Unnamed: 0
int64 0
5k
| title
stringlengths 9
210
| abstract
stringlengths 164
1.92k
|
|---|---|---|
2,100 | Self-critiquing models for assisting human evaluators | We fine-tune large language models to write natural language critiques
(natural language critical comments) using behavioral cloning. On a topic-based
summarization task, critiques written by our models help humans find flaws in
summaries that they would have otherwise missed. Our models help find naturally
occurring flaws in both model and human written summaries, and intentional
flaws in summaries written by humans to be deliberately misleading. We study
scaling properties of critiquing with both topic-based summarization and
synthetic tasks. Larger models write more helpful critiques, and on most tasks,
are better at self-critiquing, despite having harder-to-critique outputs.
Larger models can also integrate their own self-critiques as feedback, refining
their own summaries into better ones. Finally, we motivate and introduce a
framework for comparing critiquing ability to generation and discrimination
ability. Our measurements suggest that even large models may still have
relevant knowledge they cannot or do not articulate as critiques. These results
are a proof of concept for using AI-assisted human feedback to scale the
supervision of machine learning systems to tasks that are difficult for humans
to evaluate directly. We release our training datasets, as well as samples from
our critique assistance experiments. |
2,101 | SGD Noise and Implicit Low-Rank Bias in Deep Neural Networks | We analyze deep ReLU neural networks trained with mini-batch Stochastic
Gradient Descent (SGD) and weight decay. We study the source of SGD noise and
prove that when training with weight decay, the only solutions of SGD at
convergence are zero functions. Furthermore, we show, both theoretically and
empirically, that when training a neural network using SGD with weight decay
and small batch size, the resulting weight matrices are expected to be of small
rank. Our analysis relies on a minimal set of assumptions and the neural
networks may be arbitrarily wide or deep, and may include residual connections,
as well as batch normalization layers. |
2,102 | Neurotoxin: Durable Backdoors in Federated Learning | Due to their decentralized nature, federated learning (FL) systems have an
inherent vulnerability during their training to adversarial backdoor attacks.
In this type of attack, the goal of the attacker is to use poisoned updates to
implant so-called backdoors into the learned model such that, at test time, the
model's outputs can be fixed to a given target for certain inputs. (As a simple
toy example, if a user types "people from New York" into a mobile keyboard app
that uses a backdoored next word prediction model, then the model could
autocomplete the sentence to "people from New York are rude"). Prior work has
shown that backdoors can be inserted into FL models, but these backdoors are
often not durable, i.e., they do not remain in the model after the attacker
stops uploading poisoned updates. Thus, since training typically continues
progressively in production FL systems, an inserted backdoor may not survive
until deployment. Here, we propose Neurotoxin, a simple one-line modification
to existing backdoor attacks that acts by attacking parameters that are changed
less in magnitude during training. We conduct an exhaustive evaluation across
ten natural language processing and computer vision tasks, and we find that we
can double the durability of state of the art backdoors. |
2,103 | Dual-Stream Transformer with Cross-Attention on Whole-Slide Image Pyramids for Cancer Prognosis | The cancer prognosis on gigapixel Whole-Slide Images (WSIs) has always been a
challenging task. Most existing approaches focus solely on single-resolution
images. The multi-resolution schemes, utilizing image pyramids to enhance WSI
visual representations, have not yet been paid enough attention to. In order to
explore a multi-resolution solution for improving cancer prognosis accuracy,
this paper proposes a dual-stream architecture to model WSIs by an image
pyramid strategy. This architecture consists of two sub-streams: one for
low-resolution WSIs, and the other especially for high-resolution ones.
Compared to other approaches, our scheme has three highlights: (i) there exists
a one-to-one relation between stream and resolution; (ii) a square pooling
layer is added to align the patches from two resolution streams, largely
reducing computation cost and enabling a natural stream feature fusion; (iii) a
cross-attention-based method is proposed to pool high-resolution patches
spatially under the guidance of low-resolution ones. We validate our scheme on
three publicly-available datasets with a total number of 3,101 WSIs from 1,911
patients. Experimental results verify that (i) hierarchical dual-stream
representation is more effective than single-stream ones for cancer prognosis,
gaining an average C-Index rise of 5.0% and 1.8% on a single low-resolution and
high-resolution stream, respectively; (ii) our dual-stream scheme could
outperform current state-of-the-art ones, by an average C-Index improvement of
5.1%; (iii) the cancer diseases with observable survival differences could have
different preferences for model complexity. Our scheme could serve as an
alternative tool for further facilitating WSI prognosis research. |
2,104 | Learning-Based Data Storage [Vision] (Technical Report) | Deep neural network (DNN) and its variants have been extensively used for a
wide spectrum of real applications such as image classification, face/speech
recognition, fraud detection, and so on. In addition to many important machine
learning tasks, as artificial networks emulating the way brain cells function,
DNNs also show the capability of storing non-linear relationships between input
and output data, which exhibits the potential of storing data via DNNs. We
envision a new paradigm of data storage, "DNN-as-a-Database", where data are
encoded in well-trained machine learning models. Compared with conventional
data storage that directly records data in raw formats, learning-based
structures (e.g., DNN) can implicitly encode data pairs of inputs and outputs
and compute/materialize actual output data of different resolutions only if
input data are provided. This new paradigm can greatly enhance the data
security by allowing flexible data privacy settings on different levels,
achieve low space consumption and fast computation with the acceleration of new
hardware (e.g., Diffractive Neural Network and AI chips), and can be
generalized to distributed DNN-based storage/computing. In this paper, we
propose this novel concept of learning-based data storage, which utilizes a
learning structure called learning-based memory unit (LMU), to store, organize,
and retrieve data. As a case study, we use DNNs as the engine in the LMU, and
study the data capacity and accuracy of the DNN-based data storage. Our
preliminary experimental results show the feasibility of the learning-based
data storage by achieving high (100%) accuracy of the DNN storage. We explore
and design effective solutions to utilize the DNN-based data storage to manage
and query relational tables. We discuss how to generalize our solutions to
other data types (e.g., graphs) and environments such as distributed DNN
storage/computing. |
2,105 | The Rough Topology for Numerical Data | In this paper, we give a generalization of the rough topology and the core to
numerical data by classifying objects in terms of the attribute values. New
approach to find the core for numerical data is discussed. Then a measurement
to find whether an attribute is in the core or not is given. This new method
for finding the core is used for attribute reduction. It is tested and compared
by using machine learning algorithms. Finally, the algorithms and codes to
convert a data to pertinent data and to find core is also provided. |
2,106 | Distributed Differential Privacy in Multi-Armed Bandits | We consider the standard $K$-armed bandit problem under a distributed trust
model of differential privacy (DP), which enables to guarantee privacy without
a trustworthy server. Under this trust model, previous work largely focus on
achieving privacy using a shuffle protocol, where a batch of users data are
randomly permuted before sending to a central server. This protocol achieves
($\epsilon,\delta$) or approximate-DP guarantee by sacrificing an additional
additive $O\!\left(\!\frac{K\log T\sqrt{\log(1/\delta)}}{\epsilon}\!\right)\!$
cost in $T$-step cumulative regret. In contrast, the optimal privacy cost for
achieving a stronger ($\epsilon,0$) or pure-DP guarantee under the widely used
central trust model is only $\Theta\!\left(\!\frac{K\log
T}{\epsilon}\!\right)\!$, where, however, a trusted server is required. In this
work, we aim to obtain a pure-DP guarantee under distributed trust model while
sacrificing no more regret than that under central trust model. We achieve this
by designing a generic bandit algorithm based on successive arm elimination,
where privacy is guaranteed by corrupting rewards with an equivalent discrete
Laplace noise ensured by a secure computation protocol. We also show that our
algorithm, when instantiated with Skellam noise and the secure protocol,
ensures \emph{R\'{e}nyi differential privacy} -- a stronger notion than
approximate DP -- under distributed trust model with a privacy cost of
$O\!\left(\!\frac{K\sqrt{\log T}}{\epsilon}\!\right)\!$. |
2,107 | Mining Multi-Label Samples from Single Positive Labels | Conditional generative adversarial networks (cGANs) have shown superior
results in class-conditional generation tasks. In order to simultaneously
control multiple conditions, cGANs require multi-label training datasets, where
multiple labels can be assigned to each data instance. Nevertheless, the
tremendous annotation cost limits the accessibility of multi-label datasets in
the real-world scenarios. Hence, we explore the practical setting called single
positive setting, where each data instance is annotated by only one positive
label with no explicit negative labels. To generate multi-label data in the
single positive setting, we propose a novel sampling approach called
single-to-multi-label (S2M) sampling, based on the Markov chain Monte Carlo
method. As a widely applicable "add-on" method, our proposed S2M sampling
enables existing unconditional and conditional GANs to draw high-quality
multi-label data with a minimal annotation cost. Extensive experiments on real
image datasets verify the effectiveness and correctness of our method, even
when compared to a model trained with fully annotated datasets. |
2,108 | Consistent Attack: Universal Adversarial Perturbation on Embodied Vision Navigation | Embodied agents in vision navigation coupled with deep neural networks have
attracted increasing attention. However, deep neural networks are vulnerable to
malicious adversarial noises, which may potentially cause catastrophic failures
in Embodied Vision Navigation. Among these adversarial noises, universal
adversarial perturbations (UAP), i.e., the image-agnostic perturbation applied
on each frame received by the agent, are more critical for Embodied Vision
Navigation since they are computation-efficient and application-practical
during the attack. However, existing UAP methods do not consider the system
dynamics of Embodied Vision Navigation. For extending UAP in the sequential
decision setting, we formulate the disturbed environment under the universal
noise $\delta$, as a $\delta$-disturbed Markov Decision Process ($\delta$-MDP).
Based on the formulation, we analyze the properties of $\delta$-MDP and propose
two novel Consistent Attack methods for attacking Embodied agents, which first
consider the dynamic of the MDP by estimating the disturbed Q function and the
disturbed distribution. In spite of victim models, our Consistent Attack can
cause a significant drop in the performance for the Goalpoint task in habitat.
Extensive experimental results indicate that there exist potential risks for
applying Embodied Vision Navigation methods to the real world. |
2,109 | Matching options to tasks using Option-Indexed Hierarchical Reinforcement Learning | The options framework in Hierarchical Reinforcement Learning breaks down
overall goals into a combination of options or simpler tasks and associated
policies, allowing for abstraction in the action space. Ideally, these options
can be reused across different higher-level goals; indeed, such reuse is
necessary to realize the vision of a continual learning agent that can
effectively leverage its prior experience. Previous approaches have only
proposed limited forms of transfer of prelearned options to new task settings.
We propose a novel option indexing approach to hierarchical learning (OI-HRL),
where we learn an affinity function between options and the items present in
the environment. This allows us to effectively reuse a large library of
pretrained options, in zero-shot generalization at test time, by restricting
goal-directed learning to only those options relevant to the task at hand. We
develop a meta-training loop that learns the representations of options and
environments over a series of HRL problems, by incorporating feedback about the
relevance of retrieved options to the higher-level goal. We evaluate OI-HRL in
two simulated settings - the CraftWorld and AI2THOR environments - and show
that we achieve performance competitive with oracular baselines, and
substantial gains over a baseline that has the entire option pool available for
learning the hierarchical policy. |
2,110 | Regularization Penalty Optimization for Addressing Data Quality Variance in OoD Algorithms | Due to the poor generalization performance of traditional empirical risk
minimization (ERM) in the case of distributional shift, Out-of-Distribution
(OoD) generalization algorithms receive increasing attention. However, OoD
generalization algorithms overlook the great variance in the quality of
training data, which significantly compromises the accuracy of these methods.
In this paper, we theoretically reveal the relationship between training data
quality and algorithm performance and analyze the optimal regularization scheme
for Lipschitz regularized invariant risk minimization. A novel algorithm is
proposed based on the theoretical results to alleviate the influence of
low-quality data at both the sample level and the domain level. The experiments
on both the regression and classification benchmarks validate the effectiveness
of our method with statistical significance. |
2,111 | Learning to Detect with Constant False Alarm Rate | We consider the use of machine learning for hypothesis testing with an
emphasis on target detection. Classical model-based solutions rely on comparing
likelihoods. These are sensitive to imperfect models and are often
computationally expensive. In contrast, data-driven machine learning is often
more robust and yields classifiers with fixed computational complexity. Learned
detectors usually provide high accuracy with low complexity but do not have a
constant false alarm rate (CFAR) as required in many applications. To close
this gap, we propose to add a term to the loss function that promotes similar
distributions of the detector under any null hypothesis scenario. Experiments
show that our approach leads to near CFAR detectors with similar accuracy as
their competitors. |
2,112 | Finite-Time Analysis of Fully Decentralized Single-Timescale Actor-Critic | Decentralized Actor-Critic (AC) algorithms have been widely utilized for
multi-agent reinforcement learning (MARL) and have achieved remarkable success.
