable Resources Arms:
Learning Algorithms & Applications
|
Multi-player multi-armed bandits (MMAB) study how decentralized players
cooperatively play the same multi-armed bandit so as to maximize their total
cumulative rewards. Existing MMAB models mostly assume when more than one
player pulls the same arm, they either have a collision and obtain zero
rewards, or have no collision and gain independent rewards, both of which are
usually too restrictive in practical scenarios. In this paper, we propose an
MMAB with shareable resources as an extension to the collision and
non-collision settings. Each shareable arm has finite shareable resources and a
"per-load" reward random variable, both of which are unknown to players. The
reward from a shareable arm is equal to the "per-load" reward multiplied by the
minimum between the number of players pulling the arm and the arm's maximal
shareable resources. We consider two types of feedback: sharing demand
information (SDI) and sharing demand awareness (SDA), each of which provides
different signals of resource sharing. We design the DPE-SDI and SIC-SDA
algorithms to address the shareable arm problem under these two cases of
feedback respectively and prove that both algorithms have logarithmic regrets
that are tight in the number of rounds. We conduct simulations to validate both
algorithms' performance and show their utilities in wireless networking and
edge computing.
|
107,751 | 107,751 |
On tuning a mean-field model for semi-supervised classification
|
Semi-supervised learning (SSL) has become an interesting research area due to
its capacity for learning in scenarios where both labeled and unlabeled data
are available. In this work, we focus on the task of transduction - when the
objective is to label all data presented to the learner - with a mean-field
approximation to the Potts model. Aiming at this particular task we study how
classification results depend on $\beta$ and find that the optimal phase
depends highly on the amount of labeled data available. In the same study, we
also observe that more stable classifications regarding small fluctuations in
$\beta$ are related to configurations of high probability and propose a tuning
approach based on such observation. This method relies on a novel parameter
$\gamma$ and we then evaluate two different values of the said quantity in
comparison with classical methods in the field. This evaluation is conducted by
changing the amount of labeled data available and the number of nearest
neighbors in the similarity graph. Empirical results show that the tuning
method is effective and allows NMF to outperform other approaches in datasets
with fewer classes. In addition, one of the chosen values for $\gamma$ also
leads to results that are more resilient to changes in the number of neighbors,
which might be of interest to practitioners in the field of SSL.
|
107,752 | 107,752 |
Machine Learning for Violence Risk Assessment Using Dutch Clinical Notes
|
Violence risk assessment in psychiatric institutions enables interventions to
avoid violence incidents. Clinical notes written by practitioners and available
in electronic health records are valuable resources capturing unique
information, but are seldom used to their full potential. We explore
conventional and deep machine learning methods to assess violence risk in
psychiatric patients using practitioner notes. The performance of our best
models is comparable to the currently used questionnaire-based method, with an
area under the Receiver Operating Characteristic curve of approximately 0.8. We
find that the deep-learning model BERTje performs worse than conventional
machine learning methods. We also evaluate our data and our classifiers to
understand the performance of our models better. This is particularly important
for the applicability of evaluated classifiers to new data, and is also of
great interest to practitioners, due to the increased availability of new data
in electronic format.
|
107,753 | 107,753 |
Predicting batch queue job wait times for informed scheduling of urgent
HPC workloads
|
There is increasing interest in the use of HPC machines for urgent workloads
to help tackle disasters as they unfold. Whilst batch queue systems are not
ideal in supporting such workloads, many disadvantages can be worked around by
accurately predicting when a waiting job will start to run. However there are
numerous challenges in achieving such a prediction with high accuracy, not
least because the queue's state can change rapidly and depend upon many
factors. In this work we explore a novel machine learning approach for
predicting queue wait times, hypothesising that such a model can capture the
complex behaviour resulting from the queue policy and other interactions to
generate accurate job start times.
For ARCHER2 (HPE Cray EX), Cirrus (HPE 8600) and 4-cabinet (HPE Cray EX) we
explore how different machine learning approaches and techniques improve the
accuracy of our predictions, comparing against the estimation generated by
Slurm. We demonstrate that our techniques deliver the most accurate predictions
across our machines of interest, with the result of this work being the ability
to predict job start times within one minute of the actual start time for
around 65\% of jobs on ARCHER2 and 4-cabinet, and 76\% of jobs on Cirrus. When
compared against what Slurm can deliver, this represents around 3.8 times
better accuracy on ARCHER2 and 18 times better for Cirrus. Furthermore our
approach can accurately predicting the start time for three quarters of all job
within ten minutes of the actual start time on ARCHER2 and 4-cabinet, and for
90\% of jobs on Cirrus. Whilst the driver of this work has been to better
facilitate placement of urgent workloads across HPC machines, the insights
gained can be used to provide wider benefits to users and also enrich existing
batch queue systems and inform policy too.
|
107,754 | 107,754 |
Predicting single-cell perturbation responses for unseen drugs
|
Single-cell transcriptomics enabled the study of cellular heterogeneity in
response to perturbations at the resolution of individual cells. However,
scaling high-throughput screens (HTSs) to measure cellular responses for many
drugs remains a challenge due to technical limitations and, more importantly,
the cost of such multiplexed experiments. Thus, transferring information from
routinely performed bulk RNA-seq HTS is required to enrich single-cell data
meaningfully. We introduce a new encoder-decoder architecture to study the
perturbational effects of unseen drugs. We combine the model with a transfer
learning scheme and demonstrate how training on existing bulk RNA-seq HTS
datasets can improve generalisation performance. Better generalisation reduces
the need for extensive and costly screens at single-cell resolution. We
envision that our proposed method will facilitate more efficient experiment
designs through its ability to generate in-silico hypotheses, ultimately
accelerating targeted drug discovery.
|
107,755 | 107,755 |
Mixup-based Deep Metric Learning Approaches for Incomplete Supervision
|
Deep learning architectures have achieved promising results in different
areas (e.g., medicine, agriculture, and security). However, using those
powerful techniques in many real applications becomes challenging due to the
large labeled collections required during training. Several works have pursued
solutions to overcome it by proposing strategies that can learn more for less,
e.g., weakly and semi-supervised learning approaches. As these approaches do
not usually address memorization and sensitivity to adversarial examples, this
paper presents three deep metric learning approaches combined with Mixup for
incomplete-supervision scenarios. We show that some state-of-the-art approaches
in metric learning might not work well in such scenarios. Moreover, the
proposed approaches outperform most of them in different datasets.
