GitHubQjbtiger - Overview
Qjbtiger - Overview
A Ph.D. student majoring in artificial intelligence at Hong Kong University of Science and Technology (Guangzhou). - Qjbtiger
GitHubI am currently