updated 2 public sources
model compressionoptimizationacceleration

Current frame

Researcher focused on model compression, optimization, and acceleration

Extended note

Public OpenReview and DBLP records identify Xuefeng Xiao as a researcher associated with ByteDance, with OpenReview listing a "Researcher, Intelligent creation, ByteDance" role for 2018-2025 and M.S. study at South China University of Technology from 2015-2018. ByteDance Seed publication pages, DBLP, and OpenReview connect him with work on efficient generative and language model systems, including speculative decoding, diffusion acceleration, quantization, and related optimization methods.