arxiv_id large_stringlengths 9 16 | title large_stringlengths 0 387 | authors large listlengths 0 3.3k | submission_date large_stringlengths 10 117 | comments large_stringlengths 0 925 | primary_subject large_stringclasses 148 values | subjects large_stringlengths 13 440 | doi large_stringlengths 26 88 | abstract large_stringlengths 3 6.55k | file_path large_stringlengths 23 46 | filepath large_stringlengths 55 68 | subject large_stringclasses 148 values | year large_stringdate 1989-01-01 00:00:00 2025-01-01 00:00:00 | month large_stringclasses 12 values | publication_date date32 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2402.00623 | Bayesian Causal Inference with Gaussian Process Networks | [
"Enrico Giudice",
"Jack Kuipers",
"Giusi Moffa"
] | 1 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.00623 | Causal discovery and inference from observational data is an essential problem in statistics posing both modeling and computational challenges. These are typically addressed by imposing strict assumptions on the joint distribution such as linearity. We consider the problem of the Bayesian estimation of the effects of hypothetical interventions in the Gaussian Process Network (GPN) model, a flexible causal framework which allows describing the causal relationships nonparametrically. We detail how to perform causal inference on GPNs by simulating the effect of an intervention across the whole network and propagating the effect of the intervention on downstream variables. We further derive a simpler computational approximation by estimating the intervention distribution as a function of local variables only, modeling the conditional distributions via additive Gaussian processes. We extend both frameworks beyond the case of a known causal graph, incorporating uncertainty about the causal structure via Markov chain Monte Carlo methods. Simulation studies show that our approach is able to identify the effects of hypothetical interventions with non-Gaussian, non-linear observational data and accurately reflect the posterior uncertainty of the causal estimates. Finally we compare the results of our GPN-based causal inference approach to existing methods on a dataset of $A.~thaliana$ gene expressions. | stat-04.zip/2402.00623.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.00501 | Equivalence of the Empirical Risk Minimization to Regularization on the Family of f-Divergences | [
"Francisco Daunas",
"Iñaki Esnaola",
"Samir M. Perlaza",
"H. Vincent Poor"
] | 1 Feb 2024 (<a href="https://arxiv.org/abs/2402.00501v1">v1</a>), last revised 23 Oct 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Information Theory (cs.IT); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.00501 | The solution to empirical risk minimization with $f$-divergence regularization (ERM-$f$DR) is presented under mild conditions on $f$. Under such conditions, the optimal measure is shown to be unique. Examples of the solution for particular choices of the function $f$ are presented. Previously known solutions to common regularization choices are obtained by leveraging the flexibility of the family of $f$-divergences. These include the unique solutions to empirical risk minimization with relative entropy regularization (Type-I and Type-II). The analysis of the solution unveils the following properties of $f$-divergences when used in the ERM-$f$DR problem: $i\bigl)$ $f$-divergence regularization forces the support of the solution to coincide with the support of the reference measure, which introduces a strong inductive bias that dominates the evidence provided by the training data; and $ii\bigl)$ any $f$-divergence regularization is equivalent to a different $f$-divergence regularization with an appropriate transformation of the empirical risk function. | stat-04.zip/2402.00501.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.07341 | Noise-Adaptive Confidence Sets for Linear Bandits and Application to Bayesian Optimization | [
"Kwang-Sung Jun",
"Jungtaek Kim"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.07341v1">v1</a>), last revised 8 Jun 2024 (this version, v2) | accepted to ICML'24; fixed typos | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.07341 | Adapting to a priori unknown noise level is a very important but challenging problem in sequential decision-making as efficient exploration typically requires knowledge of the noise level, which is often loosely specified. We report significant progress in addressing this issue for linear bandits in two respects. First, we propose a novel confidence set that is `semi-adaptive' to the unknown sub-Gaussian parameter $\sigma_*^2$ in the sense that the (normalized) confidence width scales with $\sqrt{d\sigma_*^2 + \sigma_0^2}$ where $d$ is the dimension and $\sigma_0^2$ is the specified sub-Gaussian parameter (known) that can be much larger than $\sigma_*^2$. This is a significant improvement over $\sqrt{d\sigma_0^2}$ of the standard confidence set of Abbasi-Yadkori et al. (2011), especially when $d$ is large or $\sigma_*^2=0$. We show that this leads to an improved regret bound in linear bandits. Second, for bounded rewards, we propose a novel variance-adaptive confidence set that has much improved numerical performance upon prior art. We then apply this confidence set to develop, as we claim, the first practical variance-adaptive linear bandit algorithm via an optimistic approach, which is enabled by our novel regret analysis technique. Both of our confidence sets rely critically on `regret equality' from online learning. Our empirical evaluation in diverse Bayesian optimization tasks shows that our proposed algorithms demonstrate better or comparable performance compared to existing methods. | stat-04.zip/2402.07341.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.06151 | POTEC: Off-Policy Learning for Large Action Spaces via Two-Stage Policy Decomposition | [
"Yuta Saito",
"Jihan Yao",
"Thorsten Joachims"
] | 9 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.06151 | We study off-policy learning (OPL) of contextual bandit policies in large discrete action spaces where existing methods -- most of which rely crucially on reward-regression models or importance-weighted policy gradients -- fail due to excessive bias or variance. To overcome these issues in OPL, we propose a novel two-stage algorithm, called Policy Optimization via Two-Stage Policy Decomposition (POTEC). It leverages clustering in the action space and learns two different policies via policy- and regression-based approaches, respectively. In particular, we derive a novel low-variance gradient estimator that enables to learn a first-stage policy for cluster selection efficiently via a policy-based approach. To select a specific action within the cluster sampled by the first-stage policy, POTEC uses a second-stage policy derived from a regression-based approach within each cluster. We show that a local correctness condition, which only requires that the regression model preserves the relative expected reward differences of the actions within each cluster, ensures that our policy-gradient estimator is unbiased and the second-stage policy is optimal. We also show that POTEC provides a strict generalization of policy- and regression-based approaches and their associated assumptions. Comprehensive experiments demonstrate that POTEC provides substantial improvements in OPL effectiveness particularly in large and structured action spaces. | stat-04.zip/2402.06151.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.01000 | Multivariate Probabilistic Time Series Forecasting with Correlated Errors | [
"Vincent Zhihao Zheng",
"Lijun Sun"
] | 1 Feb 2024 (<a href="https://arxiv.org/abs/2402.01000v1">v1</a>), last revised 7 Nov 2024 (this version, v4) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01000 | Accurately modeling the correlation structure of errors is critical for reliable uncertainty quantification in probabilistic time series forecasting. While recent deep learning models for multivariate time series have developed efficient parameterizations for time-varying contemporaneous covariance, but they often assume temporal independence of errors for simplicity. However, real-world data often exhibit significant error autocorrelation and cross-lag correlation due to factors such as missing covariates. In this paper, we introduce a plug-and-play method that learns the covariance structure of errors over multiple steps for autoregressive models with Gaussian-distributed errors. To ensure scalable inference and computational efficiency, we model the contemporaneous covariance using a low-rank-plus-diagonal parameterization and capture cross-covariance through a group of independent latent temporal processes. The learned covariance matrix is then used to calibrate predictions based on observed residuals. We evaluate our method on probabilistic models built on RNNs and Transformer architectures, and the results confirm the effectiveness of our approach in improving predictive accuracy and uncertainty quantification without significantly increasing the parameter size. | stat-04.zip/2402.01000.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.05719 | Exact capacity of the \emph{wide} hidden layer treelike neural networks with generic activations | [
"Mihailo Stojnic"
] | 8 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Disordered Systems and Neural Networks (cond-mat.dis-nn); Information Theory (cs.IT); Machine Learning (cs.LG); Probability (math.PR) | https://doi.org/10.48550/arXiv.2402.05719 | Recent progress in studying \emph{treelike committee machines} (TCM) neural networks (NN) in \cite{Stojnictcmspnncaprdt23,Stojnictcmspnncapliftedrdt23,Stojnictcmspnncapdiffactrdt23} showed that the Random Duality Theory (RDT) and its a \emph{partially lifted}(pl RDT) variant are powerful tools that can be used for very precise networks capacity analysis. Here, we consider \emph{wide} hidden layer networks and uncover that certain aspects of numerical difficulties faced in \cite{Stojnictcmspnncapdiffactrdt23} miraculously disappear. In particular, we employ recently developed \emph{fully lifted} (fl) RDT to characterize the \emph{wide} ($d\rightarrow \infty$) TCM nets capacity. We obtain explicit, closed form, capacity characterizations for a very generic class of the hidden layer activations. While the utilized approach significantly lowers the amount of the needed numerical evaluations, the ultimate fl RDT usefulness and success still require a solid portion of the residual numerical work. To get the concrete capacity values, we take four very famous activations examples: \emph{\textbf{ReLU}}, \textbf{\emph{quadratic}}, \textbf{\emph{erf}}, and \textbf{\emph{tanh}}. After successfully conducting all the residual numerical work for all of them, we uncover that the whole lifting mechanism exhibits a remarkably rapid convergence with the relative improvements no better than $\sim 0.1\%$ happening already on the 3-rd level of lifting. As a convenient bonus, we also uncover that the capacity characterizations obtained on the first and second level of lifting precisely match those obtained through the statistical physics replica theory methods in \cite{ZavPeh21} for the generic and in \cite{BalMalZech19} for the ReLU activations. | stat-04.zip/2402.05719.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.05718 | REMEDI: Corrective Transformations for Improved Neural Entropy Estimation | [
"Viktor Nilsson",
"Anirban Samaddar",
"Sandeep Madireddy",
"Pierre Nyquist"
] | 8 Feb 2024 (<a href="https://arxiv.org/abs/2402.05718v1">v1</a>), last revised 20 May 2024 (this version, v2) | To appear in ICML 2024. 30 pages, 18 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.05718 | Information theoretic quantities play a central role in machine learning. The recent surge in the complexity of data and models has increased the demand for accurate estimation of these quantities. However, as the dimension grows the estimation presents significant challenges, with existing methods struggling already in relatively low dimensions. To address this issue, in this work, we introduce $\texttt{REMEDI}$ for efficient and accurate estimation of differential entropy, a fundamental information theoretic quantity. The approach combines the minimization of the cross-entropy for simple, adaptive base models and the estimation of their deviation, in terms of the relative entropy, from the data density. Our approach demonstrates improvement across a broad spectrum of estimation tasks, encompassing entropy estimation on both synthetic and natural data. Further, we extend important theoretical consistency results to a more generalized setting required by our approach. We illustrate how the framework can be naturally extended to information theoretic supervised learning models, with a specific focus on the Information Bottleneck approach. It is demonstrated that the method delivers better accuracy compared to the existing methods in Information Bottleneck. In addition, we explore a natural connection between $\texttt{REMEDI}$ and generative modeling using rejection sampling and Langevin dynamics. | stat-04.zip/2402.05718.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.00645 | Spectrally Transformed Kernel Regression | [
"Runtian Zhai",
"Rattana Pukdee",
"Roger Jin",
"Maria-Florina Balcan",
"Pradeep Ravikumar"
] | 1 Feb 2024 | ICLR 2024 spotlight. 36 pages | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.00645 | Unlabeled data is a key component of modern machine learning. In general, the role of unlabeled data is to impose a form of smoothness, usually from the similarity information encoded in a base kernel, such as the $\epsilon$-neighbor kernel or the adjacency matrix of a graph. This work revisits the classical idea of spectrally transformed kernel regression (STKR), and provides a new class of general and scalable STKR estimators able to leverage unlabeled data. Intuitively, via spectral transformation, STKR exploits the data distribution for which unlabeled data can provide additional information. First, we show that STKR is a principled and general approach, by characterizing a universal type of "target smoothness", and proving that any sufficiently smooth function can be learned by STKR. Second, we provide scalable STKR implementations for the inductive setting and a general transformation function, while prior work is mostly limited to the transductive setting. Third, we derive statistical guarantees for two scenarios: STKR with a known polynomial transformation, and STKR with kernel PCA when the transformation is unknown. Overall, we believe that this work helps deepen our understanding of how to work with unlabeled data, and its generality makes it easier to inspire new methods. | stat-04.zip/2402.00645.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.05878 | Prior-Dependent Allocations for Bayesian Fixed-Budget Best-Arm Identification in Structured Bandits | [
"Nicolas Nguyen",
"Imad Aouali",
"András György",
"Claire Vernade"
] | 8 Feb 2024 (<a href="https://arxiv.org/abs/2402.05878v1">v1</a>), last revised 24 Apr 2025 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.05878 | We study the problem of Bayesian fixed-budget best-arm identification (BAI) in structured bandits. We propose an algorithm that uses fixed allocations based on the prior information and the structure of the environment. We provide theoretical bounds on its performance across diverse models, including the first prior-dependent upper bounds for linear and hierarchical BAI. Our key contribution is introducing new proof methods that result in tighter bounds for multi-armed BAI compared to existing methods. We extensively compare our approach to other fixed-budget BAI methods, demonstrating its consistent and robust performance in various settings. Our work improves our understanding of Bayesian fixed-budget BAI in structured bandits and highlights the effectiveness of our approach in practical scenarios. | stat-04.zip/2402.05878.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.09018 | Neural Operators Meet Energy-based Theory: Operator Learning for Hamiltonian and Dissipative PDEs | [
"Yusuke Tanaka",
"Takaharu Yaguchi",
"Tomoharu Iwata",
"Naonori Ueda"
] | 14 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.09018 | The operator learning has received significant attention in recent years, with the aim of learning a mapping between function spaces. Prior works have proposed deep neural networks (DNNs) for learning such a mapping, enabling the learning of solution operators of partial differential equations (PDEs). However, these works still struggle to learn dynamics that obeys the laws of physics. This paper proposes Energy-consistent Neural Operators (ENOs), a general framework for learning solution operators of PDEs that follows the energy conservation or dissipation law from observed solution trajectories. We introduce a novel penalty function inspired by the energy-based theory of physics for training, in which the energy functional is modeled by another DNN, allowing one to bias the outputs of the DNN-based solution operators to ensure energetic consistency without explicit PDEs. Experiments on multiple physical systems show that ENO outperforms existing DNN models in predicting solutions from data, especially in super-resolution settings. | stat-04.zip/2402.09018.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.06388 | Convergence of a L2 regularized Policy Gradient Algorithm for the Multi Armed Bandit | [
"Stefana Anita",
"Gabriel Turinici"
] | 9 Feb 2024 (<a href="https://arxiv.org/abs/2402.06388v1">v1</a>), last revised 26 Nov 2024 (this version, v3) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Numerical Analysis (math.NA) | https://doi.org/10.48550/arXiv.2402.06388 | Although Multi Armed Bandit (MAB) on one hand and the policy gradient approach on the other hand are among the most used frameworks of Reinforcement Learning, the theoretical properties of the policy gradient algorithm used for MAB have not been given enough attention. We investigate in this work the convergence of such a procedure for the situation when a $L2$ regularization term is present jointly with the 'softmax' parametrization. We prove convergence under appropriate technical hypotheses and test numerically the procedure including situations beyond the theoretical setting. The tests show that a time dependent regularized procedure can improve over the canonical approach especially when the initial guess is far from the solution. | stat-04.zip/2402.06388.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.01089 | No Free Prune: Information-Theoretic Barriers to Pruning at Initialization | [
"Tanishq Kumar",
"Kevin Luo",
"Mark Sellke"