Apart from its empirical success, the theoretical convergence property of
decentralized AC algorithms is largely unexplored. The existing finite-time
convergence results are derived based on either double-loop update or
two-timescale step sizes rule, which is not often adopted in real
implementation. In this work, we introduce a fully decentralized AC algorithm,
where actor, critic, and global reward estimator are updated in an alternating
manner with step sizes being of the same order, namely, we adopt the
\emph{single-timescale} update. Theoretically, using linear approximation for
value and reward estimation, we show that our algorithm has sample complexity
of $\tilde{\mathcal{O}}(\epsilon^{-2})$ under Markovian sampling, which matches
the optimal complexity with double-loop implementation (here,
$\tilde{\mathcal{O}}$ hides a log term). The sample complexity can be improved
to ${\mathcal{O}}(\epsilon^{-2})$ under the i.i.d. sampling scheme. The central
to establishing our complexity results is \emph{the hidden smoothness of the
optimal critic variable} we revealed. We also provide a local action
privacy-preserving version of our algorithm and its analysis. Finally, we
conduct experiments to show the superiority of our algorithm over the existing
decentralized AC algorithms. |
2,113 | Darknet Traffic Classification and Adversarial Attacks | The anonymous nature of darknets is commonly exploited for illegal
activities. Previous research has employed machine learning and deep learning
techniques to automate the detection of darknet traffic in an attempt to block
these criminal activities. This research aims to improve darknet traffic
detection by assessing Support Vector Machines (SVM), Random Forest (RF),
Convolutional Neural Networks (CNN), and Auxiliary-Classifier Generative
Adversarial Networks (AC-GAN) for classification of such traffic and the
underlying application types. We find that our RF model outperforms the
state-of-the-art machine learning techniques used in prior work with the
CIC-Darknet2020 dataset. To evaluate the robustness of our RF classifier, we
obfuscate select application type classes to simulate realistic adversarial
attack scenarios. We demonstrate that our best-performing classifier can be
defeated by such attacks, and we consider ways to deal with such adversarial
attacks. |
2,114 | Machine learning based surrogate modeling with SVD enabled training for nonlinear civil structures subject to dynamic loading | The computationally expensive estimation of engineering demand parameters
(EDPs) via finite element (FE) models, while considering earthquake and
parameter uncertainty limits the use of the Performance Based Earthquake
Engineering framework. Attempts have been made to substitute FE models with
surrogate models, however, most of these models are a function of building
parameters only. This necessitates re-training for earthquakes not previously
seen by the surrogate. In this paper, the authors propose a machine learning
based surrogate model framework, which considers both these uncertainties in
order to predict for unseen earthquakes. Accordingly,earthquakes are
characterized by their projections on an orthonormal basis, computed using SVD
of a representative ground motion suite. This enables one to generate large
varieties of earthquakes by randomly sampling these weights and multiplying
them with the basis. The weights along with the constitutive parameters serve
as inputs to a machine learning model with EDPs as the desired output. Four
competing machine learning models were tested and it was observed that a deep
neural network (DNN) gave the most accurate prediction. The framework is
validated by using it to successfully predict the peak response of one-story
and three-story buildings represented using stick models, subjected to unseen
far-field ground motions. |
2,115 | Don't "research fast and break things": On the ethics of Computational Social Science | This article is concerned with setting up practical guardrails within the
research activities and environments of CSS. It aims to provide CSS scholars,
as well as policymakers and other stakeholders who apply CSS methods, with the
critical and constructive means needed to ensure that their practices are
ethical, trustworthy, and responsible. It begins by providing a taxonomy of the
ethical challenges faced by researchers in the field of CSS. These are
challenges related to (1) the treatment of research subjects, (2) the impacts
of CSS research on affected individuals and communities, (3) the quality of CSS
research and to its epistemological status, (4) research integrity, and (5)
research equity. Taking these challenges as a motivation for cultural
transformation, it then argues for the end-to-end incorporation of habits of
responsible research and innovation (RRI) into CSS practices, focusing on the
role that contextual considerations, anticipatory reflection, impact
assessment, public engagement, and justifiable and well-documented action
should play across the research lifecycle. In proposing the inclusion of habits
of RRI in CSS practices, the chapter lays out several practical steps needed
for ethical, trustworthy, and responsible CSS research activities. These
include stakeholder engagement processes, research impact assessments, data
lifecycle documentation, bias self-assessments, and transparent research
reporting protocols. |
2,116 | PAC-Net: A Model Pruning Approach to Inductive Transfer Learning | Inductive transfer learning aims to learn from a small amount of training
data for the target task by utilizing a pre-trained model from the source task.
Most strategies that involve large-scale deep learning models adopt
initialization with the pre-trained model and fine-tuning for the target task.
However, when using over-parameterized models, we can often prune the model
without sacrificing the accuracy of the source task. This motivates us to adopt
model pruning for transfer learning with deep learning models. In this paper,
we propose PAC-Net, a simple yet effective approach for transfer learning based
on pruning. PAC-Net consists of three steps: Prune, Allocate, and Calibrate
(PAC). The main idea behind these steps is to identify essential weights for
the source task, fine-tune on the source task by updating the essential
weights, and then calibrate on the target task by updating the remaining
redundant weights. Under the various and extensive set of inductive transfer
learning experiments, we show that our method achieves state-of-the-art
performance by a large margin. |
2,117 | A Functional Information Perspective on Model Interpretation | Contemporary predictive models are hard to interpret as their deep nets
exploit numerous complex relations between input elements. This work suggests a
theoretical framework for model interpretability by measuring the contribution
of relevant features to the functional entropy of the network with respect to
the input. We rely on the log-Sobolev inequality that bounds the functional
entropy by the functional Fisher information with respect to the covariance of
the data. This provides a principled way to measure the amount of information
contribution of a subset of features to the decision function. Through
extensive experiments, we show that our method surpasses existing
interpretability sampling-based methods on various data signals such as image,
text, and audio. |
2,118 | tBDFS: Temporal Graph Neural Network Leveraging DFS | Temporal graph neural networks (temporal GNNs) have been widely researched,
reaching state-of-the-art results on multiple prediction tasks. A common
approach employed by most previous works is to apply a layer that aggregates
information from the historical neighbors of a node. Taking a different
research direction, in this work, we propose tBDFS -- a novel temporal GNN
architecture. tBDFS applies a layer that efficiently aggregates information
from temporal paths to a given (target) node in the graph. For each given node,
the aggregation is applied in two stages: (1) A single representation is
learned for each temporal path ending in that node, and (2) all path
representations are aggregated into a final node representation. Overall, our
goal is not to add new information to a node, but rather observe the same exact
information in a new perspective. This allows our model to directly observe
patterns that are path-oriented rather than neighborhood-oriented. This can be
thought as a Depth-First Search (DFS) traversal over the temporal graph,
compared to the popular Breath-First Search (BFS) traversal that is applied in
previous works. We evaluate tBDFS over multiple link prediction tasks and show
its favorable performance compared to state-of-the-art baselines. To the best
of our knowledge, we are the first to apply a temporal-DFS neural network. |
2,119 | Balancing Bias and Variance for Active Weakly Supervised Learning | As a widely used weakly supervised learning scheme, modern multiple instance
learning (MIL) models achieve competitive performance at the bag level.
However, instance-level prediction, which is essential for many important
applications, remains largely unsatisfactory. We propose to conduct novel
active deep multiple instance learning that samples a small subset of
informative instances for annotation, aiming to significantly boost the
instance-level prediction. A variance regularized loss function is designed to
properly balance the bias and variance of instance-level predictions, aiming to
effectively accommodate the highly imbalanced instance distribution in MIL and
other fundamental challenges. Instead of directly minimizing the variance
regularized loss that is non-convex, we optimize a distributionally robust bag
level likelihood as its convex surrogate. The robust bag likelihood provides a
good approximation of the variance based MIL loss with a strong theoretical
guarantee. It also automatically balances bias and variance, making it
effective to identify the potentially positive instances to support active
sampling. The robust bag likelihood can be naturally integrated with a deep
architecture to support deep model training using mini-batches of
positive-negative bag pairs. Finally, a novel P-F sampling function is
developed that combines a probability vector and predicted instance scores,
obtained by optimizing the robust bag likelihood. By leveraging the key MIL
assumption, the sampling function can explore the most challenging bags and
effectively detect their positive instances for annotation, which significantly
improves the instance-level prediction. Experiments conducted over multiple
real-world datasets clearly demonstrate the state-of-the-art instance-level
prediction achieved by the proposed model. |
2,120 | A Survey on Uncertainty Reasoning and Quantification for Decision Making: Belief Theory Meets Deep Learning | An in-depth understanding of uncertainty is the first step to making
effective decisions under uncertainty. Deep/machine learning (ML/DL) has been
hugely leveraged to solve complex problems involved with processing
high-dimensional data. However, reasoning and quantifying different types of
uncertainties to achieve effective decision-making have been much less explored
in ML/DL than in other Artificial Intelligence (AI) domains. In particular,
belief/evidence theories have been studied in KRR since the 1960s to reason and
measure uncertainties to enhance decision-making effectiveness. We found that
only a few studies have leveraged the mature uncertainty research in
belief/evidence theories in ML/DL to tackle complex problems under different
types of uncertainty. In this survey paper, we discuss several popular belief
theories and their core ideas dealing with uncertainty causes and types and
quantifying them, along with the discussions of their applicability in ML/DL.
In addition, we discuss three main approaches that leverage belief theories in
Deep Neural Networks (DNNs), including Evidential DNNs, Fuzzy DNNs, and Rough
DNNs, in terms of their uncertainty causes, types, and quantification methods
along with their applicability in diverse problem domains. Based on our
in-depth survey, we discuss insights, lessons learned, limitations of the
current state-of-the-art bridging belief theories and ML/DL, and finally,
future research directions. |
2,121 | Universality and approximation bounds for echo state networks with random weights | We study the uniform approximation of echo state networks with randomly
generated internal weights. These models, in which only the readout weights are
optimized during training, have made empirical success in learning dynamical
systems. We address the representational capacity of these models by showing
that they are universal under weak conditions. Our main result gives a
sufficient condition for the activation function and a sampling procedure for
the internal weights so that echo state networks can approximate any continuous
casual time-invariant operators with high probability. In particular, for ReLU
activation, we quantify the approximation error of echo state networks for
sufficiently regular operators. |
2,122 | Federated Learning on Riemannian Manifolds | Federated learning (FL) has found many important applications in
smart-phone-APP based machine learning applications. Although many algorithms
have been studied for FL, to the best of our knowledge, algorithms for FL with
nonconvex constraints have not been studied. This paper studies FL over
Riemannian manifolds, which finds important applications such as federated PCA
and federated kPCA. We propose a Riemannian federated SVRG (RFedSVRG) method to
solve federated optimization over Riemannian manifolds. We analyze its
convergence rate under different scenarios. Numerical experiments are conducted
to compare RFedSVRG with the Riemannian counterparts of FedAvg and FedProx. We
observed from the numerical experiments that the advantages of RFedSVRG are
significant. |
2,123 | An Efficient Method for Sample Adversarial Perturbations against Nonlinear Support Vector Machines | Adversarial perturbations have drawn great attentions in various machine
learning models. In this paper, we investigate the sample adversarial
perturbations for nonlinear support vector machines (SVMs). Due to the implicit
form of the nonlinear functions mapping data to the feature space, it is
difficult to obtain the explicit form of the adversarial perturbations. By
exploring the special property of nonlinear SVMs, we transform the optimization
problem of attacking nonlinear SVMs into a nonlinear KKT system. Such a system
can be solved by various numerical methods. Numerical results show that our
method is efficient in computing adversarial perturbations. |
2,124 | Variational Bayes Deep Operator Network: A data-driven Bayesian solver for parametric differential equations | Neural network based data-driven operator learning schemes have shown
tremendous potential in computational mechanics. DeepONet is one such neural
network architecture which has gained widespread appreciation owing to its
excellent prediction capabilities. Having said that, being set in a
deterministic framework exposes DeepONet architecture to the risk of
overfitting, poor generalization and in its unaltered form, it is incapable of
quantifying the uncertainties associated with its predictions. We propose in
this paper, a Variational Bayes DeepONet (VB-DeepONet) for operator learning,
which can alleviate these limitations of DeepONet architecture to a great
extent and give user additional information regarding the associated
uncertainty at the prediction stage. The key idea behind neural networks set in
Bayesian framework is that, the weights and bias of the neural network are
treated as probability distributions instead of point estimates and, Bayesian
inference is used to update their prior distribution. Now, to manage the
computational cost associated with approximating the posterior distribution,
the proposed VB-DeepONet uses \textit{variational inference}. Unlike Markov
Chain Monte Carlo schemes, variational inference has the capacity to take into
account high dimensional posterior distributions while keeping the associated
computational cost low. Different examples covering mechanics problems like
diffusion reaction, gravity pendulum, advection diffusion have been shown to
illustrate the performance of the proposed VB-DeepONet and comparisons have
also been drawn against DeepONet set in deterministic framework. |
2,125 | Dealing with Sparse Rewards in Continuous Control Robotics via Heavy-Tailed Policies | In this paper, we present a novel Heavy-Tailed Stochastic Policy Gradient
(HT-PSG) algorithm to deal with the challenges of sparse rewards in continuous
control problems. Sparse reward is common in continuous control robotics tasks
such as manipulation and navigation, and makes the learning problem hard due to
non-trivial estimation of value functions over the state space. This demands
either reward shaping or expert demonstrations for the sparse reward
environment. However, obtaining high-quality demonstrations is quite expensive
and sometimes even impossible. We propose a heavy-tailed policy parametrization
along with a modified momentum-based policy gradient tracking scheme (HT-SPG)
to induce a stable exploratory behavior to the algorithm. The proposed
algorithm does not require access to expert demonstrations. We test the
performance of HT-SPG on various benchmark tasks of continuous control with
sparse rewards such as 1D Mario, Pathological Mountain Car, Sparse Pendulum in
OpenAI Gym, and Sparse MuJoCo environments (Hopper-v2). We show consistent
performance improvement across all tasks in terms of high average cumulative
reward. HT-SPG also demonstrates improved convergence speed with minimum
samples, thereby emphasizing the sample efficiency of our proposed algorithm. |
2,126 | DeepEmotex: Classifying Emotion in Text Messages using Deep Transfer Learning | Transfer learning has been widely used in natural language processing through
deep pretrained language models, such as Bidirectional Encoder Representations
from Transformers and Universal Sentence Encoder. Despite the great success,
language models get overfitted when applied to small datasets and are prone to
forgetting when fine-tuned with a classifier. To remedy this problem of
forgetting in transferring deep pretrained language models from one domain to
another domain, existing efforts explore fine-tuning methods to forget less. We
propose DeepEmotex an effective sequential transfer learning method to detect
emotion in text. To avoid forgetting problem, the fine-tuning step is
instrumented by a large amount of emotion-labeled data collected from Twitter.