|
107,756 | 107,756 |
An Explainable Regression Framework for Predicting Remaining Useful Life
of Machines
|
Prediction of a machine's Remaining Useful Life (RUL) is one of the key tasks
in predictive maintenance. The task is treated as a regression problem where
Machine Learning (ML) algorithms are used to predict the RUL of machine
components. These ML algorithms are generally used as a black box with a total
focus on the performance without identifying the potential causes behind the
algorithms' decisions and their working mechanism. We believe, the performance
(in terms of Mean Squared Error (MSE), etc.,) alone is not enough to build the
trust of the stakeholders in ML prediction rather more insights on the causes
behind the predictions are needed. To this aim, in this paper, we explore the
potential of Explainable AI (XAI) techniques by proposing an explainable
regression framework for the prediction of machines' RUL. We also evaluate
several ML algorithms including classical and Neural Networks (NNs) based
solutions for the task. For the explanations, we rely on two model agnostic XAI
methods namely Local Interpretable Model-Agnostic Explanations (LIME) and
Shapley Additive Explanations (SHAP). We believe, this work will provide a
baseline for future research in the domain.
|
107,757 | 107,757 |
Prescriptive and Descriptive Approaches to Machine-Learning Transparency
|
Specialized documentation techniques have been developed to communicate key
facts about machine-learning (ML) systems and the datasets and models they rely
on. Techniques such as Datasheets, FactSheets, and Model Cards have taken a
mainly descriptive approach, providing various details about the system
components. While the above information is essential for product developers and
external experts to assess whether the ML system meets their requirements,
other stakeholders might find it less actionable. In particular, ML engineers
need guidance on how to mitigate potential shortcomings in order to fix bugs or
improve the system's performance. We survey approaches that aim to provide such
guidance in a prescriptive way. We further propose a preliminary approach,
called Method Cards, which aims to increase the transparency and
reproducibility of ML systems by providing prescriptive documentation of
commonly-used ML methods and techniques. We showcase our proposal with an
example in small object detection, and demonstrate how Method Cards can
communicate key considerations for model developers. We further highlight
avenues for improving the user experience of ML engineers based on Method
Cards.
|
107,758 | 107,758 |
Predicting Sleeping Quality using Convolutional Neural Networks
|
Identifying sleep stages and patterns is an essential part of diagnosing and
treating sleep disorders. With the advancement of smart technologies, sensor
data related to sleeping patterns can be captured easily. In this paper, we
propose a Convolution Neural Network (CNN) architecture that improves the
classification performance. In particular, we benchmark the classification
performance from different methods, including traditional machine learning
methods such as Logistic Regression (LR), Decision Trees (DT), k-Nearest
Neighbour (k-NN), Naive Bayes (NB) and Support Vector Machine (SVM), on 3
publicly available sleep datasets. The accuracy, sensitivity, specificity,
precision, recall, and F-score are reported and will serve as a baseline to
simulate the research in this direction in the future.
|
107,759 | 107,759 |
Nonbacktracking spectral clustering of nonuniform hypergraphs
|
Spectral methods offer a tractable, global framework for clustering in graphs
via eigenvector computations on graph matrices. Hypergraph data, in which
entities interact on edges of arbitrary size, poses challenges for matrix
representations and therefore for spectral clustering. We study spectral
clustering for nonuniform hypergraphs based on the hypergraph nonbacktracking
operator. After reviewing the definition of this operator and its basic
properties, we prove a theorem of Ihara-Bass type to enable faster computation
of eigenpairs. We then propose an alternating algorithm for inference in a
hypergraph stochastic blockmodel via linearized belief-propagation, offering
proofs that both formalize and extend several previous results. We perform
experiments in real and synthetic data that underscore the benefits of
hypergraph methods over graph-based ones when interactions of different sizes
carry different information about cluster structure. Through an analysis of our
algorithm, we pose several conjectures about the limits of spectral methods and
detectability in hypergraph stochastic blockmodels writ large.
|
107,760 | 107,760 |
Supervised machine learning classification for short straddles on the
S&P500
|
In this working paper we present our current progress in the training of
machine learning models to execute short option strategies on the S&P500. As a
first step, this paper is breaking this problem down to a supervised
classification task to decide if a short straddle on the S&P500 should be
executed or not on a daily basis. We describe our used framework and present an
overview over our evaluation metrics on different classification models. In
this preliminary work, using standard machine learning techniques and without
hyperparameter search, we find no statistically significant outperformance to a
simple "trade always" strategy, but gain additional insights on how we could
proceed in further experiments.
|
107,761 | 107,761 |
Computer Vision for Road Imaging and Pothole Detection: A
State-of-the-Art Review of Systems and Algorithms
|
Computer vision algorithms have been prevalently utilized for 3-D road
imaging and pothole detection for over two decades. Nonetheless, there is a
lack of systematic survey articles on state-of-the-art (SoTA) computer vision
techniques, especially deep learning models, developed to tackle these
problems. This article first introduces the sensing systems employed for 2-D
and 3-D road data acquisition, including camera(s), laser scanners, and
Microsoft Kinect. Afterward, it thoroughly and comprehensively reviews the SoTA
computer vision algorithms, including (1) classical 2-D image processing, (2)
3-D point cloud modeling and segmentation, and (3) machine/deep learning,
developed for road pothole detection. This article also discusses the existing
challenges and future development trends of computer vision-based road pothole
detection approaches: classical 2-D image processing-based and 3-D point cloud
modeling and segmentation-based approaches have already become history; and
Convolutional neural networks (CNNs) have demonstrated compelling road pothole
detection results and are promising to break the bottleneck with the future
advances in self/un-supervised learning for multi-modal semantic segmentation.
We believe that this survey can serve as practical guidance for developing the
next-generation road condition assessment systems.
|
107,762 | 107,762 |
Continual Learning for Peer-to-Peer Federated Learning: A Study on
Automated Brain Metastasis Identification
|
Due to data privacy constraints, data sharing among multiple centers is
restricted. Continual learning, as one approach to peer-to-peer federated
learning, can promote multicenter collaboration on deep learning algorithm
development by sharing intermediate models instead of training data. This work
aims to investigate the feasibility of continual learning for multicenter
collaboration on an exemplary application of brain metastasis identification
using DeepMedic. 920 T1 MRI contrast enhanced volumes are split to simulate
multicenter collaboration scenarios. A continual learning algorithm, synaptic
intelligence (SI), is applied to preserve important model weights for training
one center after another. In a bilateral collaboration scenario, continual
learning with SI achieves a sensitivity of 0.917, and naive continual learning
without SI achieves a sensitivity of 0.906, while two models trained on
internal data solely without continual learning achieve sensitivity of 0.853
and 0.831 only. In a seven-center multilateral collaboration scenario, the
models trained on internal datasets (100 volumes each center) without continual
learning obtain a mean sensitivity value of 0.699. With single-visit continual
learning (i.e., the shared model visits each center only once during training),
the sensitivity is improved to 0.788 and 0.849 without SI and with SI,
respectively. With iterative continual learning (i.e., the shared model
revisits each center multiple times during training), the sensitivity is
further improved to 0.914, which is identical to the sensitivity using mixed
data for training. Our experiments demonstrate that continual learning can
improve brain metastasis identification performance for centers with limited
data. This study demonstrates the feasibility of applying continual learning
for peer-to-peer federated learning in multicenter collaboration.
|
107,763 | 107,763 |
Poisoning Deep Learning Based Recommender Model in Federated Learning
Scenarios
|
Various attack methods against recommender systems have been proposed in the
past years, and the security issues of recommender systems have drawn
considerable attention. Traditional attacks attempt to make target items
recommended to as many users as possible by poisoning the training data.