] | 2 Feb 2024 (<a href="https://arxiv.org/abs/2402.01089v1">v1</a>), last revised 24 Jul 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01089 | The existence of "lottery tickets" <a href="https://arxiv.org/abs/1803.03635" data-arxiv-id="1803.03635" class="link-https">arXiv:1803.03635</a> at or near initialization raises the tantalizing question of whether large models are necessary in deep learning, or whether sparse networks can be quickly identified and trained without ever training the dense models that contain them. However, efforts to find these sparse subnetworks without training the dense model ("pruning at initialization") have been broadly unsuccessful <a href="https://arxiv.org/abs/2009.08576" data-arxiv-id="2009.08576" class="link-https">arXiv:2009.08576</a>. We put forward a theoretical explanation for this, based on the model's effective parameter count, $p_\text{eff}$, given by the sum of the number of non-zero weights in the final network and the mutual information between the sparsity mask and the data. We show the Law of Robustness of <a href="https://arxiv.org/abs/2105.12806" data-arxiv-id="2105.12806" class="link-https">arXiv:2105.12806</a> extends to sparse networks with the usual parameter count replaced by $p_\text{eff}$, meaning a sparse neural network which robustly interpolates noisy data requires a heavily data-dependent mask. We posit that pruning during and after training outputs masks with higher mutual information than those produced by pruning at initialization. Thus two networks may have the same sparsities, but differ in effective parameter count based on how they were trained. This suggests that pruning near initialization may be infeasible and explains why lottery tickets exist, but cannot be found fast (i.e. without training the full network). Experiments on neural networks confirm that information gained during training may indeed affect model capacity. | stat-04.zip/2402.01089.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.08616 | Adjustment Identification Distance: A gadjid for Causal Structure Learning | [
"Leonard Henckel",
"Theo Würtzen",
"Sebastian Weichwald"
] | 13 Feb 2024 (<a href="https://arxiv.org/abs/2402.08616v1">v1</a>), last revised 11 Jul 2024 (this version, v2) | accepted at UAI 2024 (Conference on Uncertainty in Artificial Intelligence) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.08616 | Evaluating graphs learned by causal discovery algorithms is difficult: The number of edges that differ between two graphs does not reflect how the graphs differ with respect to the identifying formulas they suggest for causal effects. We introduce a framework for developing causal distances between graphs which includes the structural intervention distance for directed acyclic graphs as a special case. We use this framework to develop improved adjustment-based distances as well as extensions to completed partially directed acyclic graphs and causal orders. We develop new reachability algorithms to compute the distances efficiently and to prove their low polynomial time complexity. In our package gadjid (open source at <a href="https://github.com/CausalDisco/gadjid" rel="external noopener nofollow" class="link-external link-https">this https URL</a>), we provide implementations of our distances; they are orders of magnitude faster with proven lower time complexity than the structural intervention distance and thereby provide a success metric for causal discovery that scales to graph sizes that were previously prohibitive. | stat-04.zip/2402.08616.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.02162 | A Bayesian cluster validity index | [
"Nathakhun Wiroonsri",
"Onthada Preedasawakul"
] | 3 Feb 2024 (<a href="https://arxiv.org/abs/2402.02162v1">v1</a>), last revised 14 Feb 2024 (this version, v2) | 23 pages | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST) | https://doi.org/10.48550/arXiv.2402.02162 | Selecting the appropriate number of clusters is a critical step in applying clustering algorithms. To assist in this process, various cluster validity indices (CVIs) have been developed. These indices are designed to identify the optimal number of clusters within a dataset. However, users may not always seek the absolute optimal number of clusters but rather a secondary option that better aligns with their specific applications. This realization has led us to introduce a Bayesian cluster validity index (BCVI), which builds upon existing indices. The BCVI utilizes either Dirichlet or generalized Dirichlet priors, resulting in the same posterior distribution. We evaluate our BCVI using the Wiroonsri index for hard clustering and the Wiroonsri-Preedasawakul index for soft clustering as underlying indices. We compare the performance of our proposed BCVI with that of the original underlying indices and several other existing CVIs, including Davies-Bouldin, Starczewski, Xie-Beni, and KWON2 indices. Our BCVI offers clear advantages in situations where user expertise is valuable, allowing users to specify their desired range for the final number of clusters. To illustrate this, we conduct experiments classified into three different scenarios. Additionally, we showcase the practical applicability of our approach through real-world datasets, such as MRI brain tumor images. These tools will be published as a new R package 'BayesCVI'. | stat-04.zip/2402.02162.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.07131 | Resampling methods for private statistical inference | [
"Karan Chadha",
"John Duchi",
"Rohith Kuditipudi"
] | 11 Feb 2024 (<a href="https://arxiv.org/abs/2402.07131v1">v1</a>), last revised 3 Jun 2024 (this version, v3) | 45 pages | Machine Learning (stat.ML) | Machine Learning (stat.ML); Cryptography and Security (cs.CR); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.07131 | We consider the task of constructing confidence intervals with differential privacy. We propose two private variants of the non-parametric bootstrap, which privately compute the median of the results of multiple "little" bootstraps run on partitions of the data and give asymptotic bounds on the coverage error of the resulting confidence intervals. For a fixed differential privacy parameter $\epsilon$, our methods enjoy the same error rates as that of the non-private bootstrap to within logarithmic factors in the sample size $n$. We empirically validate the performance of our methods for mean estimation, median estimation, and logistic regression with both real and synthetic data. Our methods achieve similar coverage accuracy to existing methods (and non-private baselines) while providing notably shorter ($\gtrsim 10$ times) confidence intervals than previous approaches. | stat-04.zip/2402.07131.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.07723 | Generalization Bounds for Heavy-Tailed SDEs through the Fractional Fokker-Planck Equation | [
"Benjamin Dupuis",
"Umut Şimşekli"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.07723v1">v1</a>), last revised 2 Jun 2025 (this version, v3) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.07723 | Understanding the generalization properties of heavy-tailed stochastic optimization algorithms has attracted increasing attention over the past years. While illuminating interesting aspects of stochastic optimizers by using heavy-tailed stochastic differential equations as proxies, prior works either provided expected generalization bounds, or introduced non-computable information theoretic terms. Addressing these drawbacks, in this work, we prove high-probability generalization bounds for heavy-tailed SDEs which do not contain any nontrivial information theoretic terms. To achieve this goal, we develop new proof techniques based on estimating the entropy flows associated with the so-called fractional Fokker-Planck equation (a partial differential equation that governs the evolution of the distribution of the corresponding heavy-tailed SDE). In addition to obtaining high-probability bounds, we show that our bounds have a better dependence on the dimension of parameters as compared to prior art. Our results further identify a phase transition phenomenon, which suggests that heavy tails can be either beneficial or harmful depending on the problem structure. We support our theory with experiments conducted in a variety of settings. | stat-04.zip/2402.07723.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.07868 | Nesting Particle Filters for Experimental Design in Dynamical Systems | [
"Sahel Iqbal",
"Adrien Corenflos",
"Simo Särkkä",
"Hany Abdulsamad"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.07868v1">v1</a>), last revised 29 May 2024 (this version, v4) | Accepted to ICML 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.07868 | In this paper, we propose a novel approach to Bayesian experimental design for non-exchangeable data that formulates it as risk-sensitive policy optimization. We develop the Inside-Out SMC$^2$ algorithm, a nested sequential Monte Carlo technique to infer optimal designs, and embed it into a particle Markov chain Monte Carlo framework to perform gradient-based policy amortization. Our approach is distinct from other amortized experimental design techniques, as it does not rely on contrastive estimators. Numerical validation on a set of dynamical systems showcases the efficacy of our method in comparison to other state-of-the-art strategies. | stat-04.zip/2402.07868.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.01050 | Distributed MCMC inference for Bayesian Non-Parametric Latent Block Model | [
"Reda Khoufache",
"Anisse Belhadj",
"Hanene Azzag",
"Mustapha Lebbah"
] | 1 Feb 2024 | Accepted to PaKDD 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Computation (stat.CO) | https://doi.org/10.48550/arXiv.2402.01050 | In this paper, we introduce a novel Distributed Markov Chain Monte Carlo (MCMC) inference method for the Bayesian Non-Parametric Latent Block Model (DisNPLBM), employing the Master/Worker architecture. Our non-parametric co-clustering algorithm divides observations and features into partitions using latent multivariate Gaussian block distributions. The workload on rows is evenly distributed among workers, who exclusively communicate with the master and not among themselves. DisNPLBM demonstrates its impact on cluster labeling accuracy and execution times through experimental results. Moreover, we present a real-use case applying our approach to co-cluster gene expression data. The code source is publicly available at <a href="https://github.com/redakhoufache/Distributed-NPLBM" rel="external noopener nofollow" class="link-external link-https">this https URL</a>. | stat-04.zip/2402.01050.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.06940 | Efficient Incremental Belief Updates Using Weighted Virtual Observations | [
"David Tolpin"
] | 10 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.06940 | We present an algorithmic solution to the problem of incremental belief updating in the context of Monte Carlo inference in Bayesian statistical models represented by probabilistic programs. Given a model and a sample-approximated posterior, our solution constructs a set of weighted observations to condition the model such that inference would result in the same posterior. This problem arises e.g. in multi-level modelling, incremental inference, inference in presence of privacy constraints. First, a set of virtual observations is selected, then, observation weights are found through a computationally efficient optimization procedure such that the reconstructed posterior coincides with or closely approximates the original posterior. We implement and apply the solution to a number of didactic examples and case studies, showing efficiency and robustness of our approach. The provided reference implementation is agnostic to the probabilistic programming language or the inference algorithm, and can be applied to most mainstream probabilistic programming environments. | stat-04.zip/2402.06940.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.08095 | Convergence Analysis of Discrete Diffusion Model: Exact Implementation through Uniformization | [
"Hongrui Chen",
"Lexing Ying"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.08095v1">v1</a>), last revised 14 Feb 2024 (this version, v2) | 19 pages | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.08095 | Diffusion models have achieved huge empirical success in data generation tasks. Recently, some efforts have been made to adapt the framework of diffusion models to discrete state space, providing a more natural approach for modeling intrinsically discrete data, such as language and graphs. This is achieved by formulating both the forward noising process and the corresponding reversed process as Continuous Time Markov Chains (CTMCs). In this paper, we investigate the theoretical properties of the discrete diffusion model. Specifically, we introduce an algorithm leveraging the uniformization of continuous Markov chains, implementing transitions on random time points. Under reasonable assumptions on the learning of the discrete score function, we derive Total Variation distance and KL divergence guarantees for sampling from any distribution on a hypercube. Our results align with state-of-the-art achievements for diffusion models in $\mathbb{R}^d$ and further underscore the advantages of discrete diffusion models in comparison to the $\mathbb{R}^d$ setting. | stat-04.zip/2402.08095.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.07626 | Stochastic Gradient Flow Dynamics of Test Risk and its Exact Solution for Weak Features | [
"Rodrigo Veiga",
"Anastasia Remizova",
"Nicolas Macris"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.07626v1">v1</a>), last revised 10 Jun 2024 (this version, v2) | Accepted to ICML 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Disordered Systems and Neural Networks (cond-mat.dis-nn); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.07626 | We investigate the test risk of continuous-time stochastic gradient flow dynamics in learning theory. Using a path integral formulation we provide, in the regime of a small learning rate, a general formula for computing the difference between test risk curves of pure gradient and stochastic gradient flows. We apply the general theory to a simple model of weak features, which displays the double descent phenomenon, and explicitly compute the corrections brought about by the added stochastic term in the dynamics, as a function of time and model parameters. The analytical results are compared to simulations of discrete-time stochastic gradient descent and show good agreement. | stat-04.zip/2402.07626.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.07357 | Regression Trees for Fast and Adaptive Prediction Intervals | [
"Luben M. C. Cabezas",
"Mateus P. Otto",
"Rafael Izbicki",
"Rafael B. Stern"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.07357v1">v1</a>), last revised 13 Feb 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.07357 | Predictive models make mistakes. Hence, there is a need to quantify the uncertainty associated with their predictions. Conformal inference has emerged as a powerful tool to create statistically valid prediction regions around point predictions, but its naive application to regression problems yields non-adaptive regions. New conformal scores, often relying upon quantile regressors or conditional density estimators, aim to address this limitation. Although they are useful for creating prediction bands, these scores are detached from the original goal of quantifying the uncertainty around an arbitrary predictive model. This paper presents a new, model-agnostic family of methods to calibrate prediction intervals for regression problems with local coverage guarantees. Our approach is based on pursuing the coarsest partition of the feature space that approximates conditional coverage. We create this partition by training regression trees and Random Forests on conformity scores. Our proposal is versatile, as it applies to various conformity scores and prediction settings and demonstrates superior scalability and performance compared to established baselines in simulated and real-world datasets. We provide a Python package clover that implements our methods using the standard scikit-learn interface. | stat-04.zip/2402.07357.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.07735 | Graph Structure Inference with BAM: Introducing the Bilinear Attention Mechanism | [
"Philipp Froehlich",
"Heinz Koeppl"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.07735v1">v1</a>), last revised 13 Feb 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.07735 | In statistics and machine learning, detecting dependencies in datasets is a central challenge. We propose a novel neural network model for supervised graph structure learning, i.e., the process of learning a mapping between observational data and their underlying dependence structure. The model is trained with variably shaped and coupled simulated input data and requires only a single forward pass through the trained network for inference. By leveraging structural equation models and employing randomly generated multivariate Chebyshev polynomials for the simulation of training data, our method demonstrates robust generalizability across both linear and various types of non-linear dependencies. We introduce a novel bilinear attention mechanism (BAM) for explicit processing of dependency information, which operates on the level of covariance matrices of transformed data and respects the geometry of the manifold of symmetric positive definite matrices. Empirical evaluation demonstrates the robustness of our method in detecting a wide range of dependencies, excelling in undirected graph estimation and proving competitive in completed partially directed acyclic graph estimation through a novel two-step approach. | stat-04.zip/2402.07735.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.08847 | Space-Time Diffusion Bridge | [
"Hamidreza Behjoo",
"Michael Chertkov"
] | 13 Feb 2024 (<a href="https://arxiv.org/abs/2402.08847v1">v1</a>), last revised 7 Jul 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.08847 | In this study, we introduce a novel method for generating new synthetic samples that are independent and identically distributed (i.i.d.) from high-dimensional real-valued probability distributions, as defined implicitly by a set of Ground Truth (GT) samples. Central to our method is the integration of space-time mixing strategies that extend across temporal and spatial dimensions. Our methodology is underpinned by three interrelated stochastic processes designed to enable optimal transport from an easily tractable initial probability distribution to the target distribution represented by the GT samples: (a) linear processes incorporating space-time mixing that yield Gaussian conditional probability densities, (b) their diffusion bridge analogs that are conditioned to the initial and final state vectors, and (c) nonlinear stochastic processes refined through score-matching techniques. The crux of our training regime involves fine-tuning the nonlinear model, and potentially the linear models -- to align closely with the GT data. We validate the efficacy of our space-time diffusion approach with numerical experiments, laying the groundwork for more extensive future theory and experiments to fully authenticate the method, particularly providing a more efficient (possibly simulation-free) inference. | stat-04.zip/2402.08847.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.08344 | Implicit Bias in Noisy-SGD: With Applications to Differentially Private Training | [
"Tom Sander",
"Maxime Sylvestre",
"Alain Durmus"
] | 13 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.08344 | Training Deep Neural Networks (DNNs) with small batches using Stochastic Gradient Descent (SGD) yields superior test performance compared to larger batches. The specific noise structure inherent to SGD is known to be responsible for this implicit bias. DP-SGD, used to ensure differential privacy (DP) in DNNs' training, adds Gaussian noise to the clipped gradients. Surprisingly, large-batch training still results in a significant decrease in performance, which poses an important challenge because strong DP guarantees necessitate the use of massive batches. We first show that the phenomenon extends to Noisy-SGD (DP-SGD without clipping), suggesting that the stochasticity (and not the clipping) is the cause of this implicit bias, even with additional isotropic Gaussian noise. We theoretically analyse the solutions obtained with continuous versions of Noisy-SGD for the Linear Least Square and Diagonal Linear Network settings, and reveal that the implicit bias is indeed amplified by the additional noise. Thus, the performance issues of large-batch DP-SGD training are rooted in the same underlying principles as SGD, offering hope for potential improvements in large batch training strategies. | stat-04.zip/2402.08344.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.01090 | Scalable Higher-Order Tensor Product Spline Models | [