We conduct an experimental study using both curated Twitter data sets and
benchmark data sets. DeepEmotex models achieve over 91% accuracy for
multi-class emotion classification on test dataset. We evaluate the performance
of the fine-tuned DeepEmotex models in classifying emotion in EmoInt and
Stimulus benchmark datasets. The models correctly classify emotion in 73% of
the instances in the benchmark datasets. The proposed DeepEmotex-BERT model
outperforms Bi-LSTM result on the benchmark datasets by 23%. We also study the
effect of the size of the fine-tuning dataset on the accuracy of our models.
Our evaluation results show that fine-tuning with a large set of
emotion-labeled data improves both the robustness and effectiveness of the
resulting target task model. |
2,127 | Density Regression and Uncertainty Quantification with Bayesian Deep Noise Neural Networks | Deep neural network (DNN) models have achieved state-of-the-art predictive
accuracy in a wide range of supervised learning applications. However,
accurately quantifying the uncertainty in DNN predictions remains a challenging
task. For continuous outcome variables, an even more difficult problem is to
estimate the predictive density function, which not only provides a natural
quantification of the predictive uncertainty, but also fully captures the
random variation in the outcome. In this work, we propose the Bayesian Deep
Noise Neural Network (B-DeepNoise), which generalizes standard Bayesian DNNs by
extending the random noise variable from the output layer to all hidden layers.
The latent random noise equips B-DeepNoise with the flexibility to approximate
highly complex predictive distributions and accurately quantify predictive
uncertainty. For posterior computation, the unique structure of B-DeepNoise
leads to a closed-form Gibbs sampling algorithm that iteratively simulates from
the posterior full conditional distributions of the model parameters,
circumventing computationally intensive Metropolis-Hastings methods. A
theoretical analysis of B-DeepNoise establishes a recursive representation of
the predictive distribution and decomposes the predictive variance with respect
to the latent parameters. We evaluate B-DeepNoise against existing methods on
benchmark regression datasets, demonstrating its superior performance in terms
of prediction accuracy, uncertainty quantification accuracy, and uncertainty
quantification efficiency. To illustrate our method's usefulness in scientific
studies, we apply B-DeepNoise to predict general intelligence from neuroimaging
features in the Adolescent Brain Cognitive Development (ABCD) project. |
2,128 | An Unsupervised Deep-Learning Method for Bone Age Assessment | The bone age, reflecting the degree of development of the bones, can be used
to predict the adult height and detect endocrine diseases of children. Both
examinations of radiologists and variability of operators have a significant
impact on bone age assessment. To decrease human intervention , machine
learning algorithms are used to assess the bone age automatically. However,
conventional supervised deep-learning methods need pre-labeled data. In this
paper, based on the convolutional auto-encoder with constraints (CCAE), an
unsupervised deep-learning model proposed in the classification of the
fingerprint, we propose this model for the classification of the bone age and
baptize it BA-CCAE. In the proposed BA-CCAE model, the key regions of the raw
X-ray images of the bone age are encoded, yielding the latent vectors. The
K-means clustering algorithm is used to obtain the final classifications by
grouping the latent vectors of the bone images. A set of experiments on the
Radiological Society of North America pediatric bone age dataset (RSNA) show
that the accuracy of classifications at 48-month intervals is 76.15%. Although
the accuracy now is lower than most of the existing supervised models, the
proposed BA-CCAE model can establish the classification of bone age without any
pre-labeled data, and to the best of our knowledge, the proposed BA-CCAE is one
of the few trails using the unsupervised deep-learning method for the bone age
assessment. |
2,129 | Mathematical Theory of Bayesian Statistics for Unknown Information Source | In statistical inference, uncertainty is unknown and all models are wrong. A
person who makes a statistical model and a prior distribution is simultaneously
aware that they are fictional and virtual candidates. In order to study such
cases, several statistical measures have been constructed, such as cross
validation, information criteria, and marginal likelihood, however, their
mathematical properties have not yet been completely clarified when statistical
models are under- and over- parametrized. In this paper, we introduce a place
of mathematical theory of Bayesian statistics for unknown uncertainty, on which
we show general properties of cross validation, information criteria, and
marginal likelihood. The derived theory holds even if an unknown uncertainty is
unrealizable by a statistical model or even if the posterior distribution
cannot be approximated by any normal distribution, hence it gives a helpful
standpoint for a person who cannot believe in any specific model and prior. The
results are followings. (1) There exists a more precise statistical measure of
the generalization loss than leave-one-out cross validation and information
criterion based on the mathematical properties of them. (2) There exists a more
efficient approximation method of the free energy, which is the minus log
marginal likelihood, even if the posterior distribution cannot be approximated
by any normal distribution. (3) And the prior distributions optimized by the
cross validation and the widely applicable information criterion are
asymptotically equivalent to each other, which are different from that by the
marginal likelihood. |
2,130 | Physics-driven Deep Learning for PET/MRI | In this paper, we review physics- and data-driven reconstruction techniques
for simultaneous positron emission tomography (PET) / magnetic resonance
imaging (MRI) systems, which have significant advantages for clinical imaging
of cancer, neurological disorders, and heart disease. These reconstruction
approaches utilize priors, either structural or statistical, together with a
physics-based description of the PET system response. However, due to the
nested representation of the forward problem, direct PET/MRI reconstruction is
a nonlinear problem. We elucidate how a multi-faceted approach accommodates
hybrid data- and physics-driven machine learning for reconstruction of 3D
PET/MRI, summarizing important deep learning developments made in the last 5
years to address attenuation correction, scattering, low photon counts, and
data consistency. We also describe how applications of these multi-modality
approaches extend beyond PET/MRI to improving accuracy in radiation therapy
planning. We conclude by discussing opportunities for extending the current
state-of-the-art following the latest trends in physics- and deep
learning-based computational imaging and next-generation detector hardware. |
2,131 | Federated Learning with Research Prototypes for Multi-Center MRI-based Detection of Prostate Cancer with Diverse Histopathology | Early prostate cancer detection and staging from MRI are extremely
challenging tasks for both radiologists and deep learning algorithms, but the
potential to learn from large and diverse datasets remains a promising avenue
to increase their generalization capability both within- and across clinics. To
enable this for prototype-stage algorithms, where the majority of existing
research remains, in this paper we introduce a flexible federated learning
framework for cross-site training, validation, and evaluation of deep prostate
cancer detection algorithms. Our approach utilizes an abstracted representation
of the model architecture and data, which allows unpolished prototype deep
learning models to be trained without modification using the NVFlare federated
learning framework. Our results show increases in prostate cancer detection and
classification accuracy using a specialized neural network model and diverse
prostate biopsy data collected at two University of California research
hospitals, demonstrating the efficacy of our approach in adapting to different
datasets and improving MR-biomarker discovery. We open-source our FLtools
system, which can be easily adapted to other deep learning projects for medical
imaging. |
2,132 | Machine learning approaches for COVID-19 detection from chest X-ray imaging: A Systematic Review | There is a necessity to develop affordable, and reliable diagnostic tools,
which allow containing the COVID-19 spreading. Machine Learning (ML) algorithms
have been proposed to design support decision-making systems to assess chest
X-ray images, which have proven to be useful to detect and evaluate disease
progression. Many research articles are published around this subject, which
makes it difficult to identify the best approaches for future work. This paper
presents a systematic review of ML applied to COVID-19 detection using chest
X-ray images, aiming to offer a baseline for researchers in terms of methods,
architectures, databases, and current limitations. |
2,133 | Gradient Boosting Performs Low-Rank Gaussian Process Inference | This paper shows that gradient boosting based on symmetric decision trees can
be equivalently reformulated as a kernel method that converges to the solution
of a certain Kernel Ridgeless Regression problem. Thus, for low-rank kernels,
we obtain the convergence to a Gaussian Process' posterior mean, which, in
turn, allows us to easily transform gradient boosting into a sampler from the
posterior to provide better knowledge uncertainty estimates through Monte-Carlo
estimation of the posterior variance. We show that the proposed sampler allows
for better knowledge uncertainty estimates leading to improved out-of-domain
detection. |
2,134 | A Theoretical Understanding of Neural Network Compression from Sparse Linear Approximation | The goal of model compression is to reduce the size of a large neural network
while retaining a comparable performance. As a result, computation and memory
costs in resource-limited applications may be significantly reduced by dropping
redundant weights, neurons, or layers. There have been many model compression
algorithms proposed that provide impressive empirical success. However, a
theoretical understanding of model compression is still limited. One problem is
understanding if a network is more compressible than another of the same
structure. Another problem is quantifying how much one can prune a network with
theoretically guaranteed accuracy degradation. In this work, we propose to use
the sparsity-sensitive $\ell_q$-norm ($0<q<1$) to characterize compressibility
and provide a relationship between soft sparsity of the weights in the network
and the degree of compression with a controlled accuracy degradation bound. We
also develop adaptive algorithms for pruning each neuron in the network
informed by our theory. Numerical studies demonstrate the promising performance
of the proposed methods compared with standard pruning algorithms. |
2,135 | RadNet: Incident Prediction in Spatio-Temporal Road Graph Networks Using Traffic Forecasting | Efficient and accurate incident prediction in spatio-temporal systems is
critical to minimize service downtime and optimize performance. This work aims
to utilize historic data to predict and diagnose incidents using
spatio-temporal forecasting. We consider the specific use case of road traffic
systems where incidents take the form of anomalous events, such as accidents or
broken-down vehicles. To tackle this, we develop a neural model, called RadNet,
which forecasts system parameters such as average vehicle speeds for a future
timestep. As such systems largely follow daily or weekly periodicity, we
compare RadNet's predictions against historical averages to label incidents.
Unlike prior work, RadNet infers spatial and temporal trends in both
permutations, finally combining the dense representations before forecasting.
This facilitates informed inference and more accurate incident detection.
Experiments with two publicly available and a new road traffic dataset
demonstrate that the proposed model gives up to 8% higher prediction F1 scores
compared to the state-of-the-art methods. |
2,136 | Federated Offline Reinforcement Learning | Evidence-based or data-driven dynamic treatment regimes are essential for
personalized medicine, which can benefit from offline reinforcement learning
(RL). Although massive healthcare data are available across medical
institutions, they are prohibited from sharing due to privacy constraints.
Besides, heterogeneity exists in different sites. As a result, federated
offline RL algorithms are necessary and promising to deal with the problems. In
this paper, we propose a multi-site Markov decision process model which allows
both homogeneous and heterogeneous effects across sites. The proposed model
makes the analysis of the site-level features possible. We design the first
federated policy optimization algorithm for offline RL with sample complexity.
The proposed algorithm is communication-efficient and privacy-preserving, which
requires only a single round of communication interaction by exchanging summary
statistics. We give a theoretical guarantee for the proposed algorithm without
the assumption of sufficient action coverage, where the suboptimality for the
learned policies is comparable to the rate as if data is not distributed.
Extensive simulations demonstrate the effectiveness of the proposed algorithm.
The method is applied to a sepsis data set in multiple sites to illustrate its
use in clinical settings. |
2,137 | Optimal Solutions for Joint Beamforming and Antenna Selection: From Branch and Bound to Machine Learning | This work revisits the joint beamforming (BF) and antenna selection (AS)
problem, as well as its robust beamforming (RBF) version under imperfect
channel state information (CSI). Such problems arise in scenarios where the
number of the radio frequency (RF) chains is smaller than that of the antenna
elements at the transmitter, which has become a critical consideration in the
era of large-scale arrays. The joint (R)BF\&AS problem is a mixed integer and
nonlinear program, and thus finding {\it optimal solutions} is often costly, if
not outright impossible. The vast majority of the prior works tackled these
problems using continuous optimization-based approximations -- yet these
approximations do not ensure optimality or even feasibility of the solutions.
