作品集

PowerFlow: Unlocking the Dual Nature of LLMs via Principled Distribution Matching ICML 2026 OPT-AIL: Provably & Practically Efficient Adversarial Imitation Learning with General Function Approximation NeurIPS 2024 TD-GFN: Beyond the Proxy — Trajectory-Distilled Guidance for Offline GFlowNet Training ICML 2026 Diffusing to Coordinate: Efficient Online Multi-Agent Diffusion Policies (OMAD) ICML 2026 Offline Diffusion Policy for Multi-User Delay-Constrained Scheduling IEEE TMC 2026 SATA: A Paradigm for LLM Jailbreak via Simple Assistive Task Linkage ACL Findings 2025 SnipSnap: A Joint Compression Format and Dataflow Co-Optimization Framework for Efficient Sparse LLM Accelerator Design ASP-DAC 2026 A Rate-Distortion View of Uncertainty Quantification Preprint, 2024 Comparing Optimization Targets for Contrast-Consistent Search Preprint, 2023 DS-1000: A Natural and Reliable Benchmark for Data Science Code Generation ICML 2023 Transformers, parallel computation, and logarithmic depth ICML 2024 HEMA — Delay-Constrained Scheduling for Dynamic Users MobiHoc 2025