"David Rügamer"
] | 2 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Computation (stat.CO) | https://doi.org/10.48550/arXiv.2402.01090 | In the current era of vast data and transparent machine learning, it is essential for techniques to operate at a large scale while providing a clear mathematical comprehension of the internal workings of the method. Although there already exist interpretable semi-parametric regression methods for large-scale applications that take into account non-linearity in the data, the complexity of the models is still often limited. One of the main challenges is the absence of interactions in these models, which are left out for the sake of better interpretability but also due to impractical computational costs. To overcome this limitation, we propose a new approach using a factorization method to derive a highly scalable higher-order tensor product spline model. Our method allows for the incorporation of all (higher-order) interactions of non-linear feature effects while having computational costs proportional to a model without interactions. We further develop a meaningful penalization scheme and examine the induced optimization problem. We conclude by evaluating the predictive and estimation performance of our method. | stat-04.zip/2402.01090.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.04114 | SCAFFLSA: Taming Heterogeneity in Federated Linear Stochastic Approximation and TD Learning | [
"Paul Mangold",
"Sergey Samsonov",
"Safwan Labbi",
"Ilya Levin",
"Reda Alami",
"Alexey Naumov",
"Eric Moulines"
] | 6 Feb 2024 (<a href="https://arxiv.org/abs/2402.04114v1">v1</a>), last revised 27 May 2024 (this version, v2) | now with linear speed-up! | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC) | https://doi.org/10.48550/arXiv.2402.04114 | In this paper, we analyze the sample and communication complexity of the federated linear stochastic approximation (FedLSA) algorithm. We explicitly quantify the effects of local training with agent heterogeneity. We show that the communication complexity of FedLSA scales polynomially with the inverse of the desired accuracy $\epsilon$. To overcome this, we propose SCAFFLSA a new variant of FedLSA that uses control variates to correct for client drift, and establish its sample and communication complexities. We show that for statistically heterogeneous agents, its communication complexity scales logarithmically with the desired accuracy, similar to Scaffnew. An important finding is that, compared to the existing results for Scaffnew, the sample complexity scales with the inverse of the number of agents, a property referred to as linear speed-up. Achieving this linear speed-up requires completely new theoretical arguments. We apply the proposed method to federated temporal difference learning with linear function approximation and analyze the corresponding complexity improvements. | stat-04.zip/2402.04114.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.04493 | A Primal-Dual Algorithm for Offline Constrained Reinforcement Learning with Linear MDPs | [
"Kihyuk Hong",
"Ambuj Tewari"
] | 7 Feb 2024 (<a href="https://arxiv.org/abs/2402.04493v1">v1</a>), last revised 2 Jun 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.04493 | We study offline reinforcement learning (RL) with linear MDPs under the infinite-horizon discounted setting which aims to learn a policy that maximizes the expected discounted cumulative reward using a pre-collected dataset. Existing algorithms for this setting either require a uniform data coverage assumptions or are computationally inefficient for finding an $\epsilon$-optimal policy with $O(\epsilon^{-2})$ sample complexity. In this paper, we propose a primal dual algorithm for offline RL with linear MDPs in the infinite-horizon discounted setting. Our algorithm is the first computationally efficient algorithm in this setting that achieves sample complexity of $O(\epsilon^{-2})$ with partial data coverage assumption. Our work is an improvement upon a recent work that requires $O(\epsilon^{-4})$ samples. Moreover, we extend our algorithm to work in the offline constrained RL setting that enforces constraints on additional reward signals. | stat-04.zip/2402.04493.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.07762 | Scalable Structure Learning for Sparse Context-Specific Systems | [
"Felix Leopoldo Rios",
"Alex Markham",
"Liam Solus"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.07762v1">v1</a>), last revised 16 Oct 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Combinatorics (math.CO) | https://doi.org/10.48550/arXiv.2402.07762 | Several approaches to graphically representing context-specific relations among jointly distributed categorical variables have been proposed, along with structure learning algorithms. While existing optimization-based methods have limited scalability due to the large number of context-specific models, the constraint-based methods are more prone to error than even constraint-based directed acyclic graph learning algorithms since more relations must be tested. We present an algorithm for learning context-specific models that scales to hundreds of variables. Scalable learning is achieved through a combination of an order-based Markov chain Monte-Carlo search and a novel, context-specific sparsity assumption that is analogous to those typically invoked for directed acyclic graphical models. Unlike previous Markov chain Monte-Carlo search methods, our Markov chain is guaranteed to have the true posterior of the variable orderings as the stationary distribution. To implement the method, we solve a first case of an open problem recently posed by Alon and Balogh. Future work solving increasingly general instances of this problem would allow our methods to learn increasingly dense models. The method is shown to perform well on synthetic data and real world examples, in terms of both accuracy and scalability. | stat-04.zip/2402.07762.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.13400 | The Dimension of Self-Directed Learning | [
"Pramith Devulapalli",
"Steve Hanneke"
] | 20 Feb 2024 | ALT 2024 Camera ready version | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.13400 | Understanding the self-directed learning complexity has been an important problem that has captured the attention of the online learning theory community since the early 1990s. Within this framework, the learner is allowed to adaptively choose its next data point in making predictions unlike the setting in adversarial online learning. <br>In this paper, we study the self-directed learning complexity in both the binary and multi-class settings, and we develop a dimension, namely $SDdim$, that exactly characterizes the self-directed learning mistake-bound for any concept class. The intuition behind $SDdim$ can be understood as a two-player game called the "labelling game". Armed with this two-player game, we calculate $SDdim$ on a whole host of examples with notable results on axis-aligned rectangles, VC dimension $1$ classes, and linear separators. We demonstrate several learnability gaps with a central focus on self-directed learning and offline sequence learning models that include either the best or worst ordering. Finally, we extend our analysis to the self-directed binary agnostic setting where we derive upper and lower bounds. | stat-04.zip/2402.13400.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.13285 | Leveraging PAC-Bayes Theory and Gibbs Distributions for Generalization Bounds with Complexity Measures | [
"Paul Viallard",
"Rémi Emonet",
"Amaury Habrard",
"Emilie Morvant",
"Valentina Zantedeschi"
] | 19 Feb 2024 | AISTATS 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.13285 | In statistical learning theory, a generalization bound usually involves a complexity measure imposed by the considered theoretical framework. This limits the scope of such bounds, as other forms of capacity measures or regularizations are used in algorithms. In this paper, we leverage the framework of disintegrated PAC-Bayes bounds to derive a general generalization bound instantiable with arbitrary complexity measures. One trick to prove such a result involves considering a commonly used family of distributions: the Gibbs distributions. Our bound stands in probability jointly over the hypothesis and the learning sample, which allows the complexity to be adapted to the generalization gap as it can be customized to fit both the hypothesis class and the task. | stat-04.zip/2402.13285.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.03819 | Do we need rebalancing strategies? A theoretical and empirical study around SMOTE and its variants | [
"Abdoulaye Sakho",
"Emmanuel Malherbe",
"Erwan Scornet"
] | 6 Feb 2024 (<a href="https://arxiv.org/abs/2402.03819v1">v1</a>), last revised 6 Jun 2025 (this version, v5) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.03819 | Synthetic Minority Oversampling Technique (SMOTE) is a common rebalancing strategy for handling imbalanced tabular data sets. However, few works analyze SMOTE theoretically. In this paper, we derive several non-asymptotic upper bound on SMOTE density. From these results, we prove that SMOTE (with default parameter) tends to copy the original minority samples asymptotically. We confirm and illustrate empirically this first theoretical behavior on a real-world <a href="http://data-set.bFurthermore" rel="external noopener nofollow" class="link-external link-http">this http URL</a>, we prove that SMOTE density vanishes near the boundary of the support of the minority class distribution. We then adapt SMOTE based on our theoretical findings to introduce two new variants. These strategies are compared on 13 tabular data sets with 10 state-of-the-art rebalancing procedures, including deep generative and diffusion models. One of our key findings is that, for most data sets, applying no rebalancing strategy is competitive in terms of predictive performances, would it be with LightGBM, tuned random forests or logistic regression. However, when the imbalance ratio is artificially augmented, one of our two modifications of SMOTE leads to promising predictive performances compared to SMOTE and other state-of-the-art strategies. | stat-04.zip/2402.03819.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.04516 | Generalized Sobolev Transport for Probability Measures on a Graph | [
"Tam Le",
"Truyen Nguyen",
"Kenji Fukumizu"
] | 7 Feb 2024 (<a href="https://arxiv.org/abs/2402.04516v1">v1</a>), last revised 29 May 2024 (this version, v2) | To appear at ICML'2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.04516 | We study the optimal transport (OT) problem for measures supported on a graph metric space. Recently, Le et al. (2022) leverage the graph structure and propose a variant of OT, namely Sobolev transport (ST), which yields a closed-form expression for a fast computation. However, ST is essentially coupled with the $L^p$ geometric structure within its definition which makes it nontrivial to utilize ST for other prior structures. In contrast, the classic OT has the flexibility to adapt to various geometric structures by modifying the underlying cost function. An important instance is the Orlicz-Wasserstein (OW) which moves beyond the $L^p$ structure by leveraging the \emph{Orlicz geometric structure}. Comparing to the usage of standard $p$-order Wasserstein, OW remarkably helps to advance certain machine learning approaches. Nevertheless, OW brings up a new challenge on its computation due to its two-level optimization formulation. In this work, we leverage a specific class of convex functions for Orlicz structure to propose the generalized Sobolev transport (GST). GST encompasses the ST as its special case, and can be utilized for prior structures beyond the $L^p$ geometry. In connection with the OW, we show that one only needs to simply solve a univariate optimization problem to compute the GST, unlike the complex two-level optimization problem in OW. We empirically illustrate that GST is several-order faster than the OW. Moreover, we provide preliminary evidences on the advantages of GST for document classification and for several tasks in topological data analysis. | stat-04.zip/2402.04516.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.12668 | Randomization Can Reduce Both Bias and Variance: A Case Study in Random Forests | [
"Brian Liu",
"Rahul Mazumder"
] | 20 Feb 2024 (<a href="https://arxiv.org/abs/2402.12668v1">v1</a>), last revised 21 Jul 2025 (this version, v4) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.12668 | We study the often overlooked phenomenon, first noted in \cite{breiman2001random}, that random forests appear to reduce bias compared to bagging. Motivated by an interesting paper by \cite{mentch2020randomization}, where the authors explain the success of random forests in low signal-to-noise ratio (SNR) settings through regularization, we explore how random forests can capture patterns in the data that bagging ensembles fail to capture. We empirically demonstrate that in the presence of such patterns, random forests reduce bias along with variance and can increasingly outperform bagging ensembles when SNR is high. Our observations offer insights into the real-world success of random forests across a range of SNRs and enhance our understanding of the difference between random forests and bagging ensembles. Our investigations also yield practical insights into the importance of tuning $mtry$ in random forests. | stat-04.zip/2402.12668.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.03527 | Consistent Validation for Predictive Methods in Spatial Settings | [
"David R. Burt",
"Yunyi Shen",
"Tamara Broderick"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.03527v1">v1</a>), last revised 22 Mar 2025 (this version, v3) | 52 pages, 14 figures, Accepted at Artificial Intelligence and Statistics (AISTATS), 2025 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.03527 | Spatial prediction tasks are key to weather forecasting, studying air pollution impacts, and other scientific endeavors. Determining how much to trust predictions made by statistical or physical methods is essential for the credibility of scientific conclusions. Unfortunately, classical approaches for validation fail to handle mismatch between locations available for validation and (test) locations where we want to make predictions. This mismatch is often not an instance of covariate shift (as commonly formalized) because the validation and test locations are fixed (e.g., on a grid or at select points) rather than i.i.d. from two distributions. In the present work, we formalize a check on validation methods: that they become arbitrarily accurate as validation data becomes arbitrarily dense. We show that classical and covariate-shift methods can fail this check. We propose a method that builds from existing ideas in the covariate-shift literature, but adapts them to the validation data at hand. We prove that our proposal passes our check. And we demonstrate its advantages empirically on simulated and real data. | stat-04.zip/2402.03527.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.03724 | Statistical Test for Anomaly Detections by Variational Auto-Encoders | [
"Daiki Miwa",
"Tomohiro Shiraishi",
"Vo Nguyen Le Duy",
"Teruyuki Katsuoka",
"Ichiro Takeuchi"
] | 6 Feb 2024 (<a href="https://arxiv.org/abs/2402.03724v1">v1</a>), last revised 2 Jun 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.03724 | In this study, we consider the reliability assessment of anomaly detection (AD) using Variational Autoencoder (VAE). Over the last decade, VAE-based AD has been actively studied in various perspective, from method development to applied research. However, when the results of ADs are used in high-stakes decision-making, such as in medical diagnosis, it is necessary to ensure the reliability of the detected anomalies. In this study, we propose the VAE-AD Test as a method for quantifying the statistical reliability of VAE-based AD within the framework of statistical testing. Using the VAE-AD Test, the reliability of the anomaly regions detected by a VAE can be quantified in the form of p-values. This means that if an anomaly is declared when the p-value is below a certain threshold, it is possible to control the probability of false detection to a desired level. Since the VAE-AD Test is constructed based on a new statistical inference framework called selective inference, its validity is theoretically guaranteed in finite samples. To demonstrate the validity and effectiveness of the proposed VAE-AD Test, numerical experiments on artificial data and applications to brain image analysis are conducted. | stat-04.zip/2402.03724.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.04355 | PQMass: Probabilistic Assessment of the Quality of Generative Models using Probability Mass Estimation | [
"Pablo Lemos",
"Sammy Sharief",
"Nikolay Malkin",
"Salma Salhi",
"Connor Stone",
"Laurence Perreault-Levasseur",
"Yashar Hezaveh"
] | 6 Feb 2024 (<a href="https://arxiv.org/abs/2402.04355v1">v1</a>), last revised 10 Sep 2025 (this version, v3) | Published as a conference paper at ICLR 2025 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.04355 | We propose a likelihood-free method for comparing two distributions given samples from each, with the goal of assessing the quality of generative models. The proposed approach, PQMass, provides a statistically rigorous method for assessing the performance of a single generative model or the comparison of multiple competing models. PQMass divides the sample space into non-overlapping regions and applies chi-squared tests to the number of data samples that fall within each region, giving a p-value that measures the probability that the bin counts derived from two sets of samples are drawn from the same multinomial distribution. PQMass does not depend on assumptions regarding the density of the true distribution, nor does it rely on training or fitting any auxiliary models. We evaluate PQMass on data of various modalities and dimensions, demonstrating its effectiveness in assessing the quality, novelty, and diversity of generated samples. We further show that PQMass scales well to moderately high-dimensional data and thus obviates the need for feature extraction in practical applications. | stat-04.zip/2402.04355.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.15213 | Statistical Agnostic Regression: a machine learning method to validate regression models | [
"Juan M Gorriz",
"J. Ramirez",
"F. Segovia",
"F. J. Martinez-Murcia",
"C. Jiménez-Mesa",
"J. Suckling"