The main contribution of this work is threefold. First, an effective {\it
branch and bound} (B\&B) framework for solving the problems of interest is
proposed. Leveraging existing BF and RBF solvers, it is shown that the B\&B
framework guarantees global optimality of the considered problems. Second, to
expedite the potentially costly B\&B algorithm, a machine learning (ML)-based
scheme is proposed to help skip intermediate states of the B\&B search tree.
The learning model features a {\it graph neural network} (GNN)-based design
that is resilient to a commonly encountered challenge in wireless
communications, namely, the change of problem size (e.g., the number of users)
across the training and test stages. Third, comprehensive performance
characterizations are presented, showing that the GNN-based method retains the
global optimality of B\&B with provably reduced complexity, under reasonable
conditions. Numerical simulations also show that the ML-based acceleration can
often achieve an order-of-magnitude speedup relative to B\&B. |
2,138 | MammoDL: Mammographic Breast Density Estimation using Federated Learning | Assessing breast cancer risk from imaging remains a subjective process, in
which radiologists employ computer aided detection (CAD) systems or qualitative
visual assessment to estimate breast percent density (PD). More advanced
machine learning (ML) models have become the most promising way to quantify
breast cancer risk for early, accurate, and equitable diagnoses, but training
such models in medical research is often restricted to small,
single-institution data. Since patient demographics and imaging characteristics
may vary considerably across imaging sites, models trained on
single-institution data tend not to generalize well. In response to this
problem, MammoDL is proposed, an open-source software tool that leverages UNet
architecture to accurately estimate breast PD and complexity from digital
mammography (DM). With the Open Federated Learning (OpenFL) library, this
solution enables secure training on datasets across multiple institutions.
MammoDL is a leaner, more flexible model than its predecessors, boasting
improved generalization due to federation-enabled training on larger, more
representative datasets. |
2,139 | NeuGuard: Lightweight Neuron-Guided Defense against Membership Inference Attacks | Membership inference attacks (MIAs) against machine learning models can lead
to serious privacy risks for the training dataset used in the model training.
In this paper, we propose a novel and effective Neuron-Guided Defense method
named NeuGuard against membership inference attacks (MIAs). We identify a key
weakness in existing defense mechanisms against MIAs wherein they cannot
simultaneously defend against two commonly used neural network based MIAs,
indicating that these two attacks should be separately evaluated to assure the
defense effectiveness. We propose NeuGuard, a new defense approach that jointly
controls the output and inner neurons' activation with the object to guide the
model output of training set and testing set to have close distributions.
NeuGuard consists of class-wise variance minimization targeting restricting the
final output neurons and layer-wise balanced output control aiming to constrain
the inner neurons in each layer. We evaluate NeuGuard and compare it with
state-of-the-art defenses against two neural network based MIAs, five strongest
metric based MIAs including the newly proposed label-only MIA on three
benchmark datasets. Results show that NeuGuard outperforms the state-of-the-art
defenses by offering much improved utility-privacy trade-off, generality, and
overhead. |
2,140 | gDDIM: Generalized denoising diffusion implicit models | Our goal is to extend the denoising diffusion implicit model (DDIM) to
general diffusion models~(DMs). Instead of constructing a non-Markov noising
process as in the original DDIM paper, we examine the mechanism of DDIM from a
numerical perspective. We discover that the DDIM can be obtained by using some
specific approximations of the score when solving the corresponding stochastic
differential equation. We present an interpretation of the accelerating effects
of DDIM that also explains the advantages of a deterministic sampling scheme
over the stochastic one for fast sampling. Building on this insight, we extend
DDIM to general DMs with a small but delicate modification in parameterizing
the score network. When applied to the critically-damped Langevin diffusion
model, a new type of diffusion model proposed recently by augmenting the
diffusion process with velocity, our algorithm achieves an FID score of 2.28,
on CIFAR10, with only 50 number of score function evaluations~(NFEs) and an FID
score of 2.87 with only 27 NFEs, better than all existing methods with the same
NFEs. Code is available at https://github.com/qsh-zh/gDDIM |
2,141 | Parameter Convex Neural Networks | Deep learning utilizing deep neural networks (DNNs) has achieved a lot of
success recently in many important areas such as computer vision, natural
language processing, and recommendation systems. The lack of convexity for DNNs
has been seen as a major disadvantage of many optimization methods, such as
stochastic gradient descent, which greatly reduces the genelization of neural
network applications. We realize that the convexity make sense in the neural
network and propose the exponential multilayer neural network (EMLP), a class
of parameter convex neural network (PCNN) which is convex with regard to the
parameters of the neural network under some conditions that can be realized.
Besides, we propose the convexity metric for the two-layer EGCN and test the
accuracy when the convexity metric changes. For late experiments, we use the
same architecture to make the exponential graph convolutional network (EGCN)
and do the experiment on the graph classificaion dataset in which our model
EGCN performs better than the graph convolutional network (GCN) and the graph
attention network (GAT). |
2,142 | Communication-Efficient Robust Federated Learning with Noisy Labels | Federated learning (FL) is a promising privacy-preserving machine learning
paradigm over distributed located data. In FL, the data is kept locally by each
user. This protects the user privacy, but also makes the server difficult to
verify data quality, especially if the data are correctly labeled. Training
with corrupted labels is harmful to the federated learning task; however,
little attention has been paid to FL in the case of label noise. In this paper,
we focus on this problem and propose a learning-based reweighting approach to
mitigate the effect of noisy labels in FL. More precisely, we tuned a weight
for each training sample such that the learned model has optimal generalization
performance over a validation set. More formally, the process can be formulated
as a Federated Bilevel Optimization problem. Bilevel optimization problem is a
type of optimization problem with two levels of entangled problems. The
non-distributed bilevel problems have witnessed notable progress recently with
new efficient algorithms. However, solving bilevel optimization problems under
the Federated Learning setting is under-investigated. We identify that the high
communication cost in hypergradient evaluation is the major bottleneck. So we
propose \textit{Comm-FedBiO} to solve the general Federated Bilevel
Optimization problems; more specifically, we propose two
communication-efficient subroutines to estimate the hypergradient. Convergence
analysis of the proposed algorithms is also provided. Finally, we apply the
proposed algorithms to solve the noisy label problem. Our approach has shown
superior performance on several real-world datasets compared to various
baselines. |
2,143 | PhML-DyR: A Physics-Informed ML framework for Dynamic Reconfiguration in Power Systems | A transformation of the US electricity sector is underway with aggressive
targets to achieve 100% carbon pollution-free electricity by 2035. To achieve
this objective while maintaining a safe and reliable power grid, new operating
paradigms are needed, of computationally fast and accurate decision making in a
dynamic and uncertain environment. We propose a novel physics-informed machine
learning framework for the decision of dynamic grid reconfiguration (PhML-DyR),
a key task in power systems. Dynamic reconfiguration (DyR) is a process by
which switch-states are dynamically set so as to lead to an optimal grid
topology that minimizes line losses. To address the underlying computational
complexities of NP-hardness due to the mixed nature of the decision variables,
we propose the use of physics-informed ML (PhML) which integrates both
operating constraints and topological and connectivity constraints into a
neural network framework. Our PhML approach learns to simultaneously optimize
grid topology and generator dispatch to meet loads, increase efficiency, and
remain within safe operating limits. We demonstrate the effectiveness of
PhML-DyR on a canonical grid, showing a reduction in electricity loss by 23%,
and improved voltage profiles. We also show a reduction in constraint
violations by an order of magnitude as well as in training time using PhML-DyR. |
2,144 | A Simplified Un-Supervised Learning Based Approach for Ink Mismatch Detection in Handwritten Hyper-Spectral Document Images | Hyper-spectral imaging has become the latest trend in the field of optical
imaging systems. Among various other applications, hyper-spectral imaging has
been widely used for analysis of printed and handwritten documents. This paper
proposes an efficient technique for estimating the number of different but
visibly similar inks present in a Hyper spectral Document Image. Our approach
is based on un-supervised learning and does not require any prior knowledge of
the dataset. The algorithm was tested on the iVision HHID dataset and has
achieved comparable results with the state of the algorithms present in the
literature. This work can prove to be effective when employed during the early
stages of forgery detection in Hyper-spectral Document Images. |
2,145 | Rare event failure test case generation in Learning-Enabled-Controllers | Machine learning models have prevalent applications in many real-world
problems, which increases the importance of correctness in the behaviour of
these trained models. Finding a good test case that can reveal the potential
failure in these trained systems can help to retrain these models to increase
their correctness. For a well-trained model, the occurrence of a failure is
rare. Consequently, searching these rare scenarios by evaluating each sample in
input search space or randomized search would be costly and sometimes
intractable due to large search space, limited computational resources, and
available time. In this paper, we tried to address this challenge of finding
these failure scenarios faster than traditional randomized search. The central
idea of our approach is to separate the input data space in region of high
failure probability and region of low/minimal failure probability based on the
observation made by training data, data drawn from real-world statistics, and
knowledge from a domain expert. Using these information, we can design a
generative model from which we can generate scenarios that have a high
likelihood to reveal the potential failure. We evaluated this approach on two
different experimental scenarios and able to speed up the discovery of such
failures a thousand-fold faster than the traditional randomized search. |
2,146 | Memorization-Dilation: Modeling Neural Collapse Under Noise | The notion of neural collapse refers to several emergent phenomena that have
been empirically observed across various canonical classification problems.
During the terminal phase of training a deep neural network, the feature
embedding of all examples of the same class tend to collapse to a single
representation, and the features of different classes tend to separate as much
as possible. Neural collapse is often studied through a simplified model,
called the unconstrained feature representation, in which the model is assumed
to have "infinite expressivity" and can map each data point to any arbitrary
representation. In this work, we propose a more realistic variant of the
unconstrained feature representation that takes the limited expressivity of the
network into account. Empirical evidence suggests that the memorization of
noisy data points leads to a degradation (dilation) of the neural collapse.
Using a model of the memorization-dilation (M-D) phenomenon, we show one
mechanism by which different losses lead to different performances of the
trained network on noisy data. Our proofs reveal why label smoothing, a
modification of cross-entropy empirically observed to produce a regularization
effect, leads to improved generalization in classification tasks. |
2,147 | Model-based Offline Imitation Learning with Non-expert Data | Although Behavioral Cloning (BC) in theory suffers compounding errors, its
scalability and simplicity still makes it an attractive imitation learning
algorithm. In contrast, imitation approaches with adversarial training
typically does not share the same problem, but necessitates interactions with
the environment. Meanwhile, most imitation learning methods only utilises
optimal datasets, which could be significantly more expensive to obtain than
its suboptimal counterpart. A question that arises is, can we utilise the
suboptimal dataset in a principled manner, which otherwise would have been
idle? We propose a scalable model-based offline imitation learning algorithmic
framework that leverages datasets collected by both suboptimal and optimal
policies, and show that its worst case suboptimality becomes linear in the time
horizon with respect to the expert samples. We empirically validate our
theoretical results and show that the proposed method \textit{always}
outperforms BC in the low data regime on simulated continuous control domains |
2,148 | Defending Adversarial Examples by Negative Correlation Ensemble | The security issues in DNNs, such as adversarial examples, have attracted
much attention. Adversarial examples refer to the examples which are capable to
induce the DNNs return completely predictions by introducing carefully designed
perturbations. Obviously, adversarial examples bring great security risks to
the development of deep learning. Recently, Some defense approaches against
adversarial examples have been proposed, however, in our opinion, the
performance of these approaches are still limited. In this paper, we propose a
new ensemble defense approach named the Negative Correlation Ensemble (NCEn),
which achieves compelling results by introducing gradient directions and
gradient magnitudes of each member in the ensemble negatively correlated and at
the same time, reducing the transferability of adversarial examples among them.
Extensive experiments have been conducted, and the results demonstrate that
NCEn can improve the adversarial robustness of ensembles effectively. |
2,149 | Federated Learning with GAN-based Data Synthesis for Non-IID Clients | Federated learning (FL) has recently emerged as a popular privacy-preserving
collaborative learning paradigm. However, it suffers from the non-independent
and identically distributed (non-IID) data among clients. In this paper, we
propose a novel framework, named Synthetic Data Aided Federated Learning
(SDA-FL), to resolve this non-IID challenge by sharing synthetic data.