Benifiting from the feature of protecting users' private data, federated
recommendation can effectively defend such attacks. Therefore, quite a few
works have devoted themselves to developing federated recommender systems. For
proving current federated recommendation is still vulnerable, in this work we
probe to design attack approaches targeting deep learning based recommender
models in federated learning scenarios. Specifically, our attacks generate
poisoned gradients for manipulated malicious users to upload based on two
strategies (i.e., random approximation and hard user mining). Extensive
experiments show that our well-designed attacks can effectively poison the
target models, and the attack effectiveness sets the state-of-the-art.
|
107,764 | 107,764 |
PhysioGAN: Training High Fidelity Generative Model for Physiological
Sensor Readings
|
Generative models such as the variational autoencoder (VAE) and the
generative adversarial networks (GAN) have proven to be incredibly powerful for
the generation of synthetic data that preserves statistical properties and
utility of real-world datasets, especially in the context of image and natural
language text. Nevertheless, until now, there has no successful demonstration
of how to apply either method for generating useful physiological sensory data.
The state-of-the-art techniques in this context have achieved only limited
success. We present PHYSIOGAN, a generative model to produce high fidelity
synthetic physiological sensor data readings. PHYSIOGAN consists of an encoder,
decoder, and a discriminator. We evaluate PHYSIOGAN against the
state-of-the-art techniques using two different real-world datasets: ECG
classification and activity recognition from motion sensors datasets. We
compare PHYSIOGAN to the baseline models not only the accuracy of class
conditional generation but also the sample diversity and sample novelty of the
synthetic datasets. We prove that PHYSIOGAN generates samples with higher
utility than other generative models by showing that classification models
trained on only synthetic data generated by PHYSIOGAN have only 10% and 20%
decrease in their classification accuracy relative to classification models
trained on the real data. Furthermore, we demonstrate the use of PHYSIOGAN for
sensor data imputation in creating plausible results.
|
107,765 | 107,765 |
A unified theory of information transfer and causal relation
|
Information transfer between coupled stochastic dynamics, measured by
transfer entropy and information flow, is suggested as a physical process
underlying the causal relation of systems. While information transfer analysis
has booming applications in both science and engineering fields, critical
mysteries about its foundations remain unsolved. Fundamental yet difficult
questions concern how information transfer and causal relation originate, what
they depend on, how they differ from each other, and if they are created by a
unified and general quantity. These questions essentially determine the
validity of causal relation measurement via information transfer. Here we
pursue to lay a complete theoretical basis of information transfer and causal
relation. Beyond the well-known relations between these concepts that
conditionally hold, we demonstrate that information transfer and causal
relation universally originate from specific information synergy and redundancy
phenomena characterized by high-order mutual information. More importantly, our
theory analytically explains the mechanisms for information transfer and causal
relation to originate, vanish, and differ from each other. Moreover, our theory
naturally defines the effect sizes of information transfer and causal relation
based on high-dimensional coupling events. These results may provide a unified
view of information, synergy, and causal relation to bridge Pearl's causal
inference theory in computer science and information transfer analysis in
physics.
|
107,766 | 107,766 |
Signal Recovery with Non-Expansive Generative Network Priors
|
We study compressive sensing with a deep generative network prior. Initial
theoretical guarantees for efficient recovery from compressed linear
measurements have been developed for signals in the range of a ReLU network
with Gaussian weights and logarithmic expansivity: that is when each layer is
larger than the previous one by a logarithmic factor. It was later shown that
constant expansivity is sufficient for recovery. It has remained open whether
the expansivity can be relaxed, allowing for networks with contractive layers
(as often the case of real generators). In this work we answer this question,
proving that a signal in the range of a Gaussian generative network can be
recovered from few linear measurements provided that the width of the layers is
proportional to the input layer size (up to log factors). This condition allows
the generative network to have contractive layers. Our result is based on
showing that Gaussian matrices satisfy a matrix concentration inequality which
we term Range Restricted Weight Distribution Condition (R2WDC) and that weakens
the Weight Distribution Condition (WDC) upon which previous theoretical
guarantees were based. The WDC has also been used to analyze other signal
recovery problems with generative network priors. By replacing the WDC with the
R2WDC, we are able to extend previous results for signal recovery with
expansive generative network priors to non-expansive ones. We discuss these
extensions for phase retrieval, denoising, and spiked matrix recovery.
|
107,767 | 107,767 |
Bona fide Riesz projections for density estimation
|
The projection of sample measurements onto a reconstruction space represented
by a basis on a regular grid is a powerful and simple approach to estimate a
probability density function. In this paper, we focus on Riesz bases and
propose a projection operator that, in contrast to previous works, guarantees
the bona fide properties for the estimate, namely, non-negativity and total
probability mass $1$. Our bona fide projection is defined as a convex problem.
We propose solution techniques and evaluate them. Results suggest an improved
performance, specifically in circumstances prone to rippling effects.
|
107,768 | 107,768 |
Process-BERT: A Framework for Representation Learning on Educational
Process Data
|
Educational process data, i.e., logs of detailed student activities in
computerized or online learning platforms, has the potential to offer deep
insights into how students learn. One can use process data for many downstream
tasks such as learning outcome prediction and automatically delivering
personalized intervention. However, analyzing process data is challenging since
the specific format of process data varies a lot depending on different
learning/testing scenarios. In this paper, we propose a framework for learning
representations of educational process data that is applicable across many
different learning scenarios. Our framework consists of a pre-training step
that uses BERT-type objectives to learn representations from sequential process
data and a fine-tuning step that further adjusts these representations on
downstream prediction tasks. We apply our framework to the 2019 nation's report
card data mining competition dataset that consists of student problem-solving
process data and detail the specific models we use in this scenario. We conduct
both quantitative and qualitative experiments to show that our framework
results in process data representations that are both predictive and
informative.