] | 23 Feb 2024 (<a href="https://arxiv.org/abs/2402.15213v1">v1</a>), last revised 9 Nov 2024 (this version, v3) | 23 pages, 18 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST); Computation (stat.CO) | https://doi.org/10.48550/arXiv.2402.15213 | Regression analysis is a central topic in statistical modeling, aimed at estimating the relationships between a dependent variable, commonly referred to as the response variable, and one or more independent variables, i.e., explanatory variables. Linear regression is by far the most popular method for performing this task in various fields of research, such as data integration and predictive modeling when combining information from multiple sources. Classical methods for solving linear regression problems, such as Ordinary Least Squares (OLS), Ridge, or Lasso regressions, often form the foundation for more advanced machine learning (ML) techniques, which have been successfully applied, though without a formal definition of statistical significance. At most, permutation or analyses based on empirical measures (e.g., residuals or accuracy) have been conducted, leveraging the greater sensitivity of ML estimations for detection. In this paper, we introduce Statistical Agnostic Regression (SAR) for evaluating the statistical significance of ML-based linear regression models. This is achieved by analyzing concentration inequalities of the actual risk (expected loss) and considering the worst-case scenario. To this end, we define a threshold that ensures there is sufficient evidence, with a probability of at least $1-\eta$, to conclude the existence of a linear relationship in the population between the explanatory (feature) and the response (label) variables. Simulations demonstrate the ability of the proposed agnostic (non-parametric) test to provide an analysis of variance similar to the classical multivariate $F$-test for the slope parameter, without relying on the underlying assumptions of classical methods. Moreover, the residuals computed from this method represent a trade-off between those obtained from ML approaches and the classical OLS. | stat-04.zip/2402.15213.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.02969 | Towards Understanding the Word Sensitivity of Attention Layers: A Study via Random Features | [
"Simone Bombari",
"Marco Mondelli"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.02969v1">v1</a>), last revised 17 May 2024 (this version, v2) | Revision after ICML2024 reviews | Machine Learning (stat.ML) | Machine Learning (stat.ML); Computation and Language (cs.CL); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.02969 | Understanding the reasons behind the exceptional success of transformers requires a better analysis of why attention layers are suitable for NLP tasks. In particular, such tasks require predictive models to capture contextual meaning which often depends on one or few words, even if the sentence is long. Our work studies this key property, dubbed word sensitivity (WS), in the prototypical setting of random features. We show that attention layers enjoy high WS, namely, there exists a vector in the space of embeddings that largely perturbs the random attention features map. The argument critically exploits the role of the softmax in the attention layer, highlighting its benefit compared to other activations (e.g., ReLU). In contrast, the WS of standard random features is of order $1/\sqrt{n}$, $n$ being the number of words in the textual sample, and thus it decays with the length of the context. We then translate these results on the word sensitivity into generalization bounds: due to their low WS, random features provably cannot learn to distinguish between two sentences that differ only in a single word; in contrast, due to their high WS, random attention features have higher generalization capabilities. We validate our theoretical results with experimental evidence over the BERT-Base word embeddings of the imdb review dataset. | stat-04.zip/2402.02969.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.04038 | PAC-Bayesian Adversarially Robust Generalization Bounds for Graph Neural Network | [
"Tan Sun",
"Junhong Lin"
] | 6 Feb 2024 (<a href="https://arxiv.org/abs/2402.04038v1">v1</a>), last revised 6 Jul 2024 (this version, v2) | 38pages | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.04038 | Graph neural networks (GNNs) have gained popularity for various graph-related tasks. However, similar to deep neural networks, GNNs are also vulnerable to adversarial attacks. Empirical studies have shown that adversarially robust generalization has a pivotal role in establishing effective defense algorithms against adversarial attacks. In this paper, we contribute by providing adversarially robust generalization bounds for two kinds of popular GNNs, graph convolutional network (GCN) and message passing graph neural network, using the PAC-Bayesian framework. Our result reveals that spectral norm of the diffusion matrix on the graph and spectral norm of the weights as well as the perturbation factor govern the robust generalization bounds of both models. Our bounds are nontrivial generalizations of the results developed in (Liao et al., 2020) from the standard setting to adversarial setting while avoiding exponential dependence of the maximum node degree. As corollaries, we derive better PAC-Bayesian robust generalization bounds for GCN in the standard setting, which improve the bounds in (Liao et al., 2020) by avoiding exponential dependence on the maximum node degree. | stat-04.zip/2402.04038.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.14264 | Structure-agnostic Optimality of Doubly Robust Learning for Treatment Effect Estimation | [
"Jikai Jin",
"Vasilis Syrgkanis"
] | 22 Feb 2024 (<a href="https://arxiv.org/abs/2402.14264v1">v1</a>), last revised 8 Jun 2025 (this version, v4) | 31 pages, to appear in COLT 2025 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Econometrics (econ.EM); Statistics Theory (math.ST); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.14264 | Average treatment effect estimation is the most central problem in causal inference with application to numerous disciplines. While many estimation strategies have been proposed in the literature, the statistical optimality of these methods has still remained an open area of investigation, especially in regimes where these methods do not achieve parametric rates. In this paper, we adopt the recently introduced structure-agnostic framework of statistical lower bounds, which poses no structural properties on the nuisance functions other than access to black-box estimators that achieve some statistical estimation rate. This framework is particularly appealing when one is only willing to consider estimation strategies that use non-parametric regression and classification oracles as black-box sub-processes. Within this framework, we prove the statistical optimality of the celebrated and widely used doubly robust estimators for both the Average Treatment Effect (ATE) and the Average Treatment Effect on the Treated (ATT), as well as weighted variants of the former, which arise in policy evaluation. | stat-04.zip/2402.14264.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.03447 | Challenges in Variable Importance Ranking Under Correlation | [
"Annie Liang",
"Thomas Jemielita",
"Andy Liaw",
"Vladimir Svetnik",
"Lingkang Huang",
"Richard Baumgartner",
"Jason M. Klusowski"
] | 5 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.03447 | Variable importance plays a pivotal role in interpretable machine learning as it helps measure the impact of factors on the output of the prediction model. Model agnostic methods based on the generation of "null" features via permutation (or related approaches) can be applied. Such analysis is often utilized in pharmaceutical applications due to its ability to interpret black-box models, including tree-based ensembles. A major challenge and significant confounder in variable importance estimation however is the presence of between-feature correlation. Recently, several adjustments to marginal permutation utilizing feature knockoffs were proposed to address this issue, such as the variable importance measure known as conditional predictive impact (CPI). Assessment and evaluation of such approaches is the focus of our work. We first present a comprehensive simulation study investigating the impact of feature correlation on the assessment of variable importance. We then theoretically prove the limitation that highly correlated features pose for the CPI through the knockoff construction. While we expect that there is always no correlation between knockoff variables and its corresponding predictor variables, we prove that the correlation increases linearly beyond a certain correlation threshold between the predictor variables. Our findings emphasize the absence of free lunch when dealing with high feature correlation, as well as the necessity of understanding the utility and limitations behind methods in variable importance estimation. | stat-04.zip/2402.03447.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.03104 | High-dimensional Bayesian Optimization via Covariance Matrix Adaptation Strategy | [
"Lam Ngo",
"Huong Ha",
"Jeffrey Chan",
"Vu Nguyen",
"Hongyu Zhang"
] | 5 Feb 2024 | 31 pages, 17 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.03104 | Bayesian Optimization (BO) is an effective method for finding the global optimum of expensive black-box functions. However, it is well known that applying BO to high-dimensional optimization problems is challenging. To address this issue, a promising solution is to use a local search strategy that partitions the search domain into local regions with high likelihood of containing the global optimum, and then use BO to optimize the objective function within these regions. In this paper, we propose a novel technique for defining the local regions using the Covariance Matrix Adaptation (CMA) strategy. Specifically, we use CMA to learn a search distribution that can estimate the probabilities of data points being the global optimum of the objective function. Based on this search distribution, we then define the local regions consisting of data points with high probabilities of being the global optimum. Our approach serves as a meta-algorithm as it can incorporate existing black-box BO optimizers, such as BO, TuRBO, and BAxUS, to find the global optimum of the objective function within our derived local regions. We evaluate our proposed method on various benchmark synthetic and real-world problems. The results demonstrate that our method outperforms existing state-of-the-art techniques. | stat-04.zip/2402.03104.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.15625 | Learning Cyclic Causal Models from Incomplete Data | [
"Muralikrishnna G. Sethuraman",
"Faramarz Fekri"
] | 23 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.15625 | Causal learning is a fundamental problem in statistics and science, offering insights into predicting the effects of unseen treatments on a system. Despite recent advances in this topic, most existing causal discovery algorithms operate under two key assumptions: (i) the underlying graph is acyclic, and (ii) the available data is complete. These assumptions can be problematic as many real-world systems contain feedback loops (e.g., biological systems), and practical scenarios frequently involve missing data. In this work, we propose a novel framework, named MissNODAGS, for learning cyclic causal graphs from partially missing data. Under the additive noise model, MissNODAGS learns the causal graph by alternating between imputing the missing data and maximizing the expected log-likelihood of the visible part of the data in each training step, following the principles of the expectation-maximization (EM) framework. Through synthetic experiments and real-world single-cell perturbation data, we demonstrate improved performance when compared to using state-of-the-art imputation techniques followed by causal learning on partially missing interventional data. | stat-04.zip/2402.15625.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.02952 | On Least Square Estimation in Softmax Gating Mixture of Experts | [
"Huy Nguyen",
"Nhat Ho",
"Alessandro Rinaldo"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.02952v1">v1</a>), last revised 24 Jun 2024 (this version, v2) | Accepted to ICML 2024, 29 pages, 2 figures, 2 tables | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.02952 | Mixture of experts (MoE) model is a statistical machine learning design that aggregates multiple expert networks using a softmax gating function in order to form a more intricate and expressive model. Despite being commonly used in several applications owing to their scalability, the mathematical and statistical properties of MoE models are complex and difficult to analyze. As a result, previous theoretical works have primarily focused on probabilistic MoE models by imposing the impractical assumption that the data are generated from a Gaussian MoE model. In this work, we investigate the performance of the least squares estimators (LSE) under a deterministic MoE model where the data are sampled according to a regression model, a setting that has remained largely unexplored. We establish a condition called strong identifiability to characterize the convergence behavior of various types of expert functions. We demonstrate that the rates for estimating strongly identifiable experts, namely the widely used feed-forward networks with activation functions $\mathrm{sigmoid}(\cdot)$ and $\tanh(\cdot)$, are substantially faster than those of polynomial experts, which we show to exhibit a surprising slow estimation rate. Our findings have important practical implications for expert selection. | stat-04.zip/2402.02952.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.03485 | Attention Meets Post-hoc Interpretability: A Mathematical Perspective | [
"Gianluigi Lopardo",
"Frederic Precioso",
"Damien Garreau"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.03485v1">v1</a>), last revised 17 Jun 2024 (this version, v2) | Accepted at ICML 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Computation and Language (cs.CL); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.03485 | Attention-based architectures, in particular transformers, are at the heart of a technological revolution. Interestingly, in addition to helping obtain state-of-the-art results on a wide range of applications, the attention mechanism intrinsically provides meaningful insights on the internal behavior of the model. Can these insights be used as explanations? Debate rages on. In this paper, we mathematically study a simple attention-based architecture and pinpoint the differences between post-hoc and attention-based explanations. We show that they provide quite different results, and that, despite their limitations, post-hoc methods are capable of capturing more useful insights than merely examining the attention weights. | stat-04.zip/2402.03485.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.02459 | On Minimum Trace Factor Analysis -- An Old Song Sung to a New Tune | [
"C. Li",
"A. Shkolnik"
] | 4 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.02459 | Dimensionality reduction methods, such as principal component analysis (PCA) and factor analysis, are central to many problems in data science. There are, however, serious and well-understood challenges to finding robust low dimensional approximations for data with significant heteroskedastic noise. This paper introduces a relaxed version of Minimum Trace Factor Analysis (MTFA), a convex optimization method with roots dating back to the work of Ledermann in 1940. This relaxation is particularly effective at not overfitting to heteroskedastic perturbations and addresses the commonly cited Heywood cases in factor analysis and the recently identified "curse of ill-conditioning" for existing spectral methods. We provide theoretical guarantees on the accuracy of the resulting low rank subspace and the convergence rate of the proposed algorithm to compute that matrix. We develop a number of interesting connections to existing methods, including HeteroPCA, Lasso, and Soft-Impute, to fill an important gap in the already large literature on low rank matrix estimation. Numerical experiments benchmark our results against several recent proposals for dealing with heteroskedastic noise. | stat-04.zip/2402.02459.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.03254 | Minimum Description Length and Generalization Guarantees for Representation Learning | [
"Milad Sefidgaran",
"Abdellatif Zaidi",
"Piotr Krasnowski"
] | 5 Feb 2024 | Accepted and presented at NeurIPS 2023 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Information Theory (cs.IT); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.03254 | A major challenge in designing efficient statistical supervised learning algorithms is finding representations that perform well not only on available training samples but also on unseen data. While the study of representation learning has spurred much interest, most existing such approaches are heuristic; and very little is known about theoretical generalization guarantees. <br>In this paper, we establish a compressibility framework that allows us to derive upper bounds on the generalization error of a representation learning algorithm in terms of the "Minimum Description Length" (MDL) of the labels or the latent variables (representations). Rather than the mutual information between the encoder's input and the representation, which is often believed to reflect the algorithm's generalization capability in the related literature but in fact, falls short of doing so, our new bounds involve the "multi-letter" relative entropy between the distribution of the representations (or labels) of the training and test sets and a fixed prior. In particular, these new bounds reflect the structure of the encoder and are not vacuous for deterministic algorithms. Our compressibility approach, which is information-theoretic in nature, builds upon that of Blum-Langford for PAC-MDL bounds and introduces two essential ingredients: block-coding and lossy-compression. The latter allows our approach to subsume the so-called geometrical compressibility as a special case. To the best knowledge of the authors, the established generalization bounds are the first of their kind for Information Bottleneck (IB) type encoders and representation learning. Finally, we partly exploit the theoretical results by introducing a new data-dependent prior. Numerical simulations illustrate the advantages of well-chosen such priors over classical priors used in IB. | stat-04.zip/2402.03254.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.14777 | Causal Imputation for Counterfactual SCMs: Bridging Graphs and Latent Factor Models | [
"Alvaro Ribot",
"Chandler Squires",
"Caroline Uhler"
] | 22 Feb 2024 | 35 pages, 17 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.14777 | We consider the task of causal imputation, where we aim to predict the outcomes of some set of actions across a wide range of possible contexts. As a running example, we consider predicting how different drugs affect cells from different cell types. We study the index-only setting, where the actions and contexts are categorical variables with a finite number of possible values. Even in this simple setting, a practical challenge arises, since often only a small subset of possible action-context pairs have been studied. Thus, models must extrapolate to novel action-context pairs, which can be framed as a form of matrix completion with rows indexed by actions, columns indexed by contexts, and matrix entries corresponding to outcomes. We introduce a novel SCM-based model class, where the outcome is expressed as a counterfactual, actions are expressed as interventions on an instrumental variable, and contexts are defined based on the initial state of the system. We show that, under a linearity assumption, this setup induces a latent factor model over the matrix of outcomes, with an additional fixed effect term. To perform causal prediction based on this model class, we introduce simple extension to the Synthetic Interventions estimator (Agarwal et al., 2020). We evaluate several matrix completion approaches on the PRISM drug repurposing dataset, showing that our method outperforms all other considered matrix completion approaches. | stat-04.zip/2402.14777.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.02857 | Non-asymptotic Analysis of Biased Adaptive Stochastic Approximation | [