Specifically, each client pretrains a local generative adversarial network
(GAN) to generate differentially private synthetic data, which are uploaded to
the parameter server (PS) to construct a global shared synthetic dataset. To
generate confident pseudo labels for the synthetic dataset, we also propose an
iterative pseudo labeling mechanism performed by the PS. A combination of the
local private dataset and synthetic dataset with confident pseudo labels leads
to nearly identical data distributions among clients, which improves the
consistency among local models and benefits the global aggregation. Extensive
experiments evidence that the proposed framework outperforms the baseline
methods by a large margin in several benchmark datasets under both the
supervised and semi-supervised settings. |
2,150 | Soft-mask: Adaptive Substructure Extractions for Graph Neural Networks | For learning graph representations, not all detailed structures within a
graph are relevant to the given graph tasks. Task-relevant structures can be
$localized$ or $sparse$ which are only involved in subgraphs or characterized
by the interactions of subgraphs (a hierarchical perspective). A graph neural
network should be able to efficiently extract task-relevant structures and be
invariant to irrelevant parts, which is challenging for general message passing
GNNs. In this work, we propose to learn graph representations from a sequence
of subgraphs of the original graph to better capture task-relevant
substructures or hierarchical structures and skip $noisy$ parts. To this end,
we design soft-mask GNN layer to extract desired subgraphs through the mask
mechanism. The soft-mask is defined in a continuous space to maintain the
differentiability and characterize the weights of different parts. Compared
with existing subgraph or hierarchical representation learning methods and
graph pooling operations, the soft-mask GNN layer is not limited by the fixed
sample or drop ratio, and therefore is more flexible to extract subgraphs with
arbitrary sizes. Extensive experiments on public graph benchmarks show that
soft-mask mechanism brings performance improvements. And it also provides
interpretability where visualizing the values of masks in each layer allows us
to have an insight into the structures learned by the model. |
2,151 | Learning to Generate Levels by Imitating Evolution | Search-based procedural content generation (PCG) is a well-known method used
for level generation in games. Its key advantage is that it is generic and able
to satisfy functional constraints. However, due to the heavy computational
costs to run these algorithms online, search-based PCG is rarely utilized for
real-time generation. In this paper, we introduce a new type of iterative level
generator using machine learning. We train a model to imitate the evolutionary
process and use the model to generate levels. This trained model is able to
modify noisy levels sequentially to create better levels without the need for a
fitness function during inference. We evaluate our trained models on a 2D maze
generation task. We compare several different versions of the method: training
the models either at the end of evolution (normal evolution) or every 100
generations (assisted evolution) and using the model as a mutation function
during evolution. Using the assisted evolution process, the final trained
models are able to generate mazes with a success rate of 99% and high diversity
of 86%. This work opens the door to a new way of learning level generators
guided by the evolutionary process and perhaps will increase the adoption of
search-based PCG in the game industry. |
2,152 | DRAformer: Differentially Reconstructed Attention Transformer for Time-Series Forecasting | Time-series forecasting plays an important role in many real-world scenarios,
such as equipment life cycle forecasting, weather forecasting, and traffic flow
forecasting. It can be observed from recent research that a variety of
transformer-based models have shown remarkable results in time-series
forecasting. However, there are still some issues that limit the ability of
transformer-based models on time-series forecasting tasks: (i) learning
directly on raw data is susceptible to noise due to its complex and unstable
feature representation; (ii) the self-attention mechanisms pay insufficient
attention to changing features and temporal dependencies. In order to solve
these two problems, we propose a transformer-based differentially reconstructed
attention model DRAformer. Specifically, DRAformer has the following
innovations: (i) learning against differenced sequences, which preserves clear
and stable sequence features by differencing and highlights the changing
properties of sequences; (ii) the reconstructed attention: integrated distance
attention exhibits sequential distance through a learnable Gaussian kernel,
distributed difference attention calculates distribution difference by mapping
the difference sequence to the adaptive feature space, and the combination of
the two effectively focuses on the sequences with prominent associations; (iii)
the reconstructed decoder input, which extracts sequence features by
integrating variation information and temporal correlations, thereby obtaining
a more comprehensive sequence representation. Extensive experiments on four
large-scale datasets demonstrate that DRAformer outperforms state-of-the-art
baselines. |
2,153 | Discovery and density estimation of latent confounders in Bayesian networks with evidence lower bound | Discovering and parameterising latent confounders represent important and
challenging problems in causal structure learning and density estimation
respectively. In this paper, we focus on both discovering and learning the
distribution of latent confounders. This task requires solutions that come from
different areas of statistics and machine learning. We combine elements of
variational Bayesian methods, expectation-maximisation, hill-climbing search,
and structure learning under the assumption of causal insufficiency. We propose
two learning strategies; one that maximises model selection accuracy, and
another that improves computational efficiency in exchange for minor reductions
in accuracy. The former strategy is suitable for small networks and the latter
for moderate size networks. Both learning strategies perform well relative to
existing solutions. |
2,154 | Scientific Inference With Interpretable Machine Learning: Analyzing Models to Learn About Real-World Phenomena | Interpretable machine learning (IML) is concerned with the behavior and the
properties of machine learning models. Scientists, however, are only interested
in the model as a gateway to understanding the modeled phenomenon. We show how
to develop IML methods such that they allow insight into relevant phenomenon
properties. We argue that current IML research conflates two goals of
model-analysis -- model audit and scientific inference. Thereby, it remains
unclear if model interpretations have corresponding phenomenon interpretation.
Building on statistical decision theory, we show that ML model analysis allows
to describe relevant aspects of the joint data probability distribution. We
provide a five-step framework for constructing IML descriptors that can help in
addressing scientific questions, including a natural way to quantify epistemic
uncertainty. Our phenomenon-centric approach to IML in science clarifies: the
opportunities and limitations of IML for inference; that conditional not
marginal sampling is required; and, the conditions under which we can trust IML
methods. |
2,155 | Bilateral Dependency Optimization: Defending Against Model-inversion Attacks | Through using only a well-trained classifier, model-inversion (MI) attacks
can recover the data used for training the classifier, leading to the privacy
leakage of the training data. To defend against MI attacks, previous work
utilizes a unilateral dependency optimization strategy, i.e., minimizing the
dependency between inputs (i.e., features) and outputs (i.e., labels) during
training the classifier. However, such a minimization process conflicts with
minimizing the supervised loss that aims to maximize the dependency between
inputs and outputs, causing an explicit trade-off between model robustness
against MI attacks and model utility on classification tasks. In this paper, we
aim to minimize the dependency between the latent representations and the
inputs while maximizing the dependency between latent representations and the
outputs, named a bilateral dependency optimization (BiDO) strategy. In
particular, we use the dependency constraints as a universally applicable
regularizer in addition to commonly used losses for deep neural networks (e.g.,
cross-entropy), which can be instantiated with appropriate dependency criteria
according to different tasks. To verify the efficacy of our strategy, we
propose two implementations of BiDO, by using two different dependency
measures: BiDO with constrained covariance (BiDO-COCO) and BiDO with
Hilbert-Schmidt Independence Criterion (BiDO-HSIC). Experiments show that BiDO
achieves the state-of-the-art defense performance for a variety of datasets,
classifiers, and MI attacks while suffering a minor classification-accuracy
drop compared to the well-trained classifier with no defense, which lights up a
novel road to defend against MI attacks. |
2,156 | Monitoring and Proactive Management of QoS Levels in Pervasive Applications | The advent of Edge Computing (EC) as a promising paradigm that provides
multiple computation and analytics capabilities close to data sources opens new
pathways for novel applications. Nonetheless, the limited computational
capabilities of EC nodes and the expectation of ensuring high levels of QoS
during tasks execution impose strict requirements for innovative management
approaches. Motivated by the need of maintaining a minimum level of QoS during
EC nodes functioning, we elaborate a distributed and intelligent
decision-making approach for tasks scheduling. Our aim is to enhance the
behavior of EC nodes making them capable of securing high QoS levels. We
propose that nodes continuously monitor QoS levels and systematically evaluate
the probability of violating them to proactively decide some tasks to be
offloaded to peer nodes or Cloud. We present, describe and evaluate the
proposed scheme through multiple experimental scenarios revealing its
performance and the benefits of the envisioned monitoring mechanism when
serving processing requests in very dynamic environments like the EC. |
2,157 | Reducing Capacity Gap in Knowledge Distillation with Review Mechanism for Crowd Counting | The lightweight crowd counting models, in particular knowledge distillation
(KD) based models, have attracted rising attention in recent years due to their
superiority on computational efficiency and hardware requirement. However,
existing KD based models usually suffer from the capacity gap issue, resulting
in the performance of the student network being limited by the teacher network.
In this paper, we address this issue by introducing a novel review mechanism
following KD models, motivated by the review mechanism of human-beings during
the study. Thus, the proposed model is dubbed ReviewKD. The proposed model
consists of an instruction phase and a review phase, where we firstly exploit a
well-trained heavy teacher network to transfer its latent feature to a
lightweight student network in the instruction phase, then in the review phase
yield a refined estimate of the density map based on the learned feature
through a review mechanism. The effectiveness of ReviewKD is demonstrated by a
set of experiments over six benchmark datasets via comparing to the
state-of-the-art models. Numerical results show that ReviewKD outperforms
existing lightweight models for crowd counting, and can effectively alleviate
the capacity gap issue, and particularly has the performance beyond the teacher
network. Besides the lightweight models, we also show that the suggested review
mechanism can be used as a plug-and-play module to further boost the
performance of a kind of heavy crowd counting models without modifying the
neural network architecture and introducing any additional model parameter. |
2,158 | Svadhyaya system for the Second Diagnosing COVID-19 using Acoustics Challenge 2021 | This report describes the system used for detecting COVID-19 positives using
three different acoustic modalities, namely speech, breathing, and cough in the
second DiCOVA challenge. The proposed system is based on the combination of 4
different approaches, each focusing more on one aspect of the problem, and
reaches the blind test AUCs of 86.41, 77.60, and 84.55, in the breathing,
cough, and speech tracks, respectively, and the AUC of 85.37 in the fusion of
these three tracks. |
2,159 | Hierarchical Conditional Variational Autoencoder Based Acoustic Anomaly Detection | This paper aims to develop an acoustic signal-based unsupervised anomaly
detection method for automatic machine monitoring. Existing approaches such as
deep autoencoder (DAE), variational autoencoder (VAE), conditional variational
autoencoder (CVAE) etc. have limited representation capabilities in the latent
space and, hence, poor anomaly detection performance. Different models have to
be trained for each different kind of machines to accurately perform the
anomaly detection task. To solve this issue, we propose a new method named as
hierarchical conditional variational autoencoder (HCVAE). This method utilizes
available taxonomic hierarchical knowledge about industrial facility to refine
the latent space representation. This knowledge helps model to improve the
anomaly detection performance as well. We demonstrated the generalization
capability of a single HCVAE model for different types of machines by using
appropriate conditions. Additionally, to show the practicability of the
proposed approach, (i) we evaluated HCVAE model on different domain and (ii) we
checked the effect of partial hierarchical knowledge. Our results show that
HCVAE method validates both of these points, and it outperforms the baseline
system on anomaly detection task by utmost 15 % on the AUC score metric. |
2,160 | A General framework for PAC-Bayes Bounds for Meta-Learning | Meta learning automatically infers an inductive bias, that includes the
hyperparameter of the base-learning algorithm, by observing data from a finite
number of related tasks. This paper studies PAC-Bayes bounds on meta
generalization gap. The meta-generalization gap comprises two sources of
generalization gaps: the environment-level and task-level gaps resulting from
observation of a finite number of tasks and data samples per task,
respectively. In this paper, by upper bounding arbitrary convex functions,
which link the expected and empirical losses at the environment and also
per-task levels, we obtain new PAC-Bayes bounds. Using these bounds, we develop
new PAC-Bayes meta-learning algorithms. Numerical examples demonstrate the
merits of the proposed novel bounds and algorithm in comparison to prior
PAC-Bayes bounds for meta-learning. |
2,161 | Enhancing Explainability of Hyperparameter Optimization via Bayesian Algorithm Execution | Despite all the benefits of automated hyperparameter optimization (HPO), most
modern HPO algorithms are black-boxes themselves. This makes it difficult to
understand the decision process which lead to the selected configuration,
reduces trust in HPO, and thus hinders its broad adoption. Here, we study the
combination of HPO with interpretable machine learning (IML) methods such as
partial dependence plots. However, if such methods are naively applied to the
experimental data of the HPO process in a post-hoc manner, the underlying
sampling bias of the optimizer can distort interpretations. We propose a
modified HPO method which efficiently balances the search for the global
optimum w.r.t. predictive performance and the reliable estimation of IML
explanations of an underlying black-box function by coupling Bayesian
optimization and Bayesian Algorithm Execution. On benchmark cases of both
synthetic objectives and HPO of a neural network, we demonstrate that our
method returns more reliable explanations of the underlying black-box without a
loss of optimization performance. |
2,162 | A Dataset and Benchmark for Automatically Answering and Generating Machine Learning Final Exams | Can a machine learn machine learning? We propose to answer this question
using the same criteria we use to answer a similar question: can a human learn
machine learning? We automatically answer MIT final exams in Introduction to
Machine Learning at a human level. The course is a large undergraduate class
with around five hundred students each semester. Recently, program synthesis
and few-shot learning solved university-level problem set questions in
mathematics and STEM courses at a human level. In this work, we solve questions
from final exams that differ from problem sets in several ways: the questions
are longer, have multiple parts, are more complicated, and span a broader set
of topics. We provide a new dataset and benchmark of questions from eight MIT
Introduction to Machine Learning final exams between Fall 2017 and Spring 2022
and provide code for automatically answering these questions and generating new
questions. We perform ablation studies comparing zero-shot learning with
few-shot learning, chain-of-thought prompting, GPT-3 pre-trained on text and
Codex fine-tuned on code on a range of machine learning topics and find that
few-shot learning methods perform best. We make our data and code publicly
available for the machine learning community. |
2,163 | ACMP: Allen-Cahn Message Passing for Graph Neural Networks with Particle Phase Transition | Neural message passing is a basic feature extraction unit for
graph-structured data that takes account of the impact of neighboring node
features in network propagation from one layer to the next. We model such
process by an interacting particle system with attractive and repulsive forces
and the Allen-Cahn force arising in the modeling of phase transition. The
system is a reaction-diffusion process which can separate particles to
different clusters. This induces an Allen-Cahn message passing (ACMP) for graph
neural networks where the numerical iteration for the solution constitutes the
message passing propagation. The mechanism behind ACMP is phase transition of
particles which enables the formation of multi-clusters and thus GNNs
prediction for node classification. ACMP can propel the network depth to
hundreds of layers with theoretically proven strictly positive lower bound of
the Dirichlet energy. It thus provides a deep model of GNNs which circumvents
the common GNN problem of oversmoothing. Experiments for various real node
classification datasets, with possible high homophily difficulty, show the GNNs
with ACMP can achieve state of the art performance with no decay of Dirichlet
energy. |
2,164 | Learned reconstruction with convergence guarantees | In recent years, deep learning has achieved remarkable empirical success for
image reconstruction. This has catalyzed an ongoing quest for precise
characterization of correctness and reliability of data-driven methods in
critical use-cases, for instance in medical imaging. Notwithstanding the
excellent performance and efficacy of deep learning-based methods, concerns
have been raised regarding their stability, or lack thereof, with serious
practical implications. Significant advances have been made in recent years to
unravel the inner workings of data-driven image recovery methods, challenging
their widely perceived black-box nature. In this article, we will specify
relevant notions of convergence for data-driven image reconstruction, which
will form the basis of a survey of learned methods with mathematically rigorous
reconstruction guarantees. An example that is highlighted is the role of ICNN,
offering the possibility to combine the power of deep learning with classical
convex regularization theory for devising methods that are provably convergent.
