Unified framework for solving linear systems with improved convergence rates.
problem Efficiently solving linear systems with randomized batch-sampling methods.
method Developed a unified randomized batch-sampling Kaczmarz framework with concentration inequalities for analysis.
result Derived new expected linear convergence rate bounds that are tighter and more reflective of empirical behavior.
To deal with very large datasets a mini-batch version of the Monte Carlo Markov Chain Stochastic Approximation Expectation-Maximization algorithm for general latent variable models is proposed. For exponential models the algorithm is shown to be convergent under classicalconditions as the number of iterations increases…
We model how Lipschitz continuity changes during neural network training.
problem Understanding how Lipschitz continuity evolves during training.
method We use a system of stochastic differential equations to capture the dynamics of Lipschitz continuity under SGD.
result We identify three factors driving the evolution of Lipschitz continuity: gradient flow projection, gradient noise, and Hessian projection.
GCBS improves contrastive learning performance efficiently.
problem Efficiently mining hard negatives for contrastive learning.
method Global Contrastive Batch Sampling (GCBS) approximates batch assignment problem.
result GCBS improves state-of-the-art performance in sentence embedding and code-search tasks.
Bayesian online learning algorithm for one-pass data, achieving frequentist validity and uncertainty quantification.
problem Theoretical limitations in Bayesian online learning, especially in the one-pass setting.
method Proposed a new Bayesian online learning algorithm with a warm-start phase for the one-pass regime, establishing convergence rates and valid uncertainty quantification.
result The sequentially updated posterior attains optimal convergence rates and valid uncertainty quantification without diverging mini-batch sample sizes.
In this article we review computational aspects of Deep Learning (DL). Deep learning uses network architectures consisting of hierarchical layers of latent variables to construct predictors for high-dimensional input-output models. Training a deep learning architecture is computationally intensive, and efficient linear…
SGD improves generalization by using gradient variability as a proxy for data randomness.
problem Improving generalization in machine learning models trained with stochastic gradient descent.
method Bootstrap perspective on SGD, analyzing gradient variability and algorithmic variability.
result SGD avoids spurious solutions and improves generalization by implicitly regularizing the trace of the gradient covariance matrix.
EnsLoss combines multiple loss functions to prevent overfitting in classification.
problem Preventing overfitting in classification models.
method EnsLoss is an ensemble method that combines loss functions, ensuring calibration and consistency.
result EnsLoss improves classification accuracy compared to fixed loss methods.
Our research is focused on understanding and applying biological memory transfers to new AI systems that can fundamentally improve their performance, throughout their fielded lifetime experience. We leverage current understanding of biological memory transfer to arrive at AI algorithms for memory consolidation and repl…
Training neural networks is traditionally done by providing a sequence of random mini-batches sampled uniformly from the entire training data. In this work, we analyze the effect of curriculum learning, which involves the non-uniform sampling of mini-batches, on the training of deep networks, and specifically CNNs trai…
Optimizes data labeling for causal effect estimation with missing outcomes.
problem Estimating causal effects with missing outcome data and budget constraints.
method Optimizes batch sampling probability to minimize variance of causal inference estimator.
result Achieves lower mean-squared error with fewer labeled data points.
Deep Neural Networks (DNNs) thrive in recent years in which Batch Normalization (BN) plays an indispensable role. However, it has been observed that BN is costly due to the reduction operations. In this paper, we propose alleviating this problem through sampling only a small fraction of data for normalization at each i…
We present K-Means Batch Bayesian Optimization (KMBBO), a novel batch sampling algorithm for Bayesian Optimization (BO). KMBBO uses unsupervised learning to efficiently estimate peaks of the model acquisition function. We show in empirical experiments that our method outperforms the current state-of-the-art batch alloc…
Develops a contraction framework for MCMC mixing rates.
problem Proving mixing-time bounds for MCMC algorithms.
method Global and local contraction coefficients under Eγ-divergence. result Explicit global contraction coefficients for Gaussian smoothing.
This work uses RL to optimize batch experiments in SDOE.
problem Maximizing knowledge with limited resources and constraints.
method Reinforcement Learning for batch-sampling in Bayesian SDOE.
result The proposed algorithm optimizes experiment selection for multiple tasks.
New method for identifying best arm in batched multi-armed bandit problems.
problem Identifying the best arm in multi-armed bandit problems where arms are sampled in batches.
method General linear programming framework for best arm identification in batched multi-armed bandit problems.
result Demonstrated good performance in numerical studies compared to UCB-type or Thompson sampling methods.