|
107,769 | 107,769 |
Representative period selection for power system planning using
autoencoder-based dimensionality reduction
|
Power sector capacity expansion models (CEMs) that are used for studying
future low-carbon grid scenarios must incorporate detailed representation of
grid operations. Often CEMs are formulated to model grid operations over
representative periods that are sampled from the original input data using
clustering algorithms. However, such representative period selection (RPS)
methods are limited by the declining efficacy of the clustering algorithm with
increasing dimensionality of the input data and do not consider the relative
importance of input data variations on CEM outcomes. Here, we propose a RPS
method that addresses these limitations by incorporating dimensionality
reduction, accomplished via neural network based autoencoders, prior to
clustering. Such dimensionality reduction not only improves the performance of
the clustering algorithm, but also facilitates using additional features, such
as estimated outputs produced from parallel solutions of simplified versions of
the CEM for each disjoint period in the input data (e.g. 1 week). The impact of
incorporating dimensionality reduction as part of RPS methods is quantified
through the error in outcomes of the corresponding reduced-space CEM vs. the
full space CEM. Extensive numerical experimentation across various networks and
range of technology and policy scenarios establish the superiority of the
dimensionality-reduction based RPS methods.
|
107,770 | 107,770 |
Personalized Federated Learning with Multiple Known Clusters
|
We consider the problem of personalized federated learning when there are
known cluster structures within users. An intuitive approach would be to
regularize the parameters so that users in the same cluster share similar model
weights. The distances between the clusters can then be regularized to reflect
the similarity between different clusters of users. We develop an algorithm
that allows each cluster to communicate independently and derive the
convergence results. We study a hierarchical linear model to theoretically
demonstrate that our approach outperforms agents learning independently and
agents learning a single shared weight. Finally, we demonstrate the advantages
of our approach using both simulated and real-world data.
|
107,771 | 107,771 |
Standardized Evaluation of Machine Learning Methods for Evolving Data
Streams
|
Due to the unspecified and dynamic nature of data streams, online machine
learning requires powerful and flexible solutions. However, evaluating online
machine learning methods under realistic conditions is difficult. Existing work
therefore often draws on different heuristics and simulations that do not
necessarily produce meaningful and reliable results. Indeed, in the absence of
common evaluation standards, it often remains unclear how online learning
methods will perform in practice or in comparison to similar work. In this
paper, we propose a comprehensive set of properties for high-quality machine
learning in evolving data streams. In particular, we discuss sensible
performance measures and evaluation strategies for online predictive modelling,
online feature selection and concept drift detection. As one of the first
works, we also look at the interpretability of online learning methods. The
proposed evaluation standards are provided in a new Python framework called
float. Float is completely modular and allows the simultaneous integration of
common libraries, such as scikit-multiflow or river, with custom code. Float is
open-sourced and can be accessed at https://github.com/haugjo/float. In this
sense, we hope that our work will contribute to more standardized, reliable and
realistic testing and comparison of online machine learning methods.
|
107,772 | 107,772 |
Russian Texts Detoxification with Levenshtein Editing
|
Text detoxification is a style transfer task of creating neutral versions of
toxic texts. In this paper, we use the concept of text editing to build a
two-step tagging-based detoxification model using a parallel corpus of Russian
texts. With this model, we achieved the best style transfer accuracy among all
models in the RUSSE Detox shared task, surpassing larger sequence-to-sequence
models.
|
107,773 | 107,773 |
Unlocking High-Accuracy Differentially Private Image Classification
through Scale
|
Differential Privacy (DP) provides a formal privacy guarantee preventing
adversaries with access to a machine learning model from extracting information
about individual training points. Differentially Private Stochastic Gradient
Descent (DP-SGD), the most popular DP training method for deep learning,
realizes this protection by injecting noise during training. However previous
works have found that DP-SGD often leads to a significant degradation in
performance on standard image classification benchmarks. Furthermore, some
authors have postulated that DP-SGD inherently performs poorly on large models,
since the norm of the noise required to preserve privacy is proportional to the
model dimension. In contrast, we demonstrate that DP-SGD on over-parameterized
models can perform significantly better than previously thought. Combining
careful hyper-parameter tuning with simple techniques to ensure signal
propagation and improve the convergence rate, we obtain a new SOTA without
extra data on CIFAR-10 of 81.4% under (8, 10^{-5})-DP using a 40-layer
Wide-ResNet, improving over the previous SOTA of 71.7%. When fine-tuning a
pre-trained NFNet-F3, we achieve a remarkable 83.8% top-1 accuracy on ImageNet
under (0.5, 8*10^{-7})-DP. Additionally, we also achieve 86.7% top-1 accuracy
under (8, 8 \cdot 10^{-7})-DP, which is just 4.3% below the current non-private
SOTA for this task. We believe our results are a significant step towards
closing the accuracy gap between private and non-private image classification.
|
107,774 | 107,774 |
Toward Compositional Generalization in Object-Oriented World Modeling
|
Compositional generalization is a critical ability in learning and
decision-making. We focus on the setting of reinforcement learning in
object-oriented environments to study compositional generalization in world
modeling. We (1) formalize the compositional generalization problem with an
algebraic approach and (2) study how a world model can achieve that. We
introduce a conceptual environment, Object Library, and two instances, and
deploy a principled pipeline to measure the generalization ability. Motivated
by the formulation, we analyze several methods with exact} or no compositional
generalization ability using our framework, and design a differentiable
approach, Homomorphic Object-oriented World Model (HOWM), that achieves
approximate but more efficient compositional generalization.
|
107,775 | 107,775 |
Schr\"odinger's FP: Dynamic Adaptation of Floating-Point Containers for
Deep Learning Training
|
We introduce a software-hardware co-design approach to reduce memory traffic
and footprint during training with BFloat16 or FP32 boosting energy efficiency
and execution time performance. We introduce methods to dynamically adjust the
size and format of the floating-point containers used to store activations and
weights during training. The different value distributions lead us to different
approaches for exponents and mantissas. Gecko exploits the favourable exponent
distribution with a loss-less delta encoding approach to reduce the total
exponent footprint by up to $58\%$ in comparison to a 32 bit floating point
baseline. To content with the noisy mantissa distributions, we present two
lossy methods to eliminate as many as possible least significant bits while not
affecting accuracy. Quantum Mantissa, is a machine learning-first mantissa
compression method that taps on training's gradient descent algorithm to also
learn minimal mantissa bitlengths on a per-layer granularity, and obtain up to
$92\%$ reduction in total mantissa footprint. Alternatively, BitChop observes
changes in the loss function during training to adjust mantissa bit-length
network-wide yielding a reduction of $81\%$ in footprint. Schr\"{o}dinger's FP
implements hardware encoders/decoders that guided by Gecko/Quantum Mantissa or
Gecko/BitChop transparently encode/decode values when transferring to/from
off-chip memory boosting energy efficiency and reducing execution time.
|
107,776 | 107,776 |
Unaligned Supervision For Automatic Music Transcription in The Wild
|
Multi-instrument Automatic Music Transcription (AMT), or the decoding of a
musical recording into semantic musical content, is one of the holy grails of
Music Information Retrieval. Current AMT approaches are restricted to piano and
(some) guitar recordings, due to difficult data collection. In order to
overcome data collection barriers, previous AMT approaches attempt to employ
musical scores in the form of a digitized version of the same song or piece.