This survey article is aimed at both methodological researchers seeking to
advance the frontiers of our understanding of data-driven image reconstruction
methods as well as practitioners, by providing an accessible description of
convergence concepts and by placing some of the existing empirical practices on
a solid mathematical foundation. |
2,165 | Semi-Supervised Hierarchical Graph Classification | Node classification and graph classification are two graph learning problems
that predict the class label of a node and the class label of a graph
respectively. A node of a graph usually represents a real-world entity, e.g., a
user in a social network, or a document in a document citation network. In this
work, we consider a more challenging but practically useful setting, in which a
node itself is a graph instance. This leads to a hierarchical graph perspective
which arises in many domains such as social network, biological network and
document collection. We study the node classification problem in the
hierarchical graph where a 'node' is a graph instance. As labels are usually
limited, we design a novel semi-supervised solution named SEAL-CI. SEAL-CI
adopts an iterative framework that takes turns to update two modules, one
working at the graph instance level and the other at the hierarchical graph
level. To enforce a consistency among different levels of hierarchical graph,
we propose the Hierarchical Graph Mutual Information (HGMI) and further present
a way to compute HGMI with theoretical guarantee. We demonstrate the
effectiveness of this hierarchical graph modeling and the proposed SEAL-CI
method on text and social network data. |
2,166 | Multi-instrument Music Synthesis with Spectrogram Diffusion | An ideal music synthesizer should be both interactive and expressive,
generating high-fidelity audio in realtime for arbitrary combinations of
instruments and notes. Recent neural synthesizers have exhibited a tradeoff
between domain-specific models that offer detailed control of only specific
instruments, or raw waveform models that can train on all of music but with
minimal control and slow generation.
In this work, we focus on a middle ground of neural synthesizers that can
generate audio from MIDI sequences with arbitrary combinations of instruments
in realtime. This enables training on a wide range of transcription datasets
with a single model, which in turn offers note-level control of composition and
instrumentation across a wide range of instruments.
We use a simple two-stage process: MIDI to spectrograms with an
encoder-decoder Transformer, then spectrograms to audio with a generative
adversarial network (GAN) spectrogram inverter. We compare training the decoder
as an autoregressive model and as a Denoising Diffusion Probabilistic Model
(DDPM) and find that the DDPM approach is superior both qualitatively and as
measured by audio reconstruction and Fr\'echet distance metrics.
Given the interactivity and generality of this approach, we find this to be a
promising first step towards interactive and expressive neural synthesis for
arbitrary combinations of instruments and notes. |
2,167 | Rethinking the Defense Against Free-rider Attack From the Perspective of Model Weight Evolving Frequency | Federated learning (FL) is a distributed machine learning approach where
multiple clients collaboratively train a joint model without exchanging their
data. Despite FL's unprecedented success in data privacy-preserving, its
vulnerability to free-rider attacks has attracted increasing attention.
Existing defenses may be ineffective against highly camouflaged or high
percentages of free riders. To address these challenges, we reconsider the
defense from a novel perspective, i.e., model weight evolving
frequency.Empirically, we gain a novel insight that during the FL's training,
the model weight evolving frequency of free-riders and that of benign clients
are significantly different. Inspired by this insight, we propose a novel
defense method based on the model Weight Evolving Frequency, referred to as
WEF-Defense.Specifically, we first collect the weight evolving frequency
(defined as WEF-Matrix) during local training. For each client, it uploads the
local model's WEF-Matrix to the server together with its model weight for each
iteration. The server then separates free-riders from benign clients based on
the difference in the WEF-Matrix. Finally, the server uses a personalized
approach to provide different global models for corresponding clients.
Comprehensive experiments conducted on five datasets and five models
demonstrate that WEF-Defense achieves better defense effectiveness than the
state-of-the-art baselines. |
2,168 | Squeeze All: Novel Estimator and Self-Normalized Bound for Linear Contextual Bandits | We propose a novel algorithm for linear contextual bandits with $O(\sqrt{dT
\log T})$ regret bound, where $d$ is the dimension of contexts and $T$ is the
time horizon. Our proposed algorithm is equipped with a novel estimator in
which exploration is embedded through explicit randomization. Depending on the
randomization, our proposed estimator takes contribution either from contexts
of all arms or from selected contexts. We establish a self-normalized bound for
our estimator, which allows a novel decomposition of the cumulative regret into
additive dimension-dependent terms instead of multiplicative terms. We also
prove a novel lower bound of $\Omega(\sqrt{dT})$ under our problem setting.
Hence, the regret of our proposed algorithm matches the lower bound up to
logarithmic factors. The numerical experiments support the theoretical
guarantees and show that our proposed method outperforms the existing linear
bandit algorithms. |
2,169 | Feature Selection using e-values | In the context of supervised parametric models, we introduce the concept of
e-values. An e-value is a scalar quantity that represents the proximity of the
sampling distribution of parameter estimates in a model trained on a subset of
features to that of the model trained on all features (i.e. the full model).
Under general conditions, a rank ordering of e-values separates models that
contain all essential features from those that do not.
The e-values are applicable to a wide range of parametric models. We use data
depths and a fast resampling-based algorithm to implement a feature selection
procedure using e-values, providing consistency results. For a $p$-dimensional
feature space, this procedure requires fitting only the full model and
evaluating $p+1$ models, as opposed to the traditional requirement of fitting
and evaluating $2^p$ models. Through experiments across several model settings
and synthetic and real datasets, we establish that the e-values method as a
promising general alternative to existing model-specific methods of feature
selection. |
2,170 | Learning Imbalanced Datasets with Maximum Margin Loss | A learning algorithm referred to as Maximum Margin (MM) is proposed for
considering the class-imbalance data learning issue: the trained model tends to
predict the majority of classes rather than the minority ones. That is,
underfitting for minority classes seems to be one of the challenges of
generalization. For a good generalization of the minority classes, we design a
new Maximum Margin (MM) loss function, motivated by minimizing a margin-based
generalization bound through the shifting decision bound. The
theoretically-principled label-distribution-aware margin (LDAM) loss was
successfully applied with prior strategies such as re-weighting or re-sampling
along with the effective training schedule. However, they did not investigate
the maximum margin loss function yet. In this study, we investigate the
performances of two types of hard maximum margin-based decision boundary shift
with LDAM's training schedule on artificially imbalanced CIFAR-10/100 for fair
comparisons and effectiveness. |
2,171 | Fast building segmentation from satellite imagery and few local labels | Innovations in computer vision algorithms for satellite image analysis can
enable us to explore global challenges such as urbanization and land use change
at the planetary level. However, domain shift problems are a common occurrence
when trying to replicate models that drive these analyses to new areas,
particularly in the developing world. If a model is trained with imagery and
labels from one location, then it usually will not generalize well to new
locations where the content of the imagery and data distributions are
different. In this work, we consider the setting in which we have a single
large satellite imagery scene over which we want to solve an applied problem --
building footprint segmentation. Here, we do not necessarily need to worry
about creating a model that generalizes past the borders of our scene but can
instead train a local model. We show that surprisingly few labels are needed to
solve the building segmentation problem with very high-resolution (0.5m/px)
satellite imagery with this setting in mind. Our best model trained with just
527 sparse polygon annotations (an equivalent of 1500 x 1500 densely labeled
pixels) has a recall of 0.87 over held out footprints and a R2 of 0.93 on the
task of counting the number of buildings in 200 x 200-meter windows. We apply
our models over high-resolution imagery in Amman, Jordan in a case study on
urban change detection. |
2,172 | An application of neural networks to a problem in knot theory and group theory (untangling braids) | We report on our success on solving the problem of untangling braids up to
length 20 and width 4. We use feed-forward neural networks in the framework of
reinforcement learning to train the agent to choose Reidemeister moves to
untangle braids in the minimal number of moves. |
2,173 | Object Detection, Recognition, Deep Learning, and the Universal Law of Generalization | Object detection and recognition are fundamental functions underlying the
success of species. Because the appearance of an object exhibits a large
variability, the brain has to group these different stimuli under the same
object identity, a process of generalization. Does the process of
generalization follow some general principles or is it an ad-hoc
"bag-of-tricks"? The Universal Law of Generalization provided evidence that
generalization follows similar properties across a variety of species and
tasks. Here we test the hypothesis that the internal representations underlying
generalization reflect the natural properties of object detection and
recognition in our environment rather than the specifics of the system solving
these problems. By training a deep-neural-network with images of "clear" and
"camouflaged" animals, we found that with a proper choice of category
prototypes, the generalization functions are monotone decreasing, similar to
the generalization functions of biological systems. Our findings support the
hypothesis of the study. |
2,174 | Anchor-Changing Regularized Natural Policy Gradient for Multi-Objective Reinforcement Learning | We study policy optimization for Markov decision processes (MDPs) with
multiple reward value functions, which are to be jointly optimized according to
given criteria such as proportional fairness (smooth concave scalarization),
hard constraints (constrained MDP), and max-min trade-off. We propose an
Anchor-changing Regularized Natural Policy Gradient (ARNPG) framework, which
can systematically incorporate ideas from well-performing first-order methods
into the design of policy optimization algorithms for multi-objective MDP
problems. Theoretically, the designed algorithms based on the ARNPG framework
achieve $\tilde{O}(1/T)$ global convergence with exact gradients. Empirically,
the ARNPG-guided algorithms also demonstrate superior performance compared to
some existing policy gradient-based approaches in both exact gradients and
sample-based scenarios. |
2,175 | Cross-TOP: Zero-Shot Cross-Schema Task-Oriented Parsing | Deep learning methods have enabled task-oriented semantic parsing of
increasingly complex utterances. However, a single model is still typically
trained and deployed for each task separately, requiring labeled training data
for each, which makes it challenging to support new tasks, even within a single
business vertical (e.g., food-ordering or travel booking). In this paper we
describe Cross-TOP (Cross-Schema Task-Oriented Parsing), a zero-shot method for
complex semantic parsing in a given vertical. By leveraging the fact that user
requests from the same vertical share lexical and semantic similarities, a
single cross-schema parser is trained to service an arbitrary number of tasks,
seen or unseen, within a vertical. We show that Cross-TOP can achieve high
accuracy on a previously unseen task without requiring any additional training
data, thereby providing a scalable way to bootstrap semantic parsers for new
tasks. As part of this work we release the FoodOrdering dataset, a
task-oriented parsing dataset in the food-ordering vertical, with utterances
and annotations derived from five schemas, each from a different restaurant
menu. |
2,176 | Massively Increasing the number of Antibody-Virus Interactions across Studies | A central challenge in every field of biology is to use existing measurements
to predict the outcomes of future experiments. In this work, we consider the
wealth of antibody inhibition data against variants of the influenza virus. Due
to this virus's genetic diversity and evolvability, the variants examined in
one study will often have little-to-no overlap with other studies, making it
difficult to discern common patterns or unify datasets for further analysis. To
that end, we develop a computational framework that predicts how an antibody or
serum would inhibit any variant from any other study. We use this framework to
greatly expand 7 influenza datasets utilizing hemagglutination inhibition,
validating our method upon 200,000 existing measurements and predicting more
than 2,000,000 new values along with their prediction uncertainties. This
data-driven approach does not require any information beyond each virus's name
and measurements, and even datasets with as few as 5 viruses can be expanded,
making this approach widely applicable. Future influenza studies using
hemagglutination inhibition can directly utilize our curated datasets to
predict newly measured antibody responses against ~80 H3N2 influenza viruses
from 1968-2011, whereas immunological studies utilizing other viruses or a
different assay only need to find a single partially-overlapping dataset to
extend their work. In essence, this approach enables a shift in perspective
when analyzing data from "what you see is what you get" into "what anyone sees
is what everyone gets." |
2,177 | Synthetic Over-sampling for Imbalanced Node Classification with Graph Neural Networks | In recent years, graph neural networks (GNNs) have achieved state-of-the-art
performance for node classification. However, most existing GNNs would suffer
from the graph imbalance problem. In many real-world scenarios, node classes
are imbalanced, with some majority classes making up most parts of the graph.