Framework improves gradient estimation for faster training convergence.
problem Efficiently estimating noisy gradients in stochastic optimization.
method Dynamic adaptive importance sampling combining multiple distributions.
result Adaptively weighted multiple importance sampling yields superior gradient estimates.
The convergence speed of stochastic gradient descent (SGD) can be improved by actively selecting mini-batches. We explore sampling schemes where similar data points are less likely to be selected in the same mini-batch. In particular, we prove that such repulsive sampling schemes lowers the variance of the gradient est…
BatchGFN uses generative flow networks for efficient batch active learning.
problem Efficiently selecting informative batches for active learning.
method Generative flow networks to sample batches proportional to a batch reward.
result Constructs highly informative batches with a single forward pass per point.
New method improves generalization in deep learning models.
problem Improving generalization in overparameterized deep neural networks.
method Stochastic Gauss-Newton method with Levenberg-Marquardt damping and mini-batch sampling.
result Established finite-time convergence and non-asymptotic generalization bounds.
A new method improves active learning for large batch sizes.
problem Challenges in scaling Bayesian active learning to large batch sizes.
method Derives Partial Batch Label Sampling (ParBaLS) for EPIG algorithm.
result ParBaLS EPIG outperforms top-B selection and BatchBALD. HydaLearn dynamically adjusts task weights for better MTL performance.
problem Constant loss weights in MTL lead to poor results due to drifting relevance and varying mini-batch composition.
method HydaLearn uses mini-batch gradients to dynamically adjust task weights.
result HydaLearn improves performance on synthetic and real-world data.
This work investigates implicit bias in multiclass separable data using a novel geometry-aware optimizer.
problem Understanding implicit bias in overparameterized models on multiclass separable data.
method Introduces NucGD, a geometry-aware optimizer enforcing low-rank structures through nuclear norm constraints.
result NucGD enables scalable training and characterizes the impact of stochastic optimization dynamics.
A new SSL method using whitening of latent-space features.
problem Efficiency and effectiveness of self-supervised representation learning.
method Proposes a new loss function based on whitening of latent-space features, avoiding the need for negatives and asymmetric networks.
result Improves efficiency and effectiveness of SSL by avoiding the need for negatives and asymmetric networks.
The gradient noise of SGD is considered to play a central role in the observed strong generalization abilities of deep learning. While past studies confirm that the magnitude and the covariance structure of gradient noise are critical for regularization, it remains unclear whether or not the class of noise distribution…
Unified framework for high-dimensional online learning with non-divergent error bounds and adaptive gains.
problem Divergence of error bounds in high-dimensional online learning as data batches increase.
method Asynchronous decomposition framework with summary statistics and dynamic regularization.
result Non-divergent error bounds and adaptive gains in sparse online optimization.
Proposes ESCFR to estimate treatment effects from biased data.
problem Treatment selection bias in observational data.
method Stochastic optimal transport with relaxed mass-preserving and proximal factual outcome regularizers.
result Significantly better performance in estimating treatment effects.
HGT models heterogeneous graphs with dedicated node and edge representations.
problem Infeasibility of existing GNNs for heterogeneous graphs.
method Node- and edge-type dependent parameters, relative temporal encoding, heterogeneous mini-batch graph sampling.
result HGT outperforms state-of-the-art GNNs by 9%--21% on various downstream tasks.
VISION-XL improves HD video quality using latent image diffusion models.
problem Improving high-definition video quality and resolution.
method Latent image diffusion models and pseudo-batch consistent sampling.
result State-of-the-art video reconstruction across various inverse problems.
Ensembling DNNs improves minority group performance, leading to fairness.
problem Improving subgroup performances in DNN classifiers.
method Simple homogeneous ensembling of DNNs.
result Fairness naturally emerges from ensembling, improving minority group performance.
DPP-BBO diversifies batched Bayesian optimization using DPPs.
problem Efficiently proposing diverse and informative batches in batched Bayesian optimization.
method Introducing DPP-Batch Bayesian Optimization (DPP-BBO) with DPP-Thompson Sampling (DPP-TS).
result Novel Bayesian simple regret bounds for DPP-TS show improved performance over classical methods.
Structured subsampling improves privacy in deep time series forecasting.
problem Incompatible privacy guarantees with time series forecasting.
method Structured subsampling of sequential data for privacy amplification.
result Structured subsampling enables training with strong privacy guarantees.
Paper reduces uncertainty in predictive models using neural networks and Gaussian processes.
problem High epistemic uncertainty in predictive models.
method Adaptive sampling approach with prediction interval-generation neural networks and Gaussian processes.
result Method consistently converges faster to minimum epistemic uncertainty levels.
