PPL
Ziwei Li
Papers (9)
ChainCaps: Composition-Safe Tool-Using Agents via Monotonic Capability AttenuationKISS - Knowledge Infrastructure for Scientific Simulation: A Scaffolding for Agentic Earth ScienceSelf-Supervised Learning for Sparse Matrix ReorderingLearning Fill-in Reduction Ordering via Graph Policy Optimization for Sparse MatricesBridging the Gap between Sparse Matrix Reordering and Factorization: A Deep Learning Framework for Fill-in ReductionThe Granularity Mismatch in Agent Security: Argument-Level Provenance Solves Enforcement and Isolates the LLM Reasoning BottleneckSkill Drift Is Contract Violation: Proactive Maintenance for LLM Agent Skill LibrariesFuture Validity is the Missing Statistic: From Impossibility to $Φ$-Estimation for Grammar-Faithful Speculative DecodingGradient Starvation in Binary-Reward GRPO: Why Group-Mean Centering Fails and Why the Simplest Fix Works