"Sobihan Surendran",
"Antoine Godichon-Baggioni",
"Adeline Fermanian",
"Sylvain Le Corff"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.02857v1">v1</a>), last revised 14 Mar 2025 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.02857 | Stochastic Gradient Descent (SGD) with adaptive steps is widely used to train deep neural networks and generative models. Most theoretical results assume that it is possible to obtain unbiased gradient estimators, which is not the case in several recent deep learning and reinforcement learning applications that use Monte Carlo methods. This paper provides a comprehensive non-asymptotic analysis of SGD with biased gradients and adaptive steps for non-convex smooth functions. Our study incorporates time-dependent bias and emphasizes the importance of controlling the bias of the gradient estimator. In particular, we establish that Adagrad, RMSProp, and AMSGRAD, an exponential moving average variant of Adam, with biased gradients, converge to critical points for smooth non-convex functions at a rate similar to existing results in the literature for the unbiased case. Finally, we provide experimental results using Variational Autoenconders (VAE) and applications to several learning frameworks that illustrate our convergence results and show how the effect of bias can be reduced by appropriate hyperparameter tuning. | stat-04.zip/2402.02857.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.14987 | On the Performance of Empirical Risk Minimization with Smoothed Data | [
"Adam Block",
"Alexander Rakhlin",
"Abhishek Shetty"
] | 22 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.14987 | In order to circumvent statistical and computational hardness results in sequential decision-making, recent work has considered smoothed online learning, where the distribution of data at each time is assumed to have bounded likeliehood ratio with respect to a base measure when conditioned on the history. While previous works have demonstrated the benefits of smoothness, they have either assumed that the base measure is known to the learner or have presented computationally inefficient algorithms applying only in special cases. This work investigates the more general setting where the base measure is \emph{unknown} to the learner, focusing in particular on the performance of Empirical Risk Minimization (ERM) with square loss when the data are well-specified and smooth. We show that in this setting, ERM is able to achieve sublinear error whenever a class is learnable with iid data; in particular, ERM achieves error scaling as $\tilde O( \sqrt{\mathrm{comp}(\mathcal F)\cdot T} )$, where $\mathrm{comp}(\mathcal F)$ is the statistical complexity of learning $\mathcal F$ with iid data. In so doing, we prove a novel norm comparison bound for smoothed data that comprises the first sharp norm comparison for dependent data applying to arbitrary, nonlinear function classes. We complement these results with a lower bound indicating that our analysis of ERM is essentially tight, establishing a separation in the performance of ERM between smoothed and iid data. | stat-04.zip/2402.14987.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.03460 | Approximation Rates and VC-Dimension Bounds for (P)ReLU MLP Mixture of Experts | [
"Anastasis Kratsios",
"Haitz Sáez de Ocáriz Borde",
"Takashi Furuya",
"Marc T. Law"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.03460v1">v1</a>), last revised 25 May 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE); Combinatorics (math.CO); Numerical Analysis (math.NA) | https://doi.org/10.48550/arXiv.2402.03460 | Mixture-of-Experts (MoEs) can scale up beyond traditional deep learning models by employing a routing strategy in which each input is processed by a single "expert" deep learning model. This strategy allows us to scale up the number of parameters defining the MoE while maintaining sparse activation, i.e., MoEs only load a small number of their total parameters into GPU VRAM for the forward pass depending on the input. In this paper, we provide an approximation and learning-theoretic analysis of mixtures of expert MLPs with (P)ReLU activation functions. We first prove that for every error level $\varepsilon>0$ and every Lipschitz function $f:[0,1]^n\to \mathbb{R}$, one can construct a MoMLP model (a Mixture-of-Experts comprising of (P)ReLU MLPs) which uniformly approximates $f$ to $\varepsilon$ accuracy over $[0,1]^n$, while only requiring networks of $\mathcal{O}(\varepsilon^{-1})$ parameters to be loaded in memory. Additionally, we show that MoMLPs can generalize since the entire MoMLP model has a (finite) VC dimension of $\tilde{O}(L\max\{nL,JW\})$, if there are $L$ experts and each expert has a depth and width of $J$ and $W$, respectively. | stat-04.zip/2402.03460.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.03008 | Diffusive Gibbs Sampling | [
"Wenlin Chen",
"Mingtian Zhang",
"Brooks Paige",
"José Miguel Hernández-Lobato",
"David Barber"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.03008v1">v1</a>), last revised 29 May 2024 (this version, v5) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Computation (stat.CO) | https://doi.org/10.48550/arXiv.2402.03008 | The inadequate mixing of conventional Markov Chain Monte Carlo (MCMC) methods for multi-modal distributions presents a significant challenge in practical applications such as Bayesian inference and molecular dynamics. Addressing this, we propose Diffusive Gibbs Sampling (DiGS), an innovative family of sampling methods designed for effective sampling from distributions characterized by distant and disconnected modes. DiGS integrates recent developments in diffusion models, leveraging Gaussian convolution to create an auxiliary noisy distribution that bridges isolated modes in the original space and applying Gibbs sampling to alternately draw samples from both spaces. A novel Metropolis-within-Gibbs scheme is proposed to enhance mixing in the denoising sampling step. DiGS exhibits a better mixing property for sampling multi-modal distributions than state-of-the-art methods such as parallel tempering, attaining substantially improved performance across various tasks, including mixtures of Gaussians, Bayesian neural networks and molecular dynamics. | stat-04.zip/2402.03008.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.03220 | The Benefits of Reusing Batches for Gradient Descent in Two-Layer Networks: Breaking the Curse of Information and Leap Exponents | [
"Yatin Dandi",
"Emanuele Troiani",
"Luca Arnaboldi",
"Luca Pesce",
"Lenka Zdeborová",
"Florent Krzakala"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.03220v1">v1</a>), last revised 30 Jun 2024 (this version, v3) | Accepted at the International Conference on Machine Learning (ICML), 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.03220 | We investigate the training dynamics of two-layer neural networks when learning multi-index target functions. We focus on multi-pass gradient descent (GD) that reuses the batches multiple times and show that it significantly changes the conclusion about which functions are learnable compared to single-pass gradient descent. In particular, multi-pass GD with finite stepsize is found to overcome the limitations of gradient flow and single-pass GD given by the information exponent (Ben Arous et al., 2021) and leap exponent (Abbe et al., 2023) of the target function. We show that upon re-using batches, the network achieves in just two time steps an overlap with the target subspace even for functions not satisfying the staircase property (Abbe et al., 2021). We characterize the (broad) class of functions efficiently learned in finite time. The proof of our results is based on the analysis of the Dynamical Mean-Field Theory (DMFT). We further provide a closed-form description of the dynamical process of the low-dimensional projections of the weights, and numerical experiments illustrating the theory. | stat-04.zip/2402.03220.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.13622 | Analysis of Bootstrap and Subsampling in High-dimensional Regularized Regression | [
"Lucas Clarté",
"Adrien Vandenbroucque",
"Guillaume Dalle",
"Bruno Loureiro",
"Florent Krzakala",
"Lenka Zdeborová"
] | 21 Feb 2024 (<a href="https://arxiv.org/abs/2402.13622v1">v1</a>), last revised 1 Nov 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Disordered Systems and Neural Networks (cond-mat.dis-nn); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.13622 | We investigate popular resampling methods for estimating the uncertainty of statistical models, such as subsampling, bootstrap and the jackknife, and their performance in high-dimensional supervised regression tasks. We provide a tight asymptotic description of the biases and variances estimated by these methods in the context of generalized linear models, such as ridge and logistic regression, taking the limit where the number of samples $n$ and dimension $d$ of the covariates grow at a comparable fixed rate $\alpha\!=\! n/d$. Our findings are three-fold: i) resampling methods are fraught with problems in high dimensions and exhibit the double-descent-like behavior typical of these situations; ii) only when $\alpha$ is large enough do they provide consistent and reliable error estimations (we give convergence rates); iii) in the over-parametrized regime $\alpha\!<\!1$ relevant to modern machine learning practice, their predictions are not consistent, even with optimal regularization. | stat-04.zip/2402.13622.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.02817 | Bayes-Optimal Fair Classification with Linear Disparity Constraints via Pre-, In-, and Post-processing | [
"Xianli Zeng",
"Kevin Jiang",
"Guang Cheng",
"Edgar Dobriban"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.02817v1">v1</a>), last revised 27 Aug 2025 (this version, v3) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Computers and Society (cs.CY); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.02817 | Machine learning algorithms may have disparate impacts on protected groups. To address this, we develop methods for Bayes-optimal fair classification, aiming to minimize classification error subject to given group fairness constraints. We introduce the notion of \emph{linear disparity measures}, which are linear functions of a probabilistic classifier; and \emph{bilinear disparity measures}, which are also linear in the group-wise regression functions. We show that several popular disparity measures -- the deviations from demographic parity, equality of opportunity, and predictive equality -- are bilinear. <br>We find the form of Bayes-optimal fair classifiers under a single linear disparity measure, by uncovering a connection with the Neyman-Pearson lemma. For bilinear disparity measures, we are able to find the explicit form of Bayes-optimal fair classifiers as group-wise thresholding rules with explicitly characterized thresholds. We develop similar algorithms for when protected attribute cannot be used at the prediction phase. Moreover, we obtain analogous theoretical characterizations of optimal classifiers for a multi-class protected attribute and for equalized odds. <br>Leveraging our theoretical results, we design methods that learn fair Bayes-optimal classifiers under bilinear disparity constraints. Our methods cover three popular approaches to fairness-aware classification, via pre-processing (Fair Up- and Down-Sampling), in-processing (Fair cost-sensitive Classification) and post-processing (a Fair Plug-In Rule). Our methods control disparity directly while achieving near-optimal fairness-accuracy tradeoffs. We show empirically that our methods have state-of-the-art performance compared to existing algorithms. In particular, our pre-processing method can a reach higher accuracy than prior pre-processing methods at low disparity levels. | stat-04.zip/2402.02817.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.02862 | Graph Neural Machine: A New Model for Learning with Tabular Data | [
"Giannis Nikolentzos",
"Siyun Wang",
"Johannes Lutzeyer",
"Michalis Vazirgiannis"
] | 5 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.02862 | In recent years, there has been a growing interest in mapping data from different domains to graph structures. Among others, neural network models such as the multi-layer perceptron (MLP) can be modeled as graphs. In fact, MLPs can be represented as directed acyclic graphs. Graph neural networks (GNNs) have recently become the standard tool for performing machine learning tasks on graphs. In this work, we show that an MLP is equivalent to an asynchronous message passing GNN model which operates on the MLP's graph representation. We then propose a new machine learning model for tabular data, the so-called Graph Neural Machine (GNM), which replaces the MLP's directed acyclic graph with a nearly complete graph and which employs a synchronous message passing scheme. We show that a single GNM model can simulate multiple MLP models. We evaluate the proposed model in several classification and regression datasets. In most cases, the GNM model outperforms the MLP architecture. | stat-04.zip/2402.02862.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.15285 | Generative Modelling with Tensor Train approximations of Hamilton--Jacobi--Bellman equations | [
"David Sommer",
"Robert Gruhlke",
"Max Kirstein",
"Martin Eigel",
"Claudia Schillings"
] | 23 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST) | https://doi.org/10.48550/arXiv.2402.15285 | Sampling from probability densities is a common challenge in fields such as Uncertainty Quantification (UQ) and Generative Modelling (GM). In GM in particular, the use of reverse-time diffusion processes depending on the log-densities of Ornstein-Uhlenbeck forward processes are a popular sampling tool. In Berner et al. [2022] the authors point out that these log-densities can be obtained by solution of a \textit{Hamilton-Jacobi-Bellman} (HJB) equation known from stochastic optimal control. While this HJB equation is usually treated with indirect methods such as policy iteration and unsupervised training of black-box architectures like Neural Networks, we propose instead to solve the HJB equation by direct time integration, using compressed polynomials represented in the Tensor Train (TT) format for spatial discretization. Crucially, this method is sample-free, agnostic to normalization constants and can avoid the curse of dimensionality due to the TT compression. We provide a complete derivation of the HJB equation's action on Tensor Train polynomials and demonstrate the performance of the proposed time-step-, rank- and degree-adaptive integration method on a nonlinear sampling task in 20 dimensions. | stat-04.zip/2402.15285.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.01972 | Combining T-learning and DR-learning: a framework for oracle-efficient estimation of causal contrasts | [
"Lars van der Laan",
"Marco Carone",
"Alex Luedtke"
] | 3 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.01972 | We introduce efficient plug-in (EP) learning, a novel framework for the estimation of heterogeneous causal contrasts, such as the conditional average treatment effect and conditional relative risk. The EP-learning framework enjoys the same oracle-efficiency as Neyman-orthogonal learning strategies, such as DR-learning and R-learning, while addressing some of their primary drawbacks, including that (i) their practical applicability can be hindered by loss function non-convexity; and (ii) they may suffer from poor performance and instability due to inverse probability weighting and pseudo-outcomes that violate bounds. To avoid these drawbacks, EP-learner constructs an efficient plug-in estimator of the population risk function for the causal contrast, thereby inheriting the stability and robustness properties of plug-in estimation strategies like T-learning. Under reasonable conditions, EP-learners based on empirical risk minimization are oracle-efficient, exhibiting asymptotic equivalence to the minimizer of an oracle-efficient one-step debiased estimator of the population risk function. In simulation experiments, we illustrate that EP-learners of the conditional average treatment effect and conditional relative risk outperform state-of-the-art competitors, including T-learner, R-learner, and DR-learner. Open-source implementations of the proposed methods are available in our R package hte3. | stat-04.zip/2402.01972.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.02098 | Self-attention Networks Localize When QK-eigenspectrum Concentrates | [
"Han Bao",
"Ryuichiro Hataya",
"Ryo Karakida"
] | 3 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.02098 | The self-attention mechanism prevails in modern machine learning. It has an interesting functionality of adaptively selecting tokens from an input sequence by modulating the degree of attention localization, which many researchers speculate is the basis of the powerful model performance but complicates the underlying mechanism of the learning dynamics. In recent years, mainly two arguments have connected attention localization to the model performances. One is the rank collapse, where the embedded tokens by a self-attention block become very similar across different tokens, leading to a less expressive network. The other is the entropy collapse, where the attention probability approaches non-uniform and entails low entropy, making the learning dynamics more likely to be trapped in plateaus. These two failure modes may apparently contradict each other because the rank and entropy collapses are relevant to uniform and non-uniform attention, respectively. To this end, we characterize the notion of attention localization by the eigenspectrum of query-key parameter matrices and reveal that a small eigenspectrum variance leads attention to be localized. Interestingly, the small eigenspectrum variance prevents both rank and entropy collapse, leading to better model expressivity and trainability. | stat-04.zip/2402.02098.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.15115 | Physics-constrained polynomial chaos expansion for scientific machine learning and uncertainty quantification | [
"Himanshu Sharma",
"Lukáš Novák",
"Michael D. Shields"
] | 23 Feb 2024 (<a href="https://arxiv.org/abs/2402.15115v1">v1</a>), last revised 11 May 2024 (this version, v2) | 34 pages, 15 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Data Analysis, Statistics and Probability (physics.data-an) | https://doi.org/10.48550/arXiv.2402.15115 | We present a novel physics-constrained polynomial chaos expansion as a surrogate modeling method capable of performing both scientific machine learning (SciML) and uncertainty quantification (UQ) tasks. The proposed method possesses a unique capability: it seamlessly integrates SciML into UQ and vice versa, which allows it to quantify the uncertainties in SciML tasks effectively and leverage SciML for improved uncertainty assessment during UQ-related tasks. The proposed surrogate model can effectively incorporate a variety of physical constraints, such as governing partial differential equations (PDEs) with associated initial and boundary conditions constraints, inequality-type constraints (e.g., monotonicity, convexity, non-negativity, among others), and additional a priori information in the training process to supplement limited data. This ensures physically realistic predictions and significantly reduces the need for expensive computational model evaluations to train the surrogate model. Furthermore, the proposed method has a built-in uncertainty quantification (UQ) feature to efficiently estimate output uncertainties. To demonstrate the effectiveness of the proposed method, we apply it to a diverse set of problems, including linear/non-linear PDEs with deterministic and stochastic parameters, data-driven surrogate modeling of a complex physical system, and UQ of a stochastic system with parameters modeled as random fields. | stat-04.zip/2402.15115.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.14758 | Batch and match: black-box variational inference with a score-based divergence | [
