Publications

Papers

• arXiv'26 Can Coding Agents Implement Missed Compiler Optimizations? Evaluating LLM Agents on LLVM Peephole Optimizations (under review)
  Hongxu Xu, Chunhao Liao, Xintong Zhou, Chengnian Sun
• ISSTA'26 Automated Dependency Optimization for Artifact-Based Build Systems (to appear) [GitHub] [Preprint]
  Hongxu Xu, Zhenyang Xu, Shane McIntosh, Chengnian Sun
• ASPLOS'26 LPO: Discovering Missed Peephole Optimizations with Large Language Models
  Zhenyang Xu, Hongxu Xu (Co-First), Yongqiang Tian, Xintong Zhou, Chengnian Sun

Talks & Presentations

• Mar 2026 LPO: Discovering Missed Peephole Optimizations with Large Language Models, ACM International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS’26), Pittsburgh, PA, USA.

Books

• Feb 2024 CMake构建实战：项目开发卷 (CMake Build Practice: Project Development Volume)
  Hongxu Xu; 4,000+ copies sold
  Published by 人民邮电出版社 (Posts & Telecom Press, China)
  Produced by 异步图书 (epubit)

Patents

• PCT'26 Text Parsing Method and Apparatus, Device and Medium, PCT/CN2025/108666 (WO 2026/040708 A1), ByteDance. Hongxu Xu, Bing Yang.
  A streaming Markdown/LaTeX parser that avoids incomplete-syntax speech artifacts at chunk boundaries.
• US'26 Method of Text-to-Speech, Medium, and Electronic Device, US 19/307,762 (US 2026/0065893 A1), ByteDance. Yu Chen, Hongxu Xu.
  Context-aware multilingual text normalization and reading for enumerated sequences in dialogue.
• PCT'26 Speech Synthesis Method and Apparatus, Medium, Electronic Device, and Program Product, PCT/CN2025/110581 (WO 2026/056542 A1), ByteDance. Xu Cheng, Hongxu Xu, Bing Yang.
  A pipeline that parses mathematical formulas and turns them into natural spoken language.
• CN'26 Speech Synthesis Method, System, Medium, Electronic Device and Program Product, CN 202411427057.X (CN 121862072 A), ByteDance. Xu Cheng, Hongxu Xu, Yu Chen, Bing Yang.
  Rule-guided text-front-end analysis and phonetic correction for more accurate speech synthesis.
• US'25 Audio Generation Method and System, Device, and Storage Medium, US 18/948,965 (US 2025/0166612 A1), ByteDance. Hongxu Xu, Bing Yang.
  Streaming TTS scheduling and buffer synchronization for continuous, low-latency audio playback.
• CN'25 Text Segmentation Method, Apparatus and Terminal Device, CN 202311426394.2 (CN 119920233 A), ByteDance. Hongxu Xu, Bing Yang.
  Latency-aware streaming text segmentation that reduces time to first audio in voice assistants.
• US'25 Method, Device, and Medium for Speech Interaction, US 18/978,479 (US 2025/0201246 A1), ByteDance. Bing Yang, Hongxu Xu, Chen Li.
  A multimodal spoken-dialogue architecture coupling generative language models with acoustic speech representations.
• US'25 Language Determination Method and Electronic Device, US 18/927,620 (US 2025/0166611 A1), ByteDance. Bing Yang, Hongxu Xu, Lei Huang.
  Sentence-level language identification for choosing pronunciation models in mixed-language speech.