AIConfPaper
← Search

Xiangzhong Luo

2 accepted papers

2026

Flatter Tokens are More Valuable for Speculative Draft Model Training

Jiaming Fan, CAO DAMING, Xiangzhong Luo, Jiale Fu, Chonghan Liu, Xu Yang
ICLR 2026poster

Speculative Decoding (SD) is a key technique for accelerating Large Language Model (LLM) inference, but it typically requires training a draft model on a large dataset. We approach this problem from a data-centric perspective, finding that not all training samples contribute equally to the SD accept…

Cited by 0Source ↗code ↗Scholar ↗
2026

d$^2$Cache: Accelerating Diffusion-Based LLMs via Dual Adaptive Caching

Yuchu Jiang, Yue Cai, Xiangzhong Luo, Jiale Fu, Jiarui Wang, Chonghan Liu +1
ICLR 2026poster

Diffusion-based large language models (dLLMs), despite their promising performance, still suffer from inferior inference efficiency. This is because dLLMs rely on bidirectional attention and cannot directly benefit from the standard key-value (KV) cache as autoregressive models (ARMs) do. To tackle…

Cited by 0Source ↗code ↗Scholar ↗