333 papers
The paper proposes ISD-linUCB, an algorithm for stochastic non-stationary linear bandits that leverages the decomposition of reward models into stationary and non-stationary components to learn invariances from historical data, thereby reducing problem dimensionality and significantly improving dynamic regret in fast-changing environments.
This paper proposes a hierarchical Bayesian framework that integrates adaptive surrogate models, such as Fourier Neural Operators and parametric PINNs, with ensemble MALA sampling to simultaneously infer individual system parameters and learn shared unknown dynamics via ML-based closure models for ODEs and PDEs.
This paper addresses the fixed-budget best-arm identification problem in grouped bandits with feasibility constraints by establishing a theoretical lower bound on error probability and proposing the Feasibility Constrained Successive Rejects (FCSR) algorithm, which achieves optimal performance guarantees while empirically outperforming natural baselines.
This paper proposes a sparse subgradient algorithm that explicitly leverages the inherent sparsity of Max-Plus neural networks' subgradients to enable efficient and theoretically guaranteed training, overcoming the computational inefficiencies of standard backpropagation.
This paper demonstrates that applying L2 weight regularization to sparse autoencoders significantly enhances the stability and cross-seed consistency of learned features, leading to improved steering success rates and a stronger alignment between automated interpretability scores and functional controllability.
This paper establishes a principled theoretical framework for density aggregation by demonstrating that normalized generalized means with order are the only rules guaranteeing systematic improvements in log-likelihood over individual distributions, thereby providing a unified justification for the widespread use of linear and geometric pooling in Deep Ensembles.
This paper introduces Latent Space Distribution Matching (LSDM), a novel semi-supervised framework that leverages both paired and unpaired data to learn a latent space and perform joint distribution matching, thereby enhancing generation quality and providing theoretical insights into Latent Diffusion Models.
PTOPOFL is a privacy-preserving personalized federated learning framework that replaces gradient sharing with compact persistent homology descriptors to simultaneously mitigate data-reconstruction risks and improve aggregation performance on non-IID data through topology-guided clustering and weighted aggregation.
This paper reframes Quantum Neural Network design from state reachability to learnability by introducing geometric design principles and the almost Complete Local Selectivity (aCLS) criterion, demonstrating that architectures requiring joint dependence on data and trainable weights enable adaptive feature learning and outperform traditional schemes with greater efficiency.
This paper presents high-quality image synthesis using denoising diffusion probabilistic models, achieving state-of-the-art results on CIFAR10 and LSUN datasets through a novel training objective connecting diffusion models to denoising score matching with Langevin dynamics.
This paper establishes that language model performance follows predictable power-law scaling relationships with model size, dataset size, and compute, revealing that optimal training efficiency is achieved by prioritizing very large models trained on modest data amounts rather than training smaller models to convergence.
This paper proposes a new generative modeling framework based on a minimax two-player game between a generative model that captures data distribution and a discriminative model that distinguishes real from generated samples, which can be trained efficiently using backpropagation without Markov chains.
This paper introduces the Auto-Encoding Variational Bayes (AEVB) framework, which enables efficient stochastic variational inference and learning in directed probabilistic models with continuous latent variables by utilizing a reparameterized lower bound estimator and an approximate inference model to handle intractable posteriors and scale to large datasets.