Tag
This paper analyzes a single-loop, entropy-regularized Natural Actor-Critic algorithm and proves accelerated convergence rates for the unregularized objective in stochastic and deterministic regimes under linear function approximation.
This paper proposes a cross-layer polar code based federated learning scheme to address communication bottlenecks and channel impairments, providing convergence analysis and resource optimization that demonstrates performance gains over uncoded and LDPC-based benchmarks.
This paper presents ZeroLock, a backpropagation-free algorithm for concurrent memory-efficient LLM training that decouples model updates into independent chunk updates, reducing memory usage by 26.5% and improving throughput by 4.9% compared to BP-based baselines.
The paper proposes a parallel architecture that assembles static gradient methods to achieve adaptivity in stochastic gradient descent, simplifying convergence analysis while retaining parameter adaptivity.
This paper proposes F2CTO, the first distributed first-order constrained trilevel optimization method for robust coreset selection over distributed networks, with a non-asymptotic convergence guarantee of O(ε^(-3/2)).
This paper provides the first finite-time convergence guarantees for the Natural Policy Gradient algorithm in finite-horizon Markov Decision Processes, proving sublinear and linear convergence rates under different step size regimes.
This paper proposes an adaptive gradient descent method using one-sided Hölder regularity to control step sizes based on directional curvature rather than full gradient variation, providing convergence guarantees for nonconvex objectives and demonstrating empirical benefits.
This paper provides the first comprehensive convergence analysis of vanilla SGD with momentum under heavy-tailed noise without gradient clipping or normalization, revealing inferior rates compared to clipped variants and supported by experiments on synthetic functions.
This paper proposes STHTD-MP, a behavior-induced Mirror-Prox temporal-difference method for faster off-policy prediction in reinforcement learning. It replaces the covariance metric with the behavior-policy Bellman matrix and provides convergence analysis and experimental comparisons.
This paper develops a sign-separated finite-time error analysis for constant step-size Q-learning, decomposing the error into negative and positive parts and providing bounds that reveal an asymmetry related to overestimation.
This paper presents a switching-system theory for Q-learning with linear function approximation, using joint spectral radius to analyze convergence stability under deterministic, i.i.d., and Markovian observations.
This paper addresses an open problem in reinforcement learning by providing a counterexample showing that differential temporal difference learning can diverge when using a global clock, despite converging with a local clock, in average-reward settings.