Task Profiling and Draft Model Selection for Accelerating Distributed Speculative Decoding

Jan 1, 2026·

Jialin Dong

Yaodan Xu

Tan Chen

Sheng Zhou

Zhisheng Niu

· 0 min read

DOI Cite

Authors

Jialin Dong

Ph.D.

Authors

Yaodan Xu

Ph.D.

Authors

Tan Chen

Ph.D.

Authors

Sheng Zhou

Associate Professor

Authors

Zhisheng Niu

Professor

---

← SparseCoM: Fully Sparse Collaborative Perception via Mixture of BEV Experts Jan 1, 2026

UniMM-V2X: MoE-Enhanced Multi-Level Fusion for End-to-End Cooperative Autonomous Driving Jan 1, 2026 →