PPL
Xiaoming Huo
Papers (6)
Matching Rates and Optimal Allocation for Federated Probe-Logit Distillation under Heterogeneous Bandwidth BudgetsAnytime-Valid Federated Conformal RAG for LLM SwarmsConformal Selective Acting: Anytime-Valid Risk Control for RLVR-Trained LLMsCatching a Moving Subspace: Low-Rank Bandits Beyond StationarityFederated Language Models Under Bandwidth Budgets: Distillation Rates and Conformal CoverageKernel Selection is Model Selection: A Unified Complexity-Penalized Approach for MMD Two-Sample Tests