We study the finite-time training and sampling error of conditional diffusion models trained by online stochastic-gradient denoising score matching. Rather than assuming access to an exact empirical or population risk minimizer, we analyze the score network produced after a finite number of stochastic-gradient updates...
Mudit Gaur, Prashant Trivedi, Sasidhar Kunapuli et al.· 0 citations
Flow matching learns a generative velocity field by regression, but under distribution drift a field must be deployed before the current data arrive. We study this predictive problem for a fixed, possibly nonconvex, neural class, with dynamic regret against the best represented field of each round. We propose DyProd-FM...
Reinforcement learning with general utilities optimizes nonlinear functionals of policy-induced occupancy measures. We develop a natural actor-critic method for concave RLGU objectives using an exact variational dual representation, whose existence follows from concavity and differentiability, with explicit constructio...
This work introduces SPatial coHErent risk control for REplay (SPHERE), a general replay-allocation method applicable across a broad range of learning settings and demonstrates that SPHERE improves accuracy and reduces forgetting across noisy-label vision tasks, continual language-model instruction tuning, and code-gen...
Hai-Xiang Sun, Jie-Fu Zhang, Ying-Hao He et al.· arXiv.org· 0 citations
This work introduces Continual Adaptation through the Reuse of Validated Expertise (CARVE), a framework that represents accumulated expertise as a frozen base policy, immutable specialists, and task-specific credentials obtained through local validation that establishes expected-performance guarantees for repeated reus...
Jie-Fu Zhang, Hai-Xiang Sun, Yang Xu et al.· 0 citations
A hierarchical Multilevel Monte Carlo neural critic that performs debiasing simultaneously across trajectory sampling and critic optimization and establishes the first order-optimal convergence guarantees for infinite-horizon average-reward CMDPs with general policy parameterization and neural critics, while eliminatin...
This work proposes a hypergradient-based bilevel RL algorithm using the optimality of the Boltzmann policy for the entropy regularized discounted RL objective function and obtains an iteration complexity of $O(\epsilon^{-1})$ and state-of-the-art sample complexity of $\tilde{O}(\epsilon^{-2})$ under mild regularity con...
Naman Saxena, Mudit Gaur, Vaneet Aggarwal· arXiv.org· 0 citations
This work introduces Selection-Aware Semantic Stress Testing (\SASST{}), which learns a task reweighting from pre-execution features on discovery tasks and evaluates the same paired comparison on separate confirmation tasks.
Yang Xu, Chenang Li, Jiefu Zhang et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