The scores are typically aligned using audio features and strenuous human
intervention to generate training labels. We introduce NoteEM, a method for
simultaneously training a transcriber and aligning the scores to their
corresponding performances, in a fully-automated process. Using this unaligned
supervision scheme, complemented by pseudo-labels and pitch-shift augmentation,
our method enables training on in-the-wild recordings with unprecedented
accuracy and instrumental variety. Using only synthetic data and unaligned
supervision, we report SOTA note-level accuracy of the MAPS dataset, and large
favorable margins on cross-dataset evaluations. We also demonstrate robustness
and ease of use; we report comparable results when training on a small, easily
obtainable, self-collected dataset, and we propose alternative labeling to the
MusicNet dataset, which we show to be more accurate. Our project page is
available at https://benadar293.github.io
|
107,777 | 107,777 |
Unified Simulation, Perception, and Generation of Human Behavior
|
Understanding and modeling human behavior is fundamental to almost any
computer vision and robotics applications that involve humans. In this thesis,
we take a holistic approach to human behavior modeling and tackle its three
essential aspects -- simulation, perception, and generation. Throughout the
thesis, we show how the three aspects are deeply connected and how utilizing
and improving one aspect can greatly benefit the other aspects. We also discuss
the lessons learned and our vision for what is next for human behavior
modeling.
|
107,778 | 107,778 |
Curriculum Learning for Dense Retrieval Distillation
|
Recent work has shown that more effective dense retrieval models can be
obtained by distilling ranking knowledge from an existing base re-ranking
model. In this paper, we propose a generic curriculum learning based
optimization framework called CL-DRD that controls the difficulty level of
training data produced by the re-ranking (teacher) model. CL-DRD iteratively
optimizes the dense retrieval (student) model by increasing the difficulty of
the knowledge distillation data made available to it. In more detail, we
initially provide the student model coarse-grained preference pairs between
documents in the teacher's ranking and progressively move towards finer-grained
pairwise document ordering requirements. In our experiments, we apply a simple
implementation of the CL-DRD framework to enhance two state-of-the-art dense
retrieval models. Experiments on three public passage retrieval datasets
demonstrate the effectiveness of our proposed framework.
|
107,779 | 107,779 |
KING: Generating Safety-Critical Driving Scenarios for Robust Imitation
via Kinematics Gradients
|
Simulators offer the possibility of safe, low-cost development of
self-driving systems. However, current driving simulators exhibit na\"ive
behavior models for background traffic. Hand-tuned scenarios are typically
added during simulation to induce safety-critical situations. An alternative
approach is to adversarially perturb the background traffic trajectories. In
this paper, we study this approach to safety-critical driving scenario
generation using the CARLA simulator. We use a kinematic bicycle model as a
proxy to the simulator's true dynamics and observe that gradients through this
proxy model are sufficient for optimizing the background traffic trajectories.
Based on this finding, we propose KING, which generates safety-critical driving
scenarios with a 20% higher success rate than black-box optimization. By
solving the scenarios generated by KING using a privileged rule-based expert
algorithm, we obtain training data for an imitation learning policy. After
fine-tuning on this new data, we show that the policy becomes better at
avoiding collisions. Importantly, our generated data leads to reduced
collisions on both held-out scenarios generated via KING as well as traditional
hand-crafted scenarios, demonstrating improved robustness.
|
107,780 | 107,780 |
Foundations for learning from noisy quantum experiments
|
Understanding what can be learned from experiments is central to scientific
progress. In this work, we use a learning-theoretic perspective to study the
task of learning physical operations in a quantum machine when all operations
(state preparation, dynamics, and measurement) are a priori unknown. We prove
that, without any prior knowledge, if one can explore the full quantum state
space by composing the operations, then every operation can be learned. When
one cannot explore the full state space but all operations are approximately
known and noise in Clifford gates is gate-independent, we find an efficient
algorithm for learning all operations up to a single unlearnable parameter
characterizing the fidelity of the initial state. For learning a noise channel
on Clifford gates to a fixed accuracy, our algorithm uses quadratically fewer
experiments than previously known protocols. Under more general conditions, the
true description of the noise can be unlearnable; for example, we prove that no
benchmarking protocol can learn gate-dependent Pauli noise on Clifford+T gates
even under perfect state preparation and measurement. Despite not being able to
learn the noise, we show that a noisy quantum computer that performs entangled
measurements on multiple copies of an unknown state can yield a large advantage
in learning properties of the state compared to a noiseless device that
measures individual copies and then processes the measurement data using a
classical computer. Concretely, we prove that noisy quantum computers with
two-qubit gate error rate $\epsilon$ can achieve a learning task using $N$
copies of the state, while $N^{\Omega(1/\epsilon)}$ copies are required
classically.
|
107,781 | 107,781 |
Bilinear value networks
|
The dominant framework for off-policy multi-goal reinforcement learning
involves estimating goal conditioned Q-value function. When learning to achieve
multiple goals, data efficiency is intimately connected with the generalization
of the Q-function to new goals. The de-facto paradigm is to approximate Q(s, a,
g) using monolithic neural networks. To improve the generalization of the
Q-function, we propose a bilinear decomposition that represents the Q-value via
a low-rank approximation in the form of a dot product between two vector
fields. The first vector field, f(s, a), captures the environment's local
dynamics at the state s; whereas the second component, {\phi}(s, g), captures
the global relationship between the current state and the goal. We show that
our bilinear decomposition scheme substantially improves data efficiency, and
has superior transfer to out-of-distribution goals compared to prior methods.
Empirical evidence is provided on the simulated Fetch robot task-suite and
dexterous manipulation with a Shadow hand.
|
107,782 | 107,782 |
Federated Learning: Balancing the Thin Line Between Data Intelligence
and Privacy
|
Federated learning holds great promise in learning from fragmented sensitive
data and has revolutionized how machine learning models are trained. This
article provides a systematic overview and detailed taxonomy of federated
learning. We investigate the existing security challenges in federated learning
and provide a comprehensive overview of established defense techniques for data
poisoning, inference attacks, and model poisoning attacks. The work also
presents an overview of current training challenges for federated learning,
focusing on handling non-i.i.d. data, high dimensionality issues, and
heterogeneous architecture, and discusses several solutions for the associated
challenges. Finally, we discuss the remaining challenges in managing federated
learning training and suggest focused research directions to address the open
questions. Potential candidate areas for federated learning, including IoT
ecosystem, healthcare applications, are discussed with a particular focus on
banking and financial domains.
|
107,783 | 107,783 |
Identifying Critical LMS Features for Predicting At-risk Students
|
Learning management systems (LMSs) have become essential in higher education
and play an important role in helping educational institutions to promote
student success. Traditionally, LMSs have been used by postsecondary
institutions in administration, reporting, and delivery of educational content.