"Diana Cai",
"Chirag Modi",
"Loucas Pillaud-Vivien",
"Charles C. Margossian",
"Robert M. Gower",
"David M. Blei",
"Lawrence K. Saul"
] | 22 Feb 2024 (<a href="https://arxiv.org/abs/2402.14758v1">v1</a>), last revised 12 Jun 2024 (this version, v2) | 49 pages, 14 figures. To appear in the Proceedings of the 41st International Conference on Machine Learning (ICML), 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Computation (stat.CO) | https://doi.org/10.48550/arXiv.2402.14758 | Most leading implementations of black-box variational inference (BBVI) are based on optimizing a stochastic evidence lower bound (ELBO). But such approaches to BBVI often converge slowly due to the high variance of their gradient estimates and their sensitivity to hyperparameters. In this work, we propose batch and match (BaM), an alternative approach to BBVI based on a score-based divergence. Notably, this score-based divergence can be optimized by a closed-form proximal update for Gaussian variational families with full covariance matrices. We analyze the convergence of BaM when the target distribution is Gaussian, and we prove that in the limit of infinite batch size the variational parameter updates converge exponentially quickly to the target mean and covariance. We also evaluate the performance of BaM on Gaussian and non-Gaussian target distributions that arise from posterior inference in hierarchical and deep generative models. In these experiments, we find that BaM typically converges in fewer (and sometimes significantly fewer) gradient evaluations than leading implementations of BBVI based on ELBO maximization. | stat-04.zip/2402.14758.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.01382 | Emergence of heavy tails in homogenized stochastic gradient descent | [
"Zhe Jiao",
"Martin Keller-Ressel"
] | 2 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01382 | It has repeatedly been observed that loss minimization by stochastic gradient descent (SGD) leads to heavy-tailed distributions of neural network parameters. Here, we analyze a continuous diffusion approximation of SGD, called homogenized stochastic gradient descent, show that it behaves asymptotically heavy-tailed, and give explicit upper and lower bounds on its tail-index. We validate these bounds in numerical experiments and show that they are typically close approximations to the empirical tail-index of SGD iterates. In addition, their explicit form enables us to quantify the interplay between optimization parameters and the tail-index. Doing so, we contribute to the ongoing discussion on links between heavy tails and the generalization performance of neural networks as well as the ability of SGD to avoid suboptimal local minima. | stat-04.zip/2402.01382.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.16811 | Stopping Bayesian Optimization with Probabilistic Regret Bounds | [
"James T. Wilson"
] | 26 Feb 2024 (<a href="https://arxiv.org/abs/2402.16811v1">v1</a>), last revised 10 Dec 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.16811 | Bayesian optimization is a popular framework for efficiently tackling black-box search problems. As a rule, these algorithms operate by iteratively choosing what to evaluate next until some predefined budget has been exhausted. We investigate replacing this de facto stopping rule with criteria based on the probability that a point satisfies a given set of conditions. We focus on the prototypical example of an $(\epsilon, \delta)$-criterion: stop when a solution has been found whose value is within $\epsilon > 0$ of the optimum with probability at least $1 - \delta$ under the model. For Gaussian process priors, we show that Bayesian optimization satisfies this criterion under mild technical assumptions. Further, we give a practical algorithm for evaluating Monte Carlo stopping rules in a manner that is both sample efficient and robust to estimation error. These findings are accompanied by empirical results which demonstrate the strengths and weaknesses of the proposed approach. | stat-04.zip/2402.16811.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.01502 | Why do Random Forests Work? Understanding Tree Ensembles as Self-Regularizing Adaptive Smoothers | [
"Alicia Curth",
"Alan Jeffares",
"Mihaela van der Schaar"
] | 2 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01502 | Despite their remarkable effectiveness and broad application, the drivers of success underlying ensembles of trees are still not fully understood. In this paper, we highlight how interpreting tree ensembles as adaptive and self-regularizing smoothers can provide new intuition and deeper insight to this topic. We use this perspective to show that, when studied as smoothers, randomized tree ensembles not only make predictions that are quantifiably more smooth than the predictions of the individual trees they consist of, but also further regulate their smoothness at test-time based on the dissimilarity between testing and training inputs. First, we use this insight to revisit, refine and reconcile two recent explanations of forest success by providing a new way of quantifying the conjectured behaviors of tree ensembles objectively by measuring the effective degree of smoothing they imply. Then, we move beyond existing explanations for the mechanisms by which tree ensembles improve upon individual trees and challenge the popular wisdom that the superior performance of forests should be understood as a consequence of variance reduction alone. We argue that the current high-level dichotomy into bias- and variance-reduction prevalent in statistics is insufficient to understand tree ensembles -- because the prevailing definition of bias does not capture differences in the expressivity of the hypothesis classes formed by trees and forests. Instead, we show that forests can improve upon trees by three distinct mechanisms that are usually implicitly entangled. In particular, we demonstrate that the smoothing effect of ensembling can reduce variance in predictions due to noise in outcome generation, reduce variability in the quality of the learned function given fixed input data and reduce potential bias in learnable functions by enriching the available hypothesis space. | stat-04.zip/2402.01502.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.01900 | Distributional Off-policy Evaluation with Bellman Residual Minimization | [
"Sungee Hong",
"Zhengling Qi",
"Raymond K. W. Wong"
] | 2 Feb 2024 (<a href="https://arxiv.org/abs/2402.01900v1">v1</a>), last revised 12 Mar 2025 (this version, v3) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01900 | We study distributional off-policy evaluation (OPE), of which the goal is to learn the distribution of the return for a target policy using offline data generated by a different policy. The theoretical foundation of many existing work relies on the supremum-extended statistical distances such as supremum-Wasserstein distance, which are hard to estimate. In contrast, we study the more manageable expectation-extended statistical distances and provide a novel theoretical justification on their validity for learning the return distribution. Based on this attractive property, we propose a new method called Energy Bellman Residual Minimizer (EBRM) for distributional OPE. We provide corresponding in-depth theoretical analyses. We establish a finite-sample error bound for the EBRM estimator under the realizability assumption. Furthermore, we introduce a variant of our method based on a multi-step extension which improves the error bound for non-realizable settings. Notably, unlike prior distributional OPE methods, the theoretical guarantees of our method do not require the completeness assumption. | stat-04.zip/2402.01900.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.01887 | On $f$-Divergence Principled Domain Adaptation: An Improved Framework | [
"Ziqiao Wang",
"Yongyi Mao"
] | 2 Feb 2024 (<a href="https://arxiv.org/abs/2402.01887v1">v1</a>), last revised 27 Oct 2024 (this version, v2) | Accepted at NeurIPS 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01887 | Unsupervised domain adaptation (UDA) plays a crucial role in addressing distribution shifts in machine learning. In this work, we improve the theoretical foundations of UDA proposed in Acuna et al. (2021) by refining their $f$-divergence-based discrepancy and additionally introducing a new measure, $f$-domain discrepancy ($f$-DD). By removing the absolute value function and incorporating a scaling parameter, $f$-DD obtains novel target error and sample complexity bounds, allowing us to recover previous KL-based results and bridging the gap between algorithms and theory presented in Acuna et al. (2021). Using a localization technique, we also develop a fast-rate generalization bound. Empirical results demonstrate the superior performance of $f$-DD-based learning algorithms over previous works in popular UDA benchmarks. | stat-04.zip/2402.01887.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.01258 | Transformers Learn Nonlinear Features In Context: Nonconvex Mean-field Dynamics on the Attention Landscape | [
"Juno Kim",
"Taiji Suzuki"
] | 2 Feb 2024 (<a href="https://arxiv.org/abs/2402.01258v1">v1</a>), last revised 2 Jun 2024 (this version, v2) | ICML 2024 Oral | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01258 | Large language models based on the Transformer architecture have demonstrated impressive capabilities to learn in context. However, existing theoretical studies on how this phenomenon arises are limited to the dynamics of a single layer of attention trained on linear regression tasks. In this paper, we study the optimization of a Transformer consisting of a fully connected layer followed by a linear attention layer. The MLP acts as a common nonlinear representation or feature map, greatly enhancing the power of in-context learning. We prove in the mean-field and two-timescale limit that the infinite-dimensional loss landscape for the distribution of parameters, while highly nonconvex, becomes quite benign. We also analyze the second-order stability of mean-field dynamics and show that Wasserstein gradient flow almost always avoids saddle points. Furthermore, we establish novel methods for obtaining concrete improvement rates both away from and near critical points. This represents the first saddle point analysis of mean-field dynamics in general and the techniques are of independent interest. | stat-04.zip/2402.01258.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.13945 | Probabilistic Neural Networks (PNNs) for Modeling Aleatoric Uncertainty in Scientific Machine Learning | [
"Farhad Pourkamali-Anaraki",
"Jamal F. Husseini",
"Scott E. Stapleton"
] | 21 Feb 2024 | 9 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.13945 | This paper investigates the use of probabilistic neural networks (PNNs) to model aleatoric uncertainty, which refers to the inherent variability in the input-output relationships of a system, often characterized by unequal variance or heteroscedasticity. Unlike traditional neural networks that produce deterministic outputs, PNNs generate probability distributions for the target variable, allowing the determination of both predicted means and intervals in regression scenarios. Contributions of this paper include the development of a probabilistic distance metric to optimize PNN architecture, and the deployment of PNNs in controlled data sets as well as a practical material science case involving fiber-reinforced composites. The findings confirm that PNNs effectively model aleatoric uncertainty, proving to be more appropriate than the commonly employed Gaussian process regression for this purpose. Specifically, in a real-world scientific machine learning context, PNNs yield remarkably accurate output mean estimates with R-squared scores approaching 0.97, and their predicted intervals exhibit a high correlation coefficient of nearly 0.80, closely matching observed data intervals. Hence, this research contributes to the ongoing exploration of leveraging the sophisticated representational capacity of neural networks to delineate complex input-output relationships in scientific problems. | stat-04.zip/2402.13945.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.13646 | A Large Dimensional Analysis of Multi-task Semi-Supervised Learning | [
"Victor Leger",
"Romain Couillet"
] | 21 Feb 2024 | 13 pages, 9 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.13646 | This article conducts a large dimensional study of a simple yet quite versatile classification model, encompassing at once multi-task and semi-supervised learning, and taking into account uncertain labeling. Using tools from random matrix theory, we characterize the asymptotics of some key functionals, which allows us on the one hand to predict the performances of the algorithm, and on the other hand to reveal some counter-intuitive guidance on how to use it efficiently. The model, powerful enough to provide good performance guarantees, is also straightforward enough to provide strong insights into its behavior. | stat-04.zip/2402.13646.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.14578 | Multivariate Online Linear Regression for Hierarchical Forecasting | [
"Massil Hihat",
"Guillaume Garrigos",
"Adeline Fermanian",
"Simon Bussy"
] | 22 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC) | https://doi.org/10.48550/arXiv.2402.14578 | In this paper, we consider a deterministic online linear regression model where we allow the responses to be multivariate. To address this problem, we introduce MultiVAW, a method that extends the well-known Vovk-Azoury-Warmuth algorithm to the multivariate setting, and show that it also enjoys logarithmic regret in time. We apply our results to the online hierarchical forecasting problem and recover an algorithm from this literature as a special case, allowing us to relax the hypotheses usually made for its analysis. | stat-04.zip/2402.14578.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.03779 | EERO: Early Exit with Reject Option for Efficient Classification with limited budget | [
"Florian Valade",
"Mohamed Hebiri",
"Paul Gay"
] | 6 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.03779 | The increasing complexity of advanced machine learning models requires innovative approaches to manage computational resources effectively. One such method is the Early Exit strategy, which allows for adaptive computation by providing a mechanism to shorten the processing path for simpler data instances. In this paper, we propose EERO, a new methodology to translate the problem of early exiting to a problem of using multiple classifiers with reject option in order to better select the exiting head for each instance. We calibrate the probabilities of exiting at the different heads using aggregation with exponential weights to guarantee a fixed budget .We consider factors such as Bayesian risk, budget constraints, and head-specific budget consumption. Experimental results, conducted using a ResNet-18 model and a ConvNext architecture on Cifar and ImageNet datasets, demonstrate that our method not only effectively manages budget allocation but also enhances accuracy in overthinking scenarios. | stat-04.zip/2402.03779.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.13999 | Asymptotics of Learning with Deep Structured (Random) Features | [
"Dominik Schröder",
"Daniil Dmitriev",
"Hugo Cui",
"Bruno Loureiro"
] | 21 Feb 2024 (<a href="https://arxiv.org/abs/2402.13999v1">v1</a>), last revised 10 Jun 2024 (this version, v2) | ICML camera-ready version | Machine Learning (stat.ML) | Machine Learning (stat.ML); Disordered Systems and Neural Networks (cond-mat.dis-nn); Machine Learning (cs.LG); Statistics Theory (math.ST) | https://doi.org/10.48550/arXiv.2402.13999 | For a large class of feature maps we provide a tight asymptotic characterisation of the test error associated with learning the readout layer, in the high-dimensional limit where the input dimension, hidden layer widths, and number of training samples are proportionally large. This characterization is formulated in terms of the population covariance of the features. Our work is partially motivated by the problem of learning with Gaussian rainbow neural networks, namely deep non-linear fully-connected networks with random but structured weights, whose row-wise covariances are further allowed to depend on the weights of previous layers. For such networks we also derive a closed-form formula for the feature covariance in terms of the weight matrices. We further find that in some cases our results can capture feature maps learned by deep, finite-width neural networks trained under gradient descent. | stat-04.zip/2402.13999.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.04498 | Pathspace Kalman Filters with Dynamic Process Uncertainty for Analyzing Time-course Data | [
"Chaitra Agrahar",
"William Poole",
"Simone Bianco",
"Hana El-Samad"
] | 7 Feb 2024 (<a href="https://arxiv.org/abs/2402.04498v1">v1</a>), last revised 1 Apr 2024 (this version, v2) | 32 pages, 8 figures, Submitted for review | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Quantitative Methods (q-bio.QM) | https://doi.org/10.48550/arXiv.2402.04498 | Kalman Filter (KF) is an optimal linear state prediction algorithm, with applications in fields as diverse as engineering, economics, robotics, and space exploration. Here, we develop an extension of the KF, called a Pathspace Kalman Filter (PKF) which allows us to a) dynamically track the uncertainties associated with the underlying data and prior knowledge, and b) take as input an entire trajectory and an underlying mechanistic model, and using a Bayesian methodology quantify the different sources of uncertainty. An application of this algorithm is to automatically detect temporal windows where the internal mechanistic model deviates from the data in a time-dependent manner. First, we provide theorems characterizing the convergence of the PKF algorithm. Then, we numerically demonstrate that the PKF outperforms conventional KF methods on a synthetic dataset lowering the mean-squared-error by several orders of magnitude. Finally, we apply this method to biological time-course dataset involving over 1.8 million gene expression measurements. | stat-04.zip/2402.04498.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.01139 | Online conformal prediction with decaying step sizes | [
"Anastasios N. Angelopoulos",
"Rina Foygel Barber",
"Stephen Bates"
] | 2 Feb 2024 (<a href="https://arxiv.org/abs/2402.01139v1">v1</a>), last revised 28 May 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.01139 | We introduce a method for online conformal prediction with decaying step sizes. Like previous methods, ours possesses a retrospective guarantee of coverage for arbitrary sequences. However, unlike previous methods, we can simultaneously estimate a population quantile when it exists. Our theory and experiments indicate substantially improved practical properties: in particular, when the distribution is stable, the coverage is close to the desired level for every time point, not just on average over the observed sequence. | stat-04.zip/2402.01139.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.14145 | Multiply Robust Estimation for Local Distribution Shifts with Multiple Domains | [