The message propagation mechanism in GNNs would further amplify the dominance
of those majority classes, resulting in sub-optimal classification performance.
In this work, we seek to address this problem by generating pseudo instances of
minority classes to balance the training data, extending previous
over-sampling-based techniques. This task is non-trivial, as those techniques
are designed with the assumption that instances are independent. Neglection of
relation information would complicate this oversampling process. Furthermore,
the node classification task typically takes the semi-supervised setting with
only a few labeled nodes, providing insufficient supervision for the generation
of minority instances. Generated new nodes of low quality would harm the
trained classifier. In this work, we address these difficulties by synthesizing
new nodes in a constructed embedding space, which encodes both node attributes
and topology information. Furthermore, an edge generator is trained
simultaneously to model the graph structure and provide relations for new
samples. To further improve the data efficiency, we also explore synthesizing
mixed ``in-between'' nodes to utilize nodes from the majority class in this
over-sampling process. Experiments on real-world datasets validate the
effectiveness of our proposed framework. |
2,178 | Memory Classifiers: Two-stage Classification for Robustness in Machine Learning | The performance of machine learning models can significantly degrade under
distribution shifts of the data. We propose a new method for classification
which can improve robustness to distribution shifts, by combining expert
knowledge about the ``high-level" structure of the data with standard
classifiers. Specifically, we introduce two-stage classifiers called
\textit{memory classifiers}. First, these identify prototypical data points --
\textit{memories} -- to cluster the training data. This step is based on
features designed with expert guidance; for instance, for image data they can
be extracted using digital image processing algorithms. Then, within each
cluster, we learn local classifiers based on finer discriminating features, via
standard models like deep neural networks. We establish generalization bounds
for memory classifiers. We illustrate in experiments that they can improve
generalization and robustness to distribution shifts on image datasets. We show
improvements which push beyond standard data augmentation techniques. |
2,179 | Intrinsic dimensionality and generalization properties of the $\mathcal{R}$-norm inductive bias | We study the structural and statistical properties of $\mathcal{R}$-norm
minimizing interpolants of datasets labeled by specific target functions. The
$\mathcal{R}$-norm is the basis of an inductive bias for two-layer neural
networks, recently introduced to capture the functional effect of controlling
the size of network weights, independently of the network width. We find that
these interpolants are intrinsically multivariate functions, even when there
are ridge functions that fit the data, and also that the $\mathcal{R}$-norm
inductive bias is not sufficient for achieving statistically optimal
generalization for certain learning problems. Altogether, these results shed
new light on an inductive bias that is connected to practical neural network
training. |
2,180 | Large-Scale Retrieval for Reinforcement Learning | Effective decision making involves flexibly relating past experiences and
relevant contextual information to a novel situation. In deep reinforcement
learning, the dominant paradigm is for an agent to amortise information that
helps decision-making into its network weights via gradient descent on training
losses. Here, we pursue an alternative approach in which agents can utilise
large-scale context-sensitive database lookups to support their parametric
computations. This allows agents to directly learn in an end-to-end manner to
utilise relevant information to inform their outputs. In addition, new
information can be attended to by the agent, without retraining, by simply
augmenting the retrieval dataset. We study this approach in Go, a challenging
game for which the vast combinatorial state space privileges generalisation
over direct matching to past experiences. We leverage fast, approximate nearest
neighbor techniques in order to retrieve relevant data from a set of tens of
millions of expert demonstration states. Attending to this information provides
a significant boost to prediction accuracy and game-play performance over
simply using these demonstrations as training trajectories, providing a
compelling demonstration of the value of large-scale retrieval in reinforcement
learning agents. |
2,181 | Does Self-supervised Learning Really Improve Reinforcement Learning from Pixels? | We investigate whether self-supervised learning (SSL) can improve online
reinforcement learning (RL) from pixels. We extend the contrastive
reinforcement learning framework (e.g., CURL) that jointly optimizes SSL and RL
losses and conduct an extensive amount of experiments with various
self-supervised losses. Our observations suggest that the existing SSL
framework for RL fails to bring meaningful improvement over the baselines only
taking advantage of image augmentation when the same amount of data and
augmentation is used. We further perform an evolutionary search to find the
optimal combination of multiple self-supervised losses for RL, but find that
even such a loss combination fails to meaningfully outperform the methods that
only utilize carefully designed image augmentations. Often, the use of
self-supervised losses under the existing framework lowered RL performances. We
evaluate the approach in multiple different environments including a real-world
robot environment and confirm that no single self-supervised loss or image
augmentation method can dominate all environments and that the current
framework for joint optimization of SSL and RL is limited. Finally, we
empirically investigate the pretraining framework for SSL + RL and the
properties of representations learned with different approaches. |
2,182 | Tight Bounds for State Tomography with Incoherent Measurements | We consider the classic question of state tomography: given copies of an
unknown quantum state $\rho\in\mathbb{C}^{d\times d}$, output $\widehat{\rho}$
for which $\|\rho - \widehat{\rho}\|_{\mathsf{tr}} \le \varepsilon$. When one
is allowed to make coherent measurements entangled across all copies,
$\Theta(d^2/\varepsilon^2)$ copies are necessary and sufficient [Haah et al.
'17, O'Donnell-Wright '16]. Unfortunately, the protocols achieving this rate
incur large quantum memory overheads that preclude implementation on current or
near-term devices. On the other hand, the best known protocol using incoherent
(single-copy) measurements uses $O(d^3/\varepsilon^2)$ copies
[Kueng-Rauhut-Terstiege '17], and multiple papers have posed it as an open
question to understand whether or not this rate is tight. In this work, we
fully resolve this question, by showing that any protocol using incoherent
measurements, even if they are chosen adaptively, requires
$\Omega(d^3/\varepsilon^2)$ copies, matching the upper bound of
[Kueng-Rauhut-Terstiege '17].
We do so by a new proof technique which directly bounds the "tilt" of the
posterior distribution after measurements, which yields a surprisingly short
proof of our lower bound, and which we believe may be of independent interest. |
2,183 | Causal Balancing for Domain Generalization | While machine learning models rapidly advance the state-of-the-art on various
real-world tasks, out-of-domain (OOD) generalization remains a challenging
problem given the vulnerability of these models to spurious correlations. While
current domain generalization methods usually focus on enforcing certain
invariance properties across different domains by new loss function designs, we
propose a balanced mini-batch sampling strategy to reduce the domain-specific
spurious correlations in the observed training distributions. More
specifically, we propose a two-phased method that 1) identifies the source of
spurious correlations, and 2) builds balanced mini-batches free from spurious
correlations by matching on the identified source. We provide an
identifiability guarantee of the source of spuriousness and show that our
proposed approach provably samples from a balanced, spurious-free distribution
over all training environments. Experiments are conducted on three computer
vision datasets with documented spurious correlations, demonstrating
empirically that our balanced mini-batch sampling strategy improves the
performance of four different established domain generalization model baselines
compared to the random mini-batch sampling strategy. |
2,184 | Meta Optimal Transport | We study the use of amortized optimization to predict optimal transport (OT)
maps from the input measures, which we call Meta OT. This helps repeatedly
solve similar OT problems between different measures by leveraging the
knowledge and information present from past problems to rapidly predict and
solve new problems. Otherwise, standard methods ignore the knowledge of the
past solutions and suboptimally re-solve each problem from scratch. Meta OT
models surpass the standard convergence rates of log-Sinkhorn solvers in the
discrete setting and convex potentials in the continuous setting. We improve
the computational time of standard OT solvers by multiple orders of magnitude
in discrete and continuous transport settings between images, spherical data,
and color palettes. Our source code is available at
http://github.com/facebookresearch/meta-ot. |
2,185 | Balanced Product of Experts for Long-Tailed Recognition | Many real-world recognition problems suffer from an imbalanced or long-tailed
label distribution. Those distributions make representation learning more
challenging due to limited generalization over the tail classes. If the test
distribution differs from the training distribution, e.g. uniform versus
long-tailed, the problem of the distribution shift needs to be addressed. To
this aim, recent works have extended softmax cross-entropy using margin
modifications, inspired by Bayes' theorem. In this paper, we generalize several
approaches with a Balanced Product of Experts (BalPoE), which combines a family
of models with different test-time target distributions to tackle the imbalance
in the data. The proposed experts are trained in a single stage, either jointly
or independently, and fused seamlessly into a BalPoE. We show that BalPoE is
Fisher consistent for minimizing the balanced error and perform extensive
experiments to validate the effectiveness of our approach. Finally, we
investigate the effect of Mixup in this setting, discovering that
regularization is a key ingredient for learning calibrated experts. Our
experiments show that a regularized BalPoE can perform remarkably well in test
accuracy and calibration metrics, leading to state-of-the-art results on
CIFAR-100-LT, ImageNet-LT, and iNaturalist-2018 datasets. The code will be made
publicly available upon paper acceptance. |
2,186 | Is Self-Supervised Learning More Robust Than Supervised Learning? | Self-supervised contrastive learning is a powerful tool to learn visual
representation without labels. Prior work has primarily focused on evaluating
the recognition accuracy of various pre-training algorithms, but has overlooked
other behavioral aspects. In addition to accuracy, distributional robustness
plays a critical role in the reliability of machine learning models. We design
and conduct a series of robustness tests to quantify the behavioral differences
between contrastive learning and supervised learning to downstream or
pre-training data distribution changes. These tests leverage data corruptions
at multiple levels, ranging from pixel-level gamma distortion to patch-level
shuffling and to dataset-level distribution shift. Our tests unveil intriguing
robustness behaviors of contrastive and supervised learning. On the one hand,
under downstream corruptions, we generally observe that contrastive learning is
surprisingly more robust than supervised learning. On the other hand, under
pre-training corruptions, we find contrastive learning vulnerable to patch
shuffling and pixel intensity change, yet less sensitive to dataset-level
distribution change. We attempt to explain these results through the role of
data augmentation and feature space properties. Our insight has implications in
improving the downstream robustness of supervised learning. |
2,187 | Interactively Learning Preference Constraints in Linear Bandits | We study sequential decision-making with known rewards and unknown
constraints, motivated by situations where the constraints represent
expensive-to-evaluate human preferences, such as safe and comfortable driving
behavior. We formalize the challenge of interactively learning about these
constraints as a novel linear bandit problem which we call constrained linear
best-arm identification. To solve this problem, we propose the Adaptive
Constraint Learning (ACOL) algorithm. We provide an instance-dependent lower
bound for constrained linear best-arm identification and show that ACOL's
sample complexity matches the lower bound in the worst-case. In the average
case, ACOL's sample complexity bound is still significantly tighter than bounds
of simpler approaches. In synthetic experiments, ACOL performs on par with an
oracle solution and outperforms a range of baselines. As an application, we
consider learning constraints to represent human preferences in a driving
simulation. ACOL is significantly more sample efficient than alternatives for
this application. Further, we find that learning preferences as constraints is
more robust to changes in the driving scenario than encoding the preferences
directly in the reward function. |
2,188 | Rethinking Spatial Invariance of Convolutional Networks for Object Counting | Previous work generally believes that improving the spatial invariance of
convolutional networks is the key to object counting. However, after verifying
several mainstream counting networks, we surprisingly found too strict
pixel-level spatial invariance would cause overfit noise in the density map
generation. In this paper, we try to use locally connected Gaussian kernels to
replace the original convolution filter to estimate the spatial position in the
density map. The purpose of this is to allow the feature extraction process to
potentially stimulate the density map generation process to overcome the
annotation noise. Inspired by previous work, we propose a low-rank
approximation accompanied with translation invariance to favorably implement
the approximation of massive Gaussian convolution. Our work points a new
direction for follow-up research, which should investigate how to properly
relax the overly strict pixel-level spatial invariance for object counting. We
evaluate our methods on 4 mainstream object counting networks (i.e., MCNN,
CSRNet, SANet, and ResNet-50). Extensive experiments were conducted on 7
popular benchmarks for 3 applications (i.e., crowd, vehicle, and plant
counting). Experimental results show that our methods significantly outperform
other state-of-the-art methods and achieve promising learning of the spatial
position of objects. |
2,189 | Accelerated Algorithms for Monotone Inclusions and Constrained Nonconvex-Nonconcave Min-Max Optimization | We study monotone inclusions and monotone variational inequalities, as well
as their generalizations to non-monotone settings. We first show that the Extra
Anchored Gradient (EAG) algorithm, originally proposed by Yoon and Ryu [2021]
for unconstrained convex-concave min-max optimization, can be applied to solve
the more general problem of Lipschitz monotone inclusion. More specifically, we
prove that the EAG solves Lipschitz monotone inclusion problems with an
\emph{accelerated convergence rate} of $O(\frac{1}{T})$, which is \emph{optimal
among all first-order methods} [Diakonikolas, 2020, Yoon and Ryu, 2021]. Our
second result is a new algorithm, called Extra Anchored Gradient Plus (EAG+),
which not only achieves the accelerated $O(\frac{1}{T})$ convergence rate for
all monotone inclusion problems, but also exhibits the same accelerated rate
for a family of general (non-monotone) inclusion problems that concern negative
comonotone operators. As a special case of our second result, EAG+ enjoys the
$O(\frac{1}{T})$ convergence rate for solving a non-trivial class of
nonconvex-nonconcave min-max optimization problems. Our analyses are based on
simple potential function arguments, which might be useful for analysing other
accelerated algorithms. |
2,190 | List-Decodable Sparse Mean Estimation via Difference-of-Pairs Filtering | We study the problem of list-decodable sparse mean estimation. Specifically,
for a parameter $\alpha \in (0, 1/2)$, we are given $m$ points in
$\mathbb{R}^n$, $\lfloor \alpha m \rfloor$ of which are i.i.d. samples from a
distribution $D$ with unknown $k$-sparse mean $\mu$. No assumptions are made on
the remaining points, which form the majority of the dataset. The goal is to
return a small list of candidates containing a vector $\widehat \mu$ such that
$\| \widehat \mu - \mu \|_2$ is small. Prior work had studied the problem of
list-decodable mean estimation in the dense setting. In this work, we develop a
novel, conceptually simpler technique for list-decodable mean estimation. As
the main application of our approach, we provide the first sample and
computationally efficient algorithm for list-decodable sparse mean estimation.
