Selected Publications

Preprints

1. Many Right Answers, One Training Example: Removing Responder Noise from GRPO for LLM Recommenders

   Preprint 2026

   Xin Li, Hao Jiang, Jinghao Guo, Weisi Lin, Chau Yuen

   02 Train

   ProjectPaperCodeDataModels

2. Share the Symbol, Not the Vector: Training-Free Fact Communication Across Language Models

   Preprint 2026

   Xin Li, Mengbing Liu, Hao Jiang, Xin Gao, Annan Wang, Xingwei Qu, Yichi Zhang, Chau Yuen

   03 Orchestrate

   ProjectPaperCodeDataModels

3. Beyond Teacher Assignment: Domain-Normalized Multi-Teacher On-Policy Distillation

   Preprint 2026

   Xin Li, Hao Jiang, Xin Gao, Annan Wang, Yuchen Xie, Jinghao Guo, Xingwei Qu, Yichi Zhang, Chau Yuen

   02 Train

   ProjectPaperCodeDataModels

4. ErrOPD: When Does On-Policy Distillation Transfer Teacher Errors?

   Preprint 2026

   Xin Li, Hao Jiang, Annan Wang, Yichi Zhang, Chau Yuen

   02 Train

   ProjectPaperCodeDataModels

2026

1. WirelessMathBench-XL: An Auditable Benchmark for Wireless Mathematical Reasoning

   NeurIPS 2026

   Xin Li, Mengbing Liu, Yiyang Zhu, Wenhe Zhang, Li Wei, Jiancheng An, Chau Yuen

   01 EvaluateEarlier version: WirelessMathLM

   ProjectPaperCodeDataModels

2. Measuring Collapse and Correction in Homogeneous-Panel LLM Debate

   NeurIPS 2026

   Xin Li^*, Mengbing Liu^*, Chau Yuen

   03 Orchestrate

   ProjectPaperCodeData

3. Local Success Does Not Compose: Benchmarking Large Language Models for Compositional Formal Verification

   ICLR 2026

   Xu Xu^*, Xin Li^*, Xingwei Qu, Jie Fu, Binhang Yuan

   01 Evaluate

   ProjectPaper

4. Target-Local Verifier Choice in Best-of-K Reasoning Selection

   EMNLP 2026 Findings

   Xin Li, Hao Jiang, Weisi Lin

   03 Orchestrate

   ProjectCode

5. Reconstructing wireless signals for low altitude networks using small language models

   npj Wireless Technology 2026

   Xin Li, Ran Liu, Chau Yuen

   02 Train

   Paper

6. Re:Form: Reducing Human Annotations in Scalable Formal Software Verification with RL in LLMs: A Preliminary Study on Dafny

   TMLR 2026

   Chuanhao Yan^*, Fengdi Che^*, Xuhan Huang^*, Xu Xu^*, Xin Li^*, Yizhi Li^*, Xingwei Qu^*, …, Jie Fu

   02 Train

   PaperCode

7. Stacked Intelligent Metasurface-Diffractive Deep Neural Networks for Onboard Terrain Classification from SAR Level-0 Raw Data

   IEEE TSP 2026

   Mengbing Liu, Xin Li, Jiancheng An, Chau Yuen

   Wireless systems

   Project

8. RobustMAD: Evaluating Real-World Robustness of Multimodal Small Language Models for Deployable Anomaly Detection Assistants

   TMLR 2026

   Anushiya Arunan, Xin Li, Yan Qin, U-Xuan Tan, Nhu Khue Vuong, Xiaoli Li, Chau Yuen

   01 Evaluate

   ProjectPaperCode

9. GraphReduce: Coverage-Preserving LLM Aggregation for E-commerce Review Insights

   EMNLP 2026 Industry

   Hao Jiang, Xin Li, Yichi Zhang, Weisi Lin

   02 Train

   Project

10. LiveCANNBench: Benchmark SWE AI Coding for Ascend CANN

    ACL 2026 Findings

    Sijie Wang, Kai Zhao, Wee Peng Tay, Shuo Zhang, Chengwen Liu, Quanjiang Guo, Ren Junhao, Xin Li, Heng Lian, Jingdi Lei, Rui She, Huacan Wang, Ronghao Chen

    01 Evaluate

    Paper

2025

1. WirelessMathBench: A Mathematical Modeling Benchmark for LLMs in Wireless Communications

   ACL 2025 Findings

   Xin Li, Mengbing Liu, Li Wei, Jiancheng An, Mérouane Debbah, Chau Yuen

   01 Evaluate

   ProjectPaperCodeData

2. TransPathNet: A Novel Two-Stage Framework for Indoor Radio Map Prediction

   ICASSP 2025

   Xin Li, Ran Liu, Saihua Xu, Sirajudeen Gulam Razul, Chau Yuen

   Wireless systems

   ProjectPaperCode

3. LACP: LLM Agent Communication Protocol Requires Urgent Standardization

   NeurIPS 2025 Workshop (AI4NextG)

   Xin Li, Mengbing Liu, Chau Yuen

   03 Orchestrate

   ProjectPaper

2020

1. Co-Planar Parametrization for Stereo-SLAM and Visual-Inertial Odometry

   IEEE RA-L 2020

   Xin Li^*, Yanyan Li^*, Evin Pinar Örnek, Jinlong Lin, Federico Tombari

   Robot perception

   PaperCode

2. Leveraging Planar Regularities for Point-Line Visual-Inertial Odometry

   IROS 2020

   Xin Li^*, Yijia He^*, Jinlong Lin, Xiao Liu

   Robot perception

   PaperCode

Also: Onboard Terrain Classification via SIM-DNN, ML4RS workshop @ ICLR 2025 — extended into the IEEE TSP paper above.