"Steven Wilkins-Reeves",
"Xu Chen",
"Qi Ma",
"Christine Agarwal",
"Aude Hofleitner"
] | 21 Feb 2024 (<a href="https://arxiv.org/abs/2402.14145v1">v1</a>), last revised 4 Jun 2024 (this version, v2) | 9 pages, 4 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.14145 | Distribution shifts are ubiquitous in real-world machine learning applications, posing a challenge to the generalization of models trained on one data distribution to another. We focus on scenarios where data distributions vary across multiple segments of the entire population and only make local assumptions about the differences between training and test (deployment) distributions within each segment. We propose a two-stage multiply robust estimation method to improve model performance on each individual segment for tabular data analysis. The method involves fitting a linear combination of the based models, learned using clusters of training data from multiple segments, followed by a refinement step for each segment. Our method is designed to be implemented with commonly used off-the-shelf machine learning models. We establish theoretical guarantees on the generalization bound of the method on the test risk. With extensive experiments on synthetic and real datasets, we demonstrate that the proposed method substantially improves over existing alternatives in prediction accuracy and robustness on both regression and classification tasks. We also assess its effectiveness on a user city prediction dataset from Meta. | stat-04.zip/2402.14145.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.15344 | Iteration and Stochastic First-order Oracle Complexities of Stochastic Gradient Descent using Constant and Decaying Learning Rates | [
"Kento Imaizumi",
"Hideaki Iiduka"
] | 23 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.15344 | The performance of stochastic gradient descent (SGD), which is the simplest first-order optimizer for training deep neural networks, depends on not only the learning rate but also the batch size. They both affect the number of iterations and the stochastic first-order oracle (SFO) complexity needed for training. In particular, the previous numerical results indicated that, for SGD using a constant learning rate, the number of iterations needed for training decreases when the batch size increases, and the SFO complexity needed for training is minimized at a critical batch size and that it increases once the batch size exceeds that size. Here, we study the relationship between batch size and the iteration and SFO complexities needed for nonconvex optimization in deep learning with SGD using constant or decaying learning rates and show that SGD using the critical batch size minimizes the SFO complexity. We also provide numerical comparisons of SGD with the existing first-order optimizers and show the usefulness of SGD using a critical batch size. Moreover, we show that measured critical batch sizes are close to the sizes estimated from our theoretical results. | stat-04.zip/2402.15344.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.04022 | A General Theory for Kernel Packets: from state space model to compactly supported basis | [
"Liang Ding",
"Rui Tuo"
] | 6 Feb 2024 (<a href="https://arxiv.org/abs/2402.04022v1">v1</a>), last revised 10 Apr 2024 (this version, v4) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.04022 | It is well known that the state space (SS) model formulation of a Gaussian process (GP) can lower its training and prediction time both to $\CalO(n)$ for $n$ data points. We prove that an $m$-dimensional SS model formulation of GP is equivalent to a concept we introduce as the general right Kernel Packet (KP): a transformation for the GP covariance $K$ such that $\sum_{i=0}^{m}a_iD_t^{(j)}K(t,t_i)=0$ holds for any $t \leq t_1$, 0 $\leq j \leq m-1$, and $m+1$ consecutive points $t_i$, where ${D}_t^{(j)}f(t) $ denotes $j$-th derivative acting on $t$. We extend this idea to the backward SS model formulation, leading to the left KP for next $m$ consecutive points: $\sum_{i=0}^{m}b_i{D}_t^{(j)}K(t,t_{m+i})=0$ for any $t\geq t_{2m}$. By combining both left and right KPs, we can prove that a suitable linear combination of these covariance functions yields $m$ KP functions compactly supported on $(t_0,t_{2m})$. KPs improve GP prediction time to $\mathcal{O}(\log n)$ or $\mathcal{O}(1)$, enable broader applications including GP's derivatives and kernel multiplications, and can be generalized to multi-dimensional additive and product kernels for scattered data. | stat-04.zip/2402.04022.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.03169 | A Random Matrix Approach to Low-Multilinear-Rank Tensor Approximation | [
"Hugo Lebeau",
"Florent Chatelain",
"Romain Couillet"
] | 5 Feb 2024 (<a href="https://arxiv.org/abs/2402.03169v1">v1</a>), last revised 14 Jan 2025 (this version, v3) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Probability (math.PR) | https://doi.org/10.48550/arXiv.2402.03169 | This work presents a comprehensive understanding of the estimation of a planted low-rank signal from a general spiked tensor model near the computational threshold. Relying on standard tools from the theory of large random matrices, we characterize the large-dimensional spectral behavior of the unfoldings of the data tensor and exhibit relevant signal-to-noise ratios governing the detectability of the principal directions of the signal. These results allow to accurately predict the reconstruction performance of truncated multilinear SVD (MLSVD) in the non-trivial regime. This is particularly important since it serves as an initialization of the higher-order orthogonal iteration (HOOI) scheme, whose convergence to the best low-multilinear-rank approximation depends entirely on its initialization. We give a sufficient condition for the convergence of HOOI and show that the number of iterations before convergence tends to $1$ in the large-dimensional limit. | stat-04.zip/2402.03169.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.16326 | A Provably Accurate Randomized Sampling Algorithm for Logistic Regression | [
"Agniva Chowdhury",
"Pradeep Ramuhalli"
] | 26 Feb 2024 (<a href="https://arxiv.org/abs/2402.16326v1">v1</a>), last revised 31 Mar 2024 (this version, v3) | Published in the proceedings of AAAI 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.16326 | In statistics and machine learning, logistic regression is a widely-used supervised learning technique primarily employed for binary classification tasks. When the number of observations greatly exceeds the number of predictor variables, we present a simple, randomized sampling-based algorithm for logistic regression problem that guarantees high-quality approximations to both the estimated probabilities and the overall discrepancy of the model. Our analysis builds upon two simple structural conditions that boil down to randomized matrix multiplication, a fundamental and well-understood primitive of randomized numerical linear algebra. We analyze the properties of estimated probabilities of logistic regression when leverage scores are used to sample observations, and prove that accurate approximations can be achieved with a sample whose size is much smaller than the total number of observations. To further validate our theoretical findings, we conduct comprehensive empirical evaluations. Overall, our work sheds light on the potential of using randomized sampling approaches to efficiently approximate the estimated probabilities in logistic regression, offering a practical and computationally efficient solution for large-scale datasets. | stat-04.zip/2402.16326.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.01493 | Sliced-Wasserstein Estimation with Spherical Harmonics as Control Variates | [
"Rémi Leluc",
"Aymeric Dieuleveut",
"François Portier",
"Johan Segers",
"Aigerim Zhuman"
] | 2 Feb 2024 (<a href="https://arxiv.org/abs/2402.01493v1">v1</a>), last revised 15 May 2024 (this version, v2) | Accepted to ICML 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01493 | The Sliced-Wasserstein (SW) distance between probability measures is defined as the average of the Wasserstein distances resulting for the associated one-dimensional projections. As a consequence, the SW distance can be written as an integral with respect to the uniform measure on the sphere and the Monte Carlo framework can be employed for calculating the SW distance. Spherical harmonics are polynomials on the sphere that form an orthonormal basis of the set of square-integrable functions on the sphere. Putting these two facts together, a new Monte Carlo method, hereby referred to as Spherical Harmonics Control Variates (SHCV), is proposed for approximating the SW distance using spherical harmonics as control variates. The resulting approach is shown to have good theoretical properties, e.g., a no-error property for Gaussian measures under a certain form of linear dependency between the variables. Moreover, an improved rate of convergence, compared to Monte Carlo, is established for general measures. The convergence analysis relies on the Lipschitz property associated to the SW integrand. Several numerical experiments demonstrate the superior performance of SHCV against state-of-the-art methods for SW distance computation. | stat-04.zip/2402.01493.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.16408 | Stable Training of Normalizing Flows for High-dimensional Variational Inference | [
"Daniel Andrade"
] | 26 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.16408 | Variational inference with normalizing flows (NFs) is an increasingly popular alternative to MCMC methods. In particular, NFs based on coupling layers (Real NVPs) are frequently used due to their good empirical performance. In theory, increasing the depth of normalizing flows should lead to more accurate posterior approximations. However, in practice, training deep normalizing flows for approximating high-dimensional posterior distributions is often infeasible due to the high variance of the stochastic gradients. In this work, we show that previous methods for stabilizing the variance of stochastic gradient descent can be insufficient to achieve stable training of Real NVPs. As the source of the problem, we identify that, during training, samples often exhibit unusual high values. As a remedy, we propose a combination of two methods: (1) soft-thresholding of the scale in Real NVPs, and (2) a bijective soft log transformation of the samples. We evaluate these and other previously proposed modification on several challenging target distributions, including a high-dimensional horseshoe logistic regression model. Our experiments show that with our modifications, stable training of Real NVPs for posteriors with several thousand dimensions is possible, allowing for more accurate marginal likelihood estimation via importance sampling. Moreover, we evaluate several common training techniques and architecture choices and provide practical advise for training NFs for high-dimensional variational inference. | stat-04.zip/2402.16408.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.15409 | Lasso with Latents: Efficient Estimation, Covariate Rescaling, and Computational-Statistical Gaps | [
"Jonathan Kelner",
"Frederic Koehler",
"Raghu Meka",
"Dhruv Rohatgi"
] | 23 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Computational Complexity (cs.CC); Data Structures and Algorithms (cs.DS); Machine Learning (cs.LG); Statistics Theory (math.ST) | https://doi.org/10.48550/arXiv.2402.15409 | It is well-known that the statistical performance of Lasso can suffer significantly when the covariates of interest have strong correlations. In particular, the prediction error of Lasso becomes much worse than computationally inefficient alternatives like Best Subset Selection. Due to a large conjectured computational-statistical tradeoff in the problem of sparse linear regression, it may be impossible to close this gap in general. <br>In this work, we propose a natural sparse linear regression setting where strong correlations between covariates arise from unobserved latent variables. In this setting, we analyze the problem caused by strong correlations and design a surprisingly simple fix. While Lasso with standard normalization of covariates fails, there exists a heterogeneous scaling of the covariates with which Lasso will suddenly obtain strong provable guarantees for estimation. Moreover, we design a simple, efficient procedure for computing such a "smart scaling." <br>The sample complexity of the resulting "rescaled Lasso" algorithm incurs (in the worst case) quadratic dependence on the sparsity of the underlying signal. While this dependence is not information-theoretically necessary, we give evidence that it is optimal among the class of polynomial-time algorithms, via the method of low-degree polynomials. This argument reveals a new connection between sparse linear regression and a special version of sparse PCA with a near-critical negative spike. The latter problem can be thought of as a real-valued analogue of learning a sparse parity. Using it, we also establish the first computational-statistical gap for the closely related problem of learning a Gaussian Graphical Model. | stat-04.zip/2402.15409.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.02111 | Accelerating Look-ahead in Bayesian Optimization: Multilevel Monte Carlo is All you Need | [
"Shangda Yang",
"Vitaly Zankin",
"Maximilian Balandat",
"Stefan Scherer",
"Kevin Carlberg",
"Neil Walton",
"Kody J. H. Law"
] | 3 Feb 2024 (<a href="https://arxiv.org/abs/2402.02111v1">v1</a>), last revised 25 Jun 2024 (this version, v2) | Preprint ICML 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC); Probability (math.PR); Computation (stat.CO); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.02111 | We leverage multilevel Monte Carlo (MLMC) to improve the performance of multi-step look-ahead Bayesian optimization (BO) methods that involve nested expectations and maximizations. Often these expectations must be computed by Monte Carlo (MC). The complexity rate of naive MC degrades for nested operations, whereas MLMC is capable of achieving the canonical MC convergence rate for this type of problem, independently of dimension and without any smoothness assumptions. Our theoretical study focuses on the approximation improvements for twoand three-step look-ahead acquisition functions, but, as we discuss, the approach is generalizable in various ways, including beyond the context of BO. Our findings are verified numerically and the benefits of MLMC for BO are illustrated on several benchmark examples. Code is available at <a href="https://github.com/Shangda-Yang/MLMCBO" rel="external noopener nofollow" class="link-external link-https">this https URL</a> . | stat-04.zip/2402.02111.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.16688 | On the connection between Noise-Contrastive Estimation and Contrastive Divergence | [
"Amanda Olmin",
"Jakob Lindqvist",
"Lennart Svensson",
"Fredrik Lindsten"
] | 26 Feb 2024 | Accepted to AISTATS 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.16688 | Noise-contrastive estimation (NCE) is a popular method for estimating unnormalised probabilistic models, such as energy-based models, which are effective for modelling complex data distributions. Unlike classical maximum likelihood (ML) estimation that relies on importance sampling (resulting in ML-IS) or MCMC (resulting in contrastive divergence, CD), NCE uses a proxy criterion to avoid the need for evaluating an often intractable normalisation constant. <br>Despite apparent conceptual differences, we show that two NCE criteria, ranking NCE (RNCE) and conditional NCE (CNCE), can be viewed as ML estimation methods. Specifically, RNCE is equivalent to ML estimation combined with conditional importance sampling, and both RNCE and CNCE are special cases of CD. These findings bridge the gap between the two method classes and allow us to apply techniques from the ML-IS and CD literature to NCE, offering several advantageous extensions. | stat-04.zip/2402.16688.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.16792 | Rate-Optimal Rank Aggregation with Private Pairwise Rankings | [
"Shirong Xu",
"Will Wei Sun",
"Guang Cheng"
] | 26 Feb 2024 (<a href="https://arxiv.org/abs/2402.16792v1">v1</a>), last revised 2 Apr 2025 (this version, v3) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Cryptography and Security (cs.CR); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.16792 | In various real-world scenarios, such as recommender systems and political surveys, pairwise rankings are commonly collected and utilized for rank aggregation to derive an overall ranking of items. However, preference rankings can reveal individuals' personal preferences, highlighting the need to protect them from exposure in downstream analysis. In this paper, we address the challenge of preserving privacy while ensuring the utility of rank aggregation based on pairwise rankings generated from a general comparison model. A common privacy protection strategy in practice is the use of the randomized response mechanism to perturb raw pairwise rankings. However, a critical challenge arises because the privatized rankings no longer adhere to the original model, resulting in significant bias in downstream rank aggregation tasks. To address this, we propose an adaptive debiasing method for rankings from the randomized response mechanism, ensuring consistent estimation of true preferences and enhancing the utility of downstream rank aggregation. Theoretically, we provide insights into the relationship between overall privacy guarantees and estimation errors in private ranking data, and establish minimax rates for estimation errors. This enables the determination of optimal privacy guarantees that balance consistency in rank aggregation with privacy protection. We also investigate convergence rates of expected ranking errors for partial and full ranking recovery, quantifying how privacy protection affects the specification of top-$K$ item sets and complete rankings. Our findings are validated through extensive simulations and a real-world application. | stat-04.zip/2402.16792.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.16661 | Penalized Generative Variable Selection | [
"Tong Wang",
"Jian Huang",
"Shuangge Ma"
] | 26 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.16661 | Deep networks are increasingly applied to a wide variety of data, including data with high-dimensional predictors. In such analysis, variable selection can be needed along with estimation/model building. Many of the existing deep network studies that incorporate variable selection have been limited to methodological and numerical developments. In this study, we consider modeling/estimation using the conditional Wasserstein Generative Adversarial networks. Group Lasso penalization is applied for variable selection, which may improve model estimation/prediction, interpretability, stability, etc. Significantly advancing from the existing literature, the analysis of censored survival data is also considered. We establish the convergence rate for variable selection while considering the approximation error, and obtain a more efficient distribution estimation. Simulations and the analysis of real experimental data demonstrate satisfactory practical utility of the proposed analysis. | stat-04.zip/2402.16661.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.16059 | Gradient-enhanced deep Gaussian processes for multifidelity modelling | [
"Viv Bone",
"Chris van der Heide",
"Kieran Mackle",