In particular, for distributions with ``certifiably bounded'' $t$-th moments in
$k$-sparse directions and sufficiently light tails, our algorithm achieves
error of $(1/\alpha)^{O(1/t)}$ with sample complexity $m =
(k\log(n))^{O(t)}/\alpha$ and running time $\mathrm{poly}(mn^t)$. For the
special case of Gaussian inliers, our algorithm achieves the optimal error
guarantee of $\Theta (\sqrt{\log(1/\alpha)})$ with quasi-polynomial sample and
computational complexity. We complement our upper bounds with nearly-matching
statistical query and low-degree polynomial testing lower bounds. |
2,191 | ROI-Constrained Bidding via Curriculum-Guided Bayesian Reinforcement Learning | Real-Time Bidding (RTB) is an important mechanism in modern online
advertising systems. Advertisers employ bidding strategies in RTB to optimize
their advertising effects subject to various financial requirements, especially
the return-on-investment (ROI) constraint. ROIs change non-monotonically during
the sequential bidding process, and often induce a see-saw effect between
constraint satisfaction and objective optimization. While some existing
approaches show promising results in static or mildly changing ad markets, they
fail to generalize to highly dynamic ad markets with ROI constraints, due to
their inability to adaptively balance constraints and objectives amidst
non-stationarity and partial observability. In this work, we specialize in
ROI-Constrained Bidding in non-stationary markets. Based on a Partially
Observable Constrained Markov Decision Process, our method exploits an
indicator-augmented reward function free of extra trade-off parameters and
develops a Curriculum-Guided Bayesian Reinforcement Learning (CBRL) framework
to adaptively control the constraint-objective trade-off in non-stationary ad
markets. Extensive experiments on a large-scale industrial dataset with two
problem settings reveal that CBRL generalizes well in both in-distribution and
out-of-distribution data regimes, and enjoys superior learning efficiency and
stability. |
2,192 | StructCoder: Structure-Aware Transformer for Code Generation | There has been a recent surge of interest in automating software engineering
tasks using deep learning. This work addresses the problem of code generation
where the goal is to generate target code given source code in a different
language or a natural language description. Most of the state-of-the-art deep
learning models for code generation use training strategies that are primarily
designed for natural language. However, understanding and generating code
requires a more rigorous comprehension of the code syntax and semantics. With
this motivation, we develop an encoder-decoder Transformer model where both the
encoder and decoder are trained to recognize the syntax and data flow in the
source and target codes, respectively. We not only make the encoder
structure-aware by leveraging the source code's syntax tree and data flow
graph, but we also ensure that our decoder preserves the syntax and data flow
of the target code by introducing two auxiliary tasks: AST (Abstract Syntax
Tree) paths prediction and data flow prediction. To the best of our knowledge,
this is the first work to introduce a structure-aware Transformer decoder to
enhance the quality of generated code by modeling target syntax and data flow.
The proposed StructCoder model achieves state-of-the-art performance on code
translation and text-to-code generation tasks in the CodeXGLUE benchmark. |
2,193 | Measuring the Carbon Intensity of AI in Cloud Instances | By providing unprecedented access to computational resources, cloud computing
has enabled rapid growth in technologies such as machine learning, the
computational demands of which incur a high energy cost and a commensurate
carbon footprint. As a result, recent scholarship has called for better
estimates of the greenhouse gas impact of AI: data scientists today do not have
easy or reliable access to measurements of this information, precluding
development of actionable tactics. Cloud providers presenting information about
software carbon intensity to users is a fundamental stepping stone towards
minimizing emissions. In this paper, we provide a framework for measuring
software carbon intensity, and propose to measure operational carbon emissions
by using location-based and time-specific marginal emissions data per energy
unit. We provide measurements of operational software carbon intensity for a
set of modern models for natural language processing and computer vision, and a
wide range of model sizes, including pretraining of a 6.1 billion parameter
language model. We then evaluate a suite of approaches for reducing emissions
on the Microsoft Azure cloud compute platform: using cloud instances in
different geographic regions, using cloud instances at different times of day,
and dynamically pausing cloud instances when the marginal carbon intensity is
above a certain threshold. We confirm previous results that the geographic
region of the data center plays a significant role in the carbon intensity for
a given cloud instance, and find that choosing an appropriate region can have
the largest operational emissions reduction impact. We also show that the time
of day has notable impact on operational software carbon intensity. Finally, we
conclude with recommendations for how machine learning practitioners can use
software carbon intensity information to reduce environmental impact. |
2,194 | ProActive: Self-Attentive Temporal Point Process Flows for Activity Sequences | Any human activity can be represented as a temporal sequence of actions
performed to achieve a certain goal. Unlike machine-made time series, these
action sequences are highly disparate as the time taken to finish a similar
action might vary between different persons. Therefore, understanding the
dynamics of these sequences is essential for many downstream tasks such as
activity length prediction, goal prediction, etc. Existing neural approaches
that model an activity sequence are either limited to visual data or are task
specific, i.e., limited to next action or goal prediction. In this paper, we
present ProActive, a neural marked temporal point process (MTPP) framework for
modeling the continuous-time distribution of actions in an activity sequence
while simultaneously addressing three high-impact problems -- next action
prediction, sequence-goal prediction, and end-to-end sequence generation.
Specifically, we utilize a self-attention module with temporal normalizing
flows to model the influence and the inter-arrival times between actions in a
sequence. Moreover, for time-sensitive prediction, we perform an early
detection of sequence goal via a constrained margin-based optimization
procedure. This in-turn allows ProActive to predict the sequence goal using a
limited number of actions. Extensive experiments on sequences derived from
three activity recognition datasets show the significant accuracy boost of
ProActive over the state-of-the-art in terms of action and goal prediction, and
the first-ever application of end-to-end action sequence generation. |
2,195 | A new distance measurement and its application in K-Means Algorithm | K-Means clustering algorithm is one of the most commonly used clustering
algorithms because of its simplicity and efficiency. K-Means clustering
algorithm based on Euclidean distance only pays attention to the linear
distance between samples, but ignores the overall distribution structure of the
dataset (i.e. the fluid structure of dataset). Since it is difficult to
describe the internal structure of two data points by Euclidean distance in
high-dimensional data space, we propose a new distance measurement, namely,
view-distance, and apply it to the K-Means algorithm. On the classical manifold
learning datasets, S-curve and Swiss roll datasets, not only this new distance
can cluster the data according to the structure of the data itself, but also
the boundaries between categories are neat dividing lines. Moreover, we also
tested the classification accuracy and clustering effect of the K-Means
algorithm based on view-distance on some real-world datasets. The experimental
results show that, on most datasets, the K-Means algorithm based on
view-distance has a certain degree of improvement in classification accuracy
and clustering effect. |
2,196 | Extending Process Discovery with Model Complexity Optimization and Cyclic States Identification: Application to Healthcare Processes | Within Process mining, discovery techniques had made it possible to construct
business process models automatically from event logs. However, results often
do not achieve the balance between model complexity and its fitting accuracy,
so there is a need for manual model adjusting. The paper presents an approach
to process mining providing semi-automatic support to model optimization based
on the combined assessment of the model complexity and fitness. To balance
between the two ingredients, a model simplification approach is proposed, which
essentially abstracts the raw model at the desired granularity. Additionally,
we introduce a concept of meta-states, a cycle collapsing in the model, which
can potentially simplify the model and interpret it. We aim to demonstrate the
capabilities of the technological solution using three datasets from different
applications in the healthcare domain. They are remote monitoring process for
patients with arterial hypertension and workflows of healthcare workers during
the COVID-19 pandemic. A case study also investigates the use of various
complexity measures and different ways of solution application providing
insights on better practices in improving interpretability and
complexity/fitness balance in process models. |
2,197 | Hierarchical Federated Learning with Privacy | Federated learning (FL), where data remains at the federated clients, and
where only gradient updates are shared with a central aggregator, was assumed
to be private. Recent work demonstrates that adversaries with gradient-level
access can mount successful inference and reconstruction attacks. In such
settings, differentially private (DP) learning is known to provide resilience.
However, approaches used in the status quo (\ie central and local DP) introduce
disparate utility vs. privacy trade-offs. In this work, we take the first step
towards mitigating such trade-offs through {\em hierarchical FL (HFL)}. We
demonstrate that by the introduction of a new intermediary level where
calibrated DP noise can be added, better privacy vs. utility trade-offs can be
obtained; we term this {\em hierarchical DP (HDP)}. Our experiments with 3
different datasets (commonly used as benchmarks for FL) suggest that HDP
produces models as accurate as those obtained using central DP, where noise is
added at a central aggregator. Such an approach also provides comparable
benefit against inference adversaries as in the local DP case, where noise is
added at the federated clients. |
2,198 | Dynamic mean field programming | A dynamic mean field theory is developed for model based Bayesian
reinforcement learning in the large state space limit. In an analogy with the
statistical physics of disordered systems, the transition probabilities are
interpreted as couplings, and value functions as deterministic spins, and thus
the sampled transition probabilities are considered to be quenched random
variables. The results reveal that, under standard assumptions, the posterior
over Q-values is asymptotically independent and Gaussian across state-action
pairs, for infinite horizon problems. The finite horizon case exhibits the same
behaviour for all state-actions pairs at each time but has an additional
correlation across time, for each state-action pair. The results also hold for
policy evaluation. The Gaussian statistics can be computed from a set of
coupled mean field equations derived from the Bellman equation, which we call
dynamic mean field programming (DMFP). For Q-value iteration, approximate
equations are obtained by appealing to extreme value theory, and closed form
expressions are found in the independent and identically distributed case. The
Lyapunov stability of these closed form equations is studied. |
2,199 | Bayesian Estimation of Differential Privacy | Algorithms such as Differentially Private SGD enable training machine
learning models with formal privacy guarantees. However, there is a discrepancy
between the protection that such algorithms guarantee in theory and the
protection they afford in practice. An emerging strand of work empirically
estimates the protection afforded by differentially private training as a
confidence interval for the privacy budget $\varepsilon$ spent on training a
model. Existing approaches derive confidence intervals for $\varepsilon$ from
confidence intervals for the false positive and false negative rates of
membership inference attacks. Unfortunately, obtaining narrow high-confidence
intervals for $\epsilon$ using this method requires an impractically large
sample size and training as many models as samples. We propose a novel Bayesian
method that greatly reduces sample size, and adapt and validate a heuristic to
draw more than one sample per trained model. Our Bayesian method exploits the
hypothesis testing interpretation of differential privacy to obtain a posterior
for $\varepsilon$ (not just a confidence interval) from the joint posterior of
the false positive and false negative rates of membership inference attacks.
For the same sample size and confidence, we derive confidence intervals for
$\varepsilon$ around 40% narrower than prior work. The heuristic, which we
adapt from label-only DP, can be used to further reduce the number of trained
models needed to get enough samples by up to 2 orders of magnitude. |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.