In this paper, we present an additional use of LMS by using its data logs to
perform data-analytics and identify academically at-risk students. The
data-driven insights would allow educational institutions and educators to
develop and implement pedagogical interventions targeting academically at-risk
students. We used anonymized data logs created by Brightspace LMS during fall
2019, spring 2020, and fall 2020 semesters at our college. Supervised machine
learning algorithms were used to predict the final course performance of
students, and several algorithms were found to perform well with accuracy above
90%. SHAP value method was used to assess the relative importance of features
used in the predictive models. Unsupervised learning was also used to group
students into different clusters based on the similarities in their
interaction/involvement with LMS. In both of supervised and unsupervised
learning, we identified two most-important features
(Number_Of_Assignment_Submissions and Content_Completed). More importantly, our
study lays a foundation and provides a framework for developing a real-time
data analytics metric that may be incorporated into a LMS.
|
107,784 | 107,784 |
Neighbor-Based Optimized Logistic Regression Machine Learning Model For
Electric Vehicle Occupancy Detection
|
This paper presents an optimized logistic regression machine learning model
that predicts the occupancy of an Electric Vehicle (EV) charging station given
the occupancy of neighboring stations. The model was optimized for the time of
day. Trained on data from 57 EV charging stations around the University of
California San Diego campus, the model achieved an 88.43% average accuracy and
92.23% maximum accuracy in predicting occupancy, outperforming a persistence
model benchmark.
|
107,785 | 107,785 |
Hyperbolic Hierarchical Knowledge Graph Embeddings for Link Prediction
in Low Dimensions
|
Knowledge graph embeddings (KGE) have been validated as powerful methods for
inferring missing links in knowledge graphs (KGs) since they map entities into
Euclidean space and treat relations as transformations of entities. Currently,
some Euclidean KGE methods model semantic hierarchies prevalent in KGs and
promote the performance of link prediction. For hierarchical data, instead of
traditional Euclidean space, hyperbolic space as an embedding space has shown
the promise of high fidelity and low memory consumption; however, existing
hyperbolic KGE methods neglect to model them. To address this issue, we propose
a novel KGE model -- hyperbolic hierarchical KGE (HypHKGE). To be specific, we
first design the attention-based learnable curvatures for hyperbolic space to
preserve rich semantic hierarchies. Moreover, we define the hyperbolic
hierarchical transformations based on the theory of hyperbolic geometry, which
utilize hierarchies that we preserved to infer the links. Experiments show that
HypHKGE can effectively model semantic hierarchies in hyperbolic space and
outperforms the state-of-the-art hyperbolic methods, especially in low
dimensions.
|
107,786 | 107,786 |
Coupling Deep Imputation with Multitask Learning for Downstream Tasks on
Genomics Data
|
Genomics data such as RNA gene expression, methylation and micro RNA
expression are valuable sources of information for various clinical predictive
tasks. For example, predicting survival outcomes, cancer histology type and
other patients' related information is possible using not only clinical data
but molecular data as well. Moreover, using these data sources together, for
example in multitask learning, can boost the performance. However, in practice,
there are many missing data points which leads to significantly lower patient
numbers when analysing full cases, which in our setting refers to all
modalities being present.
In this paper we investigate how imputing data with missing values using deep
learning coupled with multitask learning can help to reach state-of-the-art
performance results using combined genomics modalities, RNA, micro RNA and
methylation. We propose a generalised deep imputation method to impute values
where a patient has all modalities present except one. Interestingly enough,
deep imputation alone outperforms multitask learning alone for the
classification and regression tasks across most combinations of modalities. In
contrast, when using all modalities for survival prediction we observe that
multitask learning alone outperforms deep imputation alone with statistical
significance (adjusted p-value 0.03). Thus, both approaches are complementary
when optimising performance for downstream predictive tasks.
|
107,787 | 107,787 |
Tag-assisted Multimodal Sentiment Analysis under Uncertain Missing
Modalities
|
Multimodal sentiment analysis has been studied under the assumption that all
modalities are available. However, such a strong assumption does not always
hold in practice, and most of multimodal fusion models may fail when partial
modalities are missing. Several works have addressed the missing modality
problem; but most of them only considered the single modality missing case, and
ignored the practically more general cases of multiple modalities missing. To
this end, in this paper, we propose a Tag-Assisted Transformer Encoder (TATE)
network to handle the problem of missing uncertain modalities. Specifically, we
design a tag encoding module to cover both the single modality and multiple
modalities missing cases, so as to guide the network's attention to those
missing modalities. Besides, we adopt a new space projection pattern to align
common vectors. Then, a Transformer encoder-decoder network is utilized to
learn the missing modality features. At last, the outputs of the Transformer
encoder are used for the final sentiment classification. Extensive experiments
are conducted on CMU-MOSI and IEMOCAP datasets, showing that our method can
achieve significant improvements compared with several baselines.
|
107,788 | 107,788 |
Learning cosmology and clustering with cosmic graphs
|
We train deep learning models on thousands of galaxy catalogues from the
state-of-the-art hydrodynamic simulations of the CAMELS project to perform
regression and inference. We employ Graph Neural Networks (GNNs), architectures
designed to work with irregular and sparse data, like the distribution of
galaxies in the Universe. We first show that GNNs can learn to compute the
power spectrum of galaxy catalogues with a few percent accuracy. We then train
GNNs to perform likelihood-free inference at the galaxy-field level. Our models
are able to infer the value of $\Omega_{\rm m}$ with a $\sim12\%-13\%$ accuracy
just from the positions of $\sim1000$ galaxies in a volume of $(25~h^{-1}{\rm
Mpc})^3$ at $z=0$ while accounting for astrophysical uncertainties as modelled
in CAMELS. Incorporating information from galaxy properties, such as stellar
mass, stellar metallicity, and stellar radius, increases the accuracy to
$4\%-8\%$. Our models are built to be translational and rotational invariant,
and they can extract information from any scale larger than the minimum
distance between two galaxies. However, our models are not completely robust:
testing on simulations run with a different subgrid physics than the ones used
for training does not yield as accurate results.
|
107,789 | 107,789 |
An Intriguing Property of Geophysics Inversion
|
Inversion techniques are widely used to reconstruct subsurface physical
properties (e.g., velocity, conductivity) from surface-based geophysical
measurements (e.g., seismic, electric/magnetic (EM) data). The problems are
governed by partial differential equations (PDEs) like the wave or Maxwell's
equations. Solving geophysical inversion problems is challenging due to the
ill-posedness and high computational cost. To alleviate those issues, recent
studies leverage deep neural networks to learn the inversion mappings from
measurements to the property directly. In this paper, we show that such a
mapping can be well modeled by a very shallow (but not wide) network with only
five layers. This is achieved based on our new finding of an intriguing
property: a near-linear relationship between the input and output, after
applying integral transform in high dimensional space. In particular, when
dealing with the inversion from seismic data to subsurface velocity governed by
a wave equation, the integral results of velocity with Gaussian kernels are
linearly correlated to the integral of seismic data with sine kernels.