"Ingo H.J. Jahn",
"Peter M. Dower",
"Chris Manzie"
] | 25 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.16059 | Multifidelity models integrate data from multiple sources to produce a single approximator for the underlying process. Dense low-fidelity samples are used to reduce interpolation error, while sparse high-fidelity samples are used to compensate for bias or noise in the low-fidelity samples. Deep Gaussian processes (GPs) are attractive for multifidelity modelling as they are non-parametric, robust to overfitting, perform well for small datasets, and, critically, can capture nonlinear and input-dependent relationships between data of different fidelities. Many datasets naturally contain gradient data, especially when they are generated by computational models that are compatible with automatic differentiation or have adjoint solutions. Principally, this work extends deep GPs to incorporate gradient data. We demonstrate this method on an analytical test problem and a realistic partial differential equation problem, where we predict the aerodynamic coefficients of a hypersonic flight vehicle over a range of flight conditions and geometries. In both examples, the gradient-enhanced deep GP outperforms a gradient-enhanced linear GP model and their non-gradient-enhanced counterparts. | stat-04.zip/2402.16059.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.01460 | Deep conditional distribution learning via conditional Föllmer flow | [
"Jinyuan Chang",
"Zhao Ding",
"Yuling Jiao",
"Ruoxuan Li",
"Jerry Zhijian Yang"
] | 2 Feb 2024 (<a href="https://arxiv.org/abs/2402.01460v1">v1</a>), last revised 13 Jun 2024 (this version, v2) | The original title of this paper is "Deep Conditional Generative Learning: Model and Error Analysis" | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.01460 | We introduce an ordinary differential equation (ODE) based deep generative method for learning conditional distributions, named Conditional Föllmer Flow. Starting from a standard Gaussian distribution, the proposed flow could approximate the target conditional distribution very well when the time is close to 1. For effective implementation, we discretize the flow with Euler's method where we estimate the velocity field nonparametrically using a deep neural network. Furthermore, we also establish the convergence result for the Wasserstein-2 distance between the distribution of the learned samples and the target conditional distribution, providing the first comprehensive end-to-end error analysis for conditional distribution learning via ODE flow. Our numerical experiments showcase its effectiveness across a range of scenarios, from standard nonparametric conditional density estimation problems to more intricate challenges involving image data, illustrating its superiority over various existing conditional density estimation methods. | stat-04.zip/2402.01460.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.16712 | l1-norm regularized l1-norm best-fit lines | [
"Xiao Ling",
"Paul Brooks"
] | 26 Feb 2024 (<a href="https://arxiv.org/abs/2402.16712v1">v1</a>), last revised 6 Mar 2024 (this version, v2) | 10 pages | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Optimization and Control (math.OC) | https://doi.org/10.48550/arXiv.2402.16712 | In this work, we propose an optimization framework for estimating a sparse robust one-dimensional subspace. Our objective is to minimize both the representation error and the penalty, in terms of the l1-norm criterion. Given that the problem is NP-hard, we introduce a linear relaxation-based approach. Additionally, we present a novel fitting procedure, utilizing simple ratios and sorting techniques. The proposed algorithm demonstrates a worst-case time complexity of $O(n^2 m \log n)$ and, in certain instances, achieves global optimality for the sparse robust subspace, thereby exhibiting polynomial time efficiency. Compared to extant methodologies, the proposed algorithm finds the subspace with the lowest discordance, offering a smoother trade-off between sparsity and fit. Its architecture affords scalability, evidenced by a 16-fold improvement in computational speeds for matrices of 2000x2000 over CPU version. Furthermore, this method is distinguished by several advantages, including its independence from initialization and deterministic and replicable procedures. Furthermore, this method is distinguished by several advantages, including its independence from initialization and deterministic and replicable procedures. The real-world example demonstrates the effectiveness of algorithm in achieving meaningful sparsity, underscoring its precise and useful application across various domains. | stat-04.zip/2402.16712.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.07160 | PASOA- PArticle baSed Bayesian Optimal Adaptive design | [
"Jacopo Iollo",
"Christophe Heinkelé",
"Pierre Alliez",
"Florence Forbes"
] | 11 Feb 2024 (<a href="https://arxiv.org/abs/2402.07160v1">v1</a>), last revised 4 Jun 2024 (this version, v2) | ICML 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Computation (stat.CO); Methodology (stat.ME) | https://doi.org/10.48550/arXiv.2402.07160 | We propose a new procedure named PASOA, for Bayesian experimental design, that performs sequential design optimization by simultaneously providing accurate estimates of successive posterior distributions for parameter inference. The sequential design process is carried out via a contrastive estimation principle, using stochastic optimization and Sequential Monte Carlo (SMC) samplers to maximise the Expected Information Gain (EIG). As larger information gains are obtained for larger distances between successive posterior distributions, this EIG objective may worsen classical SMC performance. To handle this issue, tempering is proposed to have both a large information gain and an accurate SMC sampling, that we show is crucial for performance. This novel combination of stochastic optimization and tempered SMC allows to jointly handle design optimization and parameter inference. We provide a proof that the obtained optimal design estimators benefit from some consistency property. Numerical experiments confirm the potential of the approach, which outperforms other recent existing procedures. | stat-04.zip/2402.07160.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.14684 | Adaptive time series forecasting with markovian variance switching | [
"Baptiste Abélès",
"Joseph de Vilmarest",
"Olivier Wintemberger"
] | 22 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Probability (math.PR) | https://doi.org/10.48550/arXiv.2402.14684 | Adaptive time series forecasting is essential for prediction under regime changes. Several classical methods assume linear Gaussian state space model (LGSSM) with variances constant in time. However, there are many real-world processes that cannot be captured by such models. We consider a state-space model with Markov switching variances. Such dynamical systems are usually intractable because of their computational complexity increasing exponentially with time; Variational Bayes (VB) techniques have been applied to this problem. In this paper, we propose a new way of estimating variances based on online learning theory; we adapt expert aggregation methods to learn the variances over time. We apply the proposed method to synthetic data and to the problem of electricity load forecasting. We show that this method is robust to misspecification and outperforms traditional expert aggregation. | stat-04.zip/2402.14684.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.14951 | In-Context Learning of a Linear Transformer Block: Benefits of the MLP Component and One-Step GD Initialization | [
"Ruiqi Zhang",
"Jingfeng Wu",
"Peter L. Bartlett"
] | 22 Feb 2024 | 39 pages | Machine Learning (stat.ML) | Machine Learning (stat.ML); Computation and Language (cs.CL); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.14951 | We study the \emph{in-context learning} (ICL) ability of a \emph{Linear Transformer Block} (LTB) that combines a linear attention component and a linear multi-layer perceptron (MLP) component. For ICL of linear regression with a Gaussian prior and a \emph{non-zero mean}, we show that LTB can achieve nearly Bayes optimal ICL risk. In contrast, using only linear attention must incur an irreducible additive approximation error. Furthermore, we establish a correspondence between LTB and one-step gradient descent estimators with learnable initialization ($\mathsf{GD}\text{-}\mathbf{\beta}$), in the sense that every $\mathsf{GD}\text{-}\mathbf{\beta}$ estimator can be implemented by an LTB estimator and every optimal LTB estimator that minimizes the in-class ICL risk is effectively a $\mathsf{GD}\text{-}\mathbf{\beta}$ estimator. Finally, we show that $\mathsf{GD}\text{-}\mathbf{\beta}$ estimators can be efficiently optimized with gradient flow, despite a non-convex training objective. Our results reveal that LTB achieves ICL by implementing $\mathsf{GD}\text{-}\mathbf{\beta}$, and they highlight the role of MLP layers in reducing approximation error. | stat-04.zip/2402.14951.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.09723 | Efficient Prompt Optimization Through the Lens of Best Arm Identification | [
"Chengshuai Shi",
"Kun Yang",
"Zihan Chen",
"Jundong Li",
"Jing Yang",
"Cong Shen"
] | 15 Feb 2024 (<a href="https://arxiv.org/abs/2402.09723v1">v1</a>), last revised 30 May 2024 (this version, v3) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.09723 | The remarkable instruction-following capability of large language models (LLMs) has sparked a growing interest in automatically finding good prompts, i.e., prompt optimization. Most existing works follow the scheme of selecting from a pre-generated pool of candidate prompts. However, these designs mainly focus on the generation strategy, while limited attention has been paid to the selection method. Especially, the cost incurred during the selection (e.g., accessing LLM and evaluating the responses) is rarely explicitly considered. To overcome this limitation, this work provides a principled framework, TRIPLE, to efficiently perform prompt selection under an explicit budget constraint. TRIPLE is built on a novel connection established between prompt optimization and fixed-budget best arm identification (BAI-FB) in multi-armed bandits (MAB); thus, it is capable of leveraging the rich toolbox from BAI-FB systematically and also incorporating unique characteristics of prompt optimization. Extensive experiments on multiple well-adopted tasks using various LLMs demonstrate the remarkable performance improvement of TRIPLE over baselines while satisfying the limited budget constraints. As an extension, variants of TRIPLE are proposed to efficiently select examples for few-shot prompts, also achieving superior empirical performance. | stat-04.zip/2402.09723.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.08201 | Off-Policy Evaluation in Markov Decision Processes under Weak Distributional Overlap | [
"Mohammad Mehrabi",
"Stefan Wager"
] | 13 Feb 2024 (<a href="https://arxiv.org/abs/2402.08201v1">v1</a>), last revised 28 Sep 2025 (this version, v2) | 59 pages, 8 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.08201 | Doubly robust methods hold considerable promise for off-policy evaluation in Markov decision processes (MDPs) under sequential ignorability: They have been shown to converge as $1/\sqrt{T}$ with the horizon $T$, to be statistically efficient in large samples, and to allow for modular implementation where preliminary estimation tasks can be executed using standard reinforcement learning techniques. Existing results, however, make heavy use of a strong distributional overlap assumption whereby the stationary distributions of the target policy and the data-collection policy are within a bounded factor of each other -- and this assumption is typically only credible when the state space of the MDP is bounded. In this paper, we re-visit the task of off-policy evaluation in MDPs under a weaker notion of distributional overlap, and introduce a class of truncated doubly robust (TDR) estimators which we find to perform well in this setting. When the distribution ratio of the target and data-collection policies is square-integrable (but not necessarily bounded), our approach recovers the large-sample behavior previously established under strong distributional overlap. When this ratio is not square-integrable, TDR is still consistent but with a slower-than-$1/\sqrt{T}$-rate; furthermore, this rate of convergence is minimax over a class of MDPs defined only using mixing conditions. We validate our approach numerically and find that, in our experiments, appropriate truncation plays a major role in enabling accurate off-policy evaluation when strong distributional overlap does not hold. | stat-04.zip/2402.08201.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.08711 | Correction to "Wasserstein distance estimates for the distributions of numerical approximations to ergodic stochastic differential equations" | [
"Daniel Paulin",
"Peter A. Whalley"
] | 13 Feb 2024 (<a href="https://arxiv.org/abs/2402.08711v1">v1</a>), last revised 2 Dec 2024 (this version, v3) | 9 pages | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Numerical Analysis (math.NA); Probability (math.PR) | https://doi.org/10.48550/arXiv.2402.08711 | A method for analyzing non-asymptotic guarantees of numerical discretizations of ergodic SDEs in Wasserstein-2 distance is presented by Sanz-Serna and Zygalakis in ``Wasserstein distance estimates for the distributions of numerical approximations to ergodic stochastic differential equations". They analyze the UBU integrator which is strong order two and only requires one gradient evaluation per step, resulting in desirable non-asymptotic guarantees, in particular $\mathcal{O}(d^{1/4}\epsilon^{-1/2})$ steps to reach a distance of $\epsilon > 0$ in Wasserstein-2 distance away from the target distribution. However, there is a mistake in the local error estimates in Sanz-Serna and Zygalakis (2021), in particular, a stronger assumption is needed to achieve these complexity estimates. This note reconciles the theory with the dimension dependence observed in practice in many applications of interest. | stat-04.zip/2402.08711.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.08082 | Score-based generative models break the curse of dimensionality in learning a family of sub-Gaussian probability distributions | [
"Frank Cole",
"Yulong Lu"
] | 12 Feb 2024 (<a href="https://arxiv.org/abs/2402.08082v1">v1</a>), last revised 23 Feb 2024 (this version, v3) | 33 pages, to appear in the proceedings of 12th International Conference on Learning Representations | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.08082 | While score-based generative models (SGMs) have achieved remarkable success in enormous image generation tasks, their mathematical foundations are still limited. In this paper, we analyze the approximation and generalization of SGMs in learning a family of sub-Gaussian probability distributions. We introduce a notion of complexity for probability distributions in terms of their relative density with respect to the standard Gaussian measure. We prove that if the log-relative density can be locally approximated by a neural network whose parameters can be suitably bounded, then the distribution generated by empirical score matching approximates the target distribution in total variation with a dimension-independent rate. We illustrate our theory through examples, which include certain mixtures of Gaussians. An essential ingredient of our proof is to derive a dimension-free deep neural network approximation rate for the true score function associated with the forward process, which is interesting in its own right. | stat-04.zip/2402.08082.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.08412 | Interacting Particle Systems on Networks: joint inference of the network and the interaction kernel | [
"Quanjun Lang",
"Xiong Wang",
"Fei Lu",
"Mauro Maggioni"
] | 13 Feb 2024 | 53 pages, 17 figures | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Dynamical Systems (math.DS); Statistics Theory (math.ST) | https://doi.org/10.48550/arXiv.2402.08412 | Modeling multi-agent systems on networks is a fundamental challenge in a wide variety of disciplines. We jointly infer the weight matrix of the network and the interaction kernel, which determine respectively which agents interact with which others and the rules of such interactions from data consisting of multiple trajectories. The estimator we propose leads naturally to a non-convex optimization problem, and we investigate two approaches for its solution: one is based on the alternating least squares (ALS) algorithm; another is based on a new algorithm named operator regression with alternating least squares (ORALS). Both algorithms are scalable to large ensembles of data trajectories. We establish coercivity conditions guaranteeing identifiability and well-posedness. The ALS algorithm appears statistically efficient and robust even in the small data regime but lacks performance and convergence guarantees. The ORALS estimator is consistent and asymptotically normal under a coercivity condition. We conduct several numerical experiments ranging from Kuramoto particle systems on networks to opinion dynamics in leader-follower models. | stat-04.zip/2402.08412.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
2402.09598 | MCMC-driven learning | [
"Alexandre Bouchard-Côté",
"Trevor Campbell",
"Geoff Pleiss",
"Nikola Surjanovic"
] | 14 Feb 2024 | Machine Learning (stat.ML) | Machine Learning (stat.ML); Machine Learning (cs.LG); Statistics Theory (math.ST); Computation (stat.CO) | https://doi.org/10.48550/arXiv.2402.09598 | This paper is intended to appear as a chapter for the Handbook of Markov Chain Monte Carlo. The goal of this chapter is to unify various problems at the intersection of Markov chain Monte Carlo (MCMC) and machine learning$\unicode{x2014}$which includes black-box variational inference, adaptive MCMC, normalizing flow construction and transport-assisted MCMC, surrogate-likelihood MCMC, coreset construction for MCMC with big data, Markov chain gradient descent, Markovian score climbing, and more$\unicode{x2014}$within one common framework. By doing so, the theory and methods developed for each may be translated and generalized. | stat-04.zip/2402.09598.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 | |
2402.08164 | On Limitations of the Transformer Architecture | [
"Binghui Peng",
"Srini Narayanan",
"Christos Papadimitriou"
] | 13 Feb 2024 (<a href="https://arxiv.org/abs/2402.08164v1">v1</a>), last revised 26 Feb 2024 (this version, v2) | Machine Learning (stat.ML) | Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) | https://doi.org/10.48550/arXiv.2402.08164 | What are the root causes of hallucinations in large language models (LLMs)? We use Communication Complexity to prove that the Transformer layer is incapable of composing functions (e.g., identify a grandparent of a person in a genealogy) if the domains of the functions are large enough; we show through examples that this inability is already empirically present when the domains are quite small. We also point out that several mathematical tasks that are at the core of the so-called compositional tasks thought to be hard for LLMs are unlikely to be solvable by Transformers, for large enough instances and assuming that certain well accepted conjectures in the field of Computational Complexity are true. | stat-04.zip/2402.08164.pdf | /content/arxiv-data/data/stat.ML/2024/02/00000000.parquet | stat.ML | 2024 | 02 | 2024-02-01 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.