Improved TD learning reduces batch sampling error.
problem TD learning can be inaccurate in batch settings due to action weighting.
method Introduced PSEC-TD(0) to correct action weighting using importance sampling.
result PSEC-TD(0) produces more accurate value function estimates.
Deep learning uses ROC cost functions to improve virtual screening accuracy.
problem Challenges in training deep learning models for virtual screening, especially class imbalance and lack of ground truth labels.
method Proposes using ROC cost functions to optimize deep learning models for virtual screening, introduces new training schemes and cost functions.
result Demonstrates improved performance of ROC-based approaches on PubChem datasets.
Recent advances in neural networks have inspired people to design hybrid recommendation algorithms that can incorporate both (1) user-item interaction information and (2) content information including image, audio, and text. Despite their promising results, neural network-based recommendation algorithms pose extensive …
Enhances financial optimization under model uncertainty using subsampling.
problem Model uncertainty in financial decision-making from limited data.
method Superimposes uncertainty measure on model space, uses subsampling for model distribution approximation, adapts SGD for efficiency.
result Uncertainty measures outperform traditional methods and achieve comparable performance to Bayesian methods.
Study mini-batch SGD noise and its limits, proving complexity guarantees.
problem Analyzing the noise in mini-batch SGD and its impact on optimization.
method Examined the conditional covariance and diffusion limits of SGD under different sampling designs.
result Proved mean-square upper bounds and Fisher van Trees lower bounds for SGD, linking them to effective dimension and condition number.
Sketch-GNN reduces GNN training time and memory usage to sublinear scales.
problem Training GNNs on large graphs is computationally expensive and memory-intensive.
method Develops a sketch-based algorithm that trains GNNs on compact sketches of graph adjacency and node embeddings.
result Training time and memory usage grow sublinearly with respect to graph size.
Stochastic Gradient Descent introduces noise in training, affecting model decision boundaries.
problem Understanding the impact of noise in SGD on model decision boundaries.
method Characterized SGD and persistent SGD dynamics in a neural network model, measuring noise magnitude in both under- and over-parametrized regimes.
result Noisier algorithms lead to wider decision boundaries in constraint satisfaction problems.
A new approach for efficient batch multiobjective optimization using Thompson sampling.
problem Inefficient batch multiobjective optimization due to expensive oracles and hard inner optimization.
method Proposes a Thompson sampling approach (qextttPOTS) that chooses Pareto optimal candidates sequentially. result Empirically superior performance compared to classical evolutionary approaches and MOBO.
This paper surveys various methods for dimensionality reduction and nearest neighbor search.
problem Efficiently reducing high-dimensional data to lower dimensions while preserving essential information.
method Linear and nonlinear random projections, including sparse random projections, random Fourier Features, and Random Kitchen Sinks.
result Various methods for dimensionality reduction and nearest neighbor search are explained and compared.
The paper studies how norms of random vectors are preserved by random projections.
problem Understanding how random matrix affects norms of random vectors.
method Proved the distribution of the norm of random vector is preserved by random projection.
result Random matrix preserves the distribution of the norm of random vectors with i.i.d. entries.
Orthogonal random features approximate a Bessel kernel, offering sharper bounds than random Fourier features.
problem Approximating Gaussian kernel efficiently for large datasets.
method Use of Haar orthogonal matrices to construct orthogonal random features and analyze their bias and variance.
result Orthogonal random features approximate a Bessel kernel, not the Gaussian kernel, with sharper bounds.
New methods improve prediction performance and reduce computation time in boosting and random forest models.
problem Improving prediction performance and reducing computation time in boosting and random forest models.
method Random tree depth injection approach for Boosting and Random Forests.
result The new methods can improve prediction performance and reduce computation time by up to 40%.
Proposes a method for modeling random objects in metric spaces using random effects.
problem Modeling random objects in non-Euclidean spaces with random effects.
method Nonlinear Fréchet-based algorithm for M-estimation.
result Consistent estimation of prediction target under random-effects formulation.
Enhances random forest performance with exogenous randomness.
problem Improving random forest performance through exogenous randomness.
method Developed non-asymptotic MSE expansions for individual trees and forests, identified two types of randomness, and conducted simulations.
result Exogenous randomness, particularly feature subsampling, reduces both bias and variance of random forests.
New random forest method provides optimal rates and confidence bands.
problem Improving random forest regression rates and constructing confidence bands.
method Proposed Ehrenfest centered purely random forests achieve optimal rates; used Gaussian approximation for supremum of empirical processes.
result Explicit asymptotic uniform confidence bands constructed for both random forest types.