Furthermore, this property can be easily turned into a light-weight
encoder-decoder network for inversion. The encoder contains the integration of
seismic data and the linear transformation without need for fine-tuning. The
decoder only consists of a single transformer block to reverse the integral of
velocity. Experiments show that this interesting property holds for two
geophysics inversion problems over four different datasets. Compared to much
deeper InversionNet, our method achieves comparable accuracy, but consumes
significantly fewer parameters.
|
107,790 | 107,790 |
GCN-FFNN: A Two-Stream Deep Model for Learning Solution to Partial
Differential Equations
|
This paper introduces a novel two-stream deep model based on graph
convolutional network (GCN) architecture and feed-forward neural networks
(FFNN) for learning the solution of nonlinear partial differential equations
(PDEs). The model aims at incorporating both graph and grid input
representations using two streams corresponding to GCN and FFNN models,
respectively. Each stream layer receives and processes its own input
representation. As opposed to FFNN which receives a grid-like structure, the
GCN stream layer operates on graph input data where the neighborhood
information is incorporated through the adjacency matrix of the graph. In this
way, the proposed GCN-FFNN model learns from two types of input
representations, i.e. grid and graph data, obtained via the discretization of
the PDE domain. The GCN-FFNN model is trained in two phases. In the first
phase, the model parameters of each stream are trained separately. Both streams
employ the same error function to adjust their parameters by enforcing the
models to satisfy the given PDE as well as its initial and boundary conditions
on grid or graph collocation (training) data. In the second phase, the learned
parameters of two-stream layers are frozen and their learned representation
solutions are fed to fully connected layers whose parameters are learned using
the previously used error function. The learned GCN-FFNN model is tested on
test data located both inside and outside the PDE domain. The obtained
numerical results demonstrate the applicability and efficiency of the proposed
GCN-FFNN model over individual GCN and FFNN models on 1D-Burgers,
1D-Schr\"odinger, 2D-Burgers and 2D-Schr\"odinger equations.
|
107,791 | 107,791 |
CAVES: A Dataset to facilitate Explainable Classification and
Summarization of Concerns towards COVID Vaccines
|
Convincing people to get vaccinated against COVID-19 is a key societal
challenge in the present times. As a first step towards this goal, many prior
works have relied on social media analysis to understand the specific concerns
that people have towards these vaccines, such as potential side-effects,
ineffectiveness, political factors, and so on. Though there are datasets that
broadly classify social media posts into Anti-vax and Pro-Vax labels, there is
no dataset (to our knowledge) that labels social media posts according to the
specific anti-vaccine concerns mentioned in the posts. In this paper, we have
curated CAVES, the first large-scale dataset containing about 10k COVID-19
anti-vaccine tweets labelled into various specific anti-vaccine concerns in a
multi-label setting. This is also the first multi-label classification dataset
that provides explanations for each of the labels. Additionally, the dataset
also provides class-wise summaries of all the tweets. We also perform
preliminary experiments on the dataset and show that this is a very challenging
dataset for multi-label explainable classification and tweet summarization, as
is evident by the moderate scores achieved by some state-of-the-art models. Our
dataset and codes are available at: https://github.com/sohampoddar26/caves-data
|
107,792 | 107,792 |
Learning to Split for Automatic Bias Detection
|
Classifiers are biased when trained on biased datasets. As a remedy, we
propose Learning to Split (ls), an algorithm for automatic bias detection.
Given a dataset with input-label pairs, ls learns to split this dataset so that
predictors trained on the training split generalize poorly to the testing
split. This performance gap provides a proxy for measuring the degree of bias
in the learned features and can therefore be used to reduce biases. Identifying
non-generalizable splits is challenging as we don't have any explicit
annotations about how to split. In this work, we show that the prediction
correctness of the testing example can be used as a source of weak supervision:
generalization performance will drop if we move examples that are predicted
correctly away from the testing split, leaving only those that are
mispredicted. We evaluate our approach on Beer Review, Waterbirds, CelebA and
MNLI. Empirical results show that ls is able to generate astonishingly
challenging splits that correlate with human-identified biases. Moreover, we
demonstrate that combining robust learning algorithms (such as group DRO) with
splits identified by ls enables automatic de-biasing. Compared with previous
state-of-the-arts, we substantially improves the worst-group performance (23.4%
on average) when the source of biases is unknown during training and
validation.
|
107,793 | 107,793 |
BEINIT: Avoiding Barren Plateaus in Variational Quantum Algorithms
|
Barren plateaus are a notorious problem in the optimization of variational
quantum algorithms and pose a critical obstacle in the quest for more efficient
quantum machine learning algorithms. Many potential reasons for barren plateaus
have been identified but few solutions have been proposed to avoid them in
practice. Existing solutions are mainly focused on the initialization of
unitary gate parameters without taking into account the changes induced by
input data. In this paper, we propose an alternative strategy which initializes
the parameters of a unitary gate by drawing from a beta distribution. The
hyperparameters of the beta distribution are estimated from the data. To
further prevent barren plateau during training we add a novel perturbation at
every gradient descent step. Taking these ideas together, we empirically show
that our proposed framework significantly reduces the possibility of a complex
quantum neural network getting stuck in a barren plateau.
|
107,794 | 107,794 |
High Dimensional Bayesian Optimization with Kernel Principal Component
Analysis
|
Bayesian Optimization (BO) is a surrogate-based global optimization strategy
that relies on a Gaussian Process regression (GPR) model to approximate the
objective function and an acquisition function to suggest candidate points. It
is well-known that BO does not scale well for high-dimensional problems because
the GPR model requires substantially more data points to achieve sufficient
accuracy and acquisition optimization becomes computationally expensive in high
dimensions. Several recent works aim at addressing these issues, e.g., methods
that implement online variable selection or conduct the search on a
lower-dimensional sub-manifold of the original search space. Advancing our
previous work of PCA-BO that learns a linear sub-manifold, this paper proposes
a novel kernel PCA-assisted BO (KPCA-BO) algorithm, which embeds a non-linear
sub-manifold in the search space and performs BO on this sub-manifold.
Intuitively, constructing the GPR model on a lower-dimensional sub-manifold
helps improve the modeling accuracy without requiring much more data from the
objective function. Also, our approach defines the acquisition function on the
lower-dimensional sub-manifold, making the acquisition optimization more
manageable.
We compare the performance of KPCA-BO to the vanilla BO and PCA-BO on the
multi-modal problems of the COCO/BBOB benchmark suite. Empirical results show
that KPCA-BO outperforms BO in terms of convergence speed on most test
problems, and this benefit becomes more significant when the dimensionality
increases. For the 60D functions, KPCA-BO surpasses PCA-BO in many test cases.
Moreover, it efficiently reduces the CPU time required to train the GPR model
and optimize the acquisition function compared to the vanilla BO.
|
107,795 | 107,795 |
Probabilistic Permutation Graph Search: Black-Box Optimization for
Fairness in Ranking
|
There are several measures for fairness in ranking, based on different
underlying assumptions and perspectives. PL optimization with the REINFORCE
algorithm can be used for optimizing black-box objective functions over
permutations. In particular, it can be used for optimizing fairness measures.
However, though effective for queries with a moderate number of repeating
sessions, PL optimization has room for improvement for queries with a small
number of repeating sessions.
In this paper, we present a novel way of representing permutation
distributions, based on the notion of permutation graphs. Similar to PL, our
distribution representation, called PPG, can be used for black-box optimization
of fairness. Different from PL, where pointwise logits are used as the
distribution parameters, in PPG pairwise inversion probabilities together with
a reference permutation construct the distribution. As such, the reference
permutation can be set to the best sampled permutation regarding the objective
function, making PPG suitable for both deterministic and stochastic rankings.
Our experiments show that PPG, while comparable to PL for larger session
repetitions (i.e., stochastic ranking), improves over PL for optimizing
fairness metrics for queries with one session (i.e., deterministic ranking).
Additionally, when accurate utility estimations are available, e.g., in tabular
models, the performance of PPG in fairness optimization is significantly
boosted compared to lower quality utility estimations from a learning to rank
model, leading to a large performance gap with PL. Finally, the pairwise
probabilities make it possible to impose pairwise constraints such as "item
$d_1$ should always be ranked higher than item $d_2$." Such constraints can be
used to simultaneously optimize the fairness metric and control another
objective such as ranking performance.
|
107,796 | 107,796 |
Triformer: Triangular, Variable-Specific Attentions for Long Sequence
Multivariate Time Series Forecasting--Full Version
|
A variety of real-world applications rely on far future information to make
decisions, thus calling for efficient and accurate long sequence multivariate
time series forecasting. While recent attention-based forecasting models show
strong abilities in capturing long-term dependencies, they still suffer from
two key limitations. First, canonical self attention has a quadratic complexity
w.r.t. the input time series length, thus falling short in efficiency. Second,
different variables' time series often have distinct temporal dynamics, which
existing studies fail to capture, as they use the same model parameter space,
e.g., projection matrices, for all variables' time series, thus falling short
in accuracy. To ensure high efficiency and accuracy, we propose Triformer, a
triangular, variable-specific attention. (i) Linear complexity: we introduce a
novel patch attention with linear complexity. When stacking multiple layers of
the patch attentions, a triangular structure is proposed such that the layer
sizes shrink exponentially, thus maintaining linear complexity. (ii)
Variable-specific parameters: we propose a light-weight method to enable
distinct sets of model parameters for different variables' time series to
enhance accuracy without compromising efficiency and memory usage. Strong
empirical evidence on four datasets from multiple domains justifies our design
choices, and it demonstrates that Triformer outperforms state-of-the-art
methods w.r.t. both accuracy and efficiency. This is an extended version of
"Triformer: Triangular, Variable-Specific Attentions for Long Sequence
Multivariate Time Series Forecasting", to appear in IJCAI 2022 [Cirstea et al.,
2022a], including additional experimental results.
|
107,797 | 107,797 |
Formulating Robustness Against Unforeseen Attacks
|
Existing defenses against adversarial examples such as adversarial training
typically assume that the adversary will conform to a specific or known threat
model, such as $\ell_p$ perturbations within a fixed budget. In this paper, we
focus on the scenario where there is a mismatch in the threat model assumed by
the defense during training, and the actual capabilities of the adversary at
test time. We ask the question: if the learner trains against a specific
"source" threat model, when can we expect robustness to generalize to a
stronger unknown "target" threat model during test-time? Our key contribution
is to formally define the problem of learning and generalization with an
unforeseen adversary, which helps us reason about the increase in adversarial
risk from the conventional perspective of a known adversary. Applying our
framework, we derive a generalization bound which relates the generalization
gap between source and target threat models to variation of the feature
extractor, which measures the expected maximum difference between extracted
features across a given threat model. Based on our generalization bound, we
propose adversarial training with variation regularization (AT-VR) which
reduces variation of the feature extractor across the source threat model
during training. We empirically demonstrate that AT-VR can lead to improved
generalization to unforeseen attacks during test-time compared to standard
adversarial training on Gaussian and image datasets.
|
107,798 | 107,798 |
AGIC: Approximate Gradient Inversion Attack on Federated Learning
|
Federated learning is a private-by-design distributed learning paradigm where
clients train local models on their own data before a central server aggregates
their local updates to compute a global model. Depending on the aggregation
method used, the local updates are either the gradients or the weights of local
learning models. Recent reconstruction attacks apply a gradient inversion
optimization on the gradient update of a single minibatch to reconstruct the
private data used by clients during training. As the state-of-the-art
reconstruction attacks solely focus on single update, realistic adversarial
scenarios are overlooked, such as observation across multiple updates and
updates trained from multiple mini-batches. A few studies consider a more
challenging adversarial scenario where only model updates based on multiple
mini-batches are observable, and resort to computationally expensive simulation
to untangle the underlying samples for each local step. In this paper, we
propose AGIC, a novel Approximate Gradient Inversion Attack that efficiently
and effectively reconstructs images from both model or gradient updates, and
across multiple epochs. In a nutshell, AGIC (i) approximates gradient updates
of used training samples from model updates to avoid costly simulation
procedures, (ii) leverages gradient/model updates collected from multiple
epochs, and (iii) assigns increasing weights to layers with respect to the
neural network structure for reconstruction quality. We extensively evaluate
AGIC on three datasets, CIFAR-10, CIFAR-100 and ImageNet. Our results show that
AGIC increases the peak signal-to-noise ratio (PSNR) by up to 50% compared to
two representative state-of-the-art gradient inversion attacks. Furthermore,
AGIC is faster than the state-of-the-art simulation based attack, e.g., it is
5x faster when attacking FedAvg with 8 local steps in between model updates.
|
107,799 | 107,799 |
Depth Estimation with Simplified Transformer
|
Transformer and its variants have shown state-of-the-art results in many
vision tasks recently, ranging from image classification to dense prediction.
Despite of their success, limited work has been reported on improving the model
efficiency for deployment in latency-critical applications, such as autonomous
driving and robotic navigation. In this paper, we aim at improving upon the
existing transformers in vision, and propose a method for self-supervised
monocular Depth Estimation with Simplified Transformer (DEST), which is
efficient and particularly suitable for deployment on GPU-based platforms.
Through strategic design choices, our model leads to significant reduction in
model size, complexity, as well as inference latency, while achieving superior
accuracy as compared to state-of-the-art. We also show that our design
generalize well to other dense prediction task without bells and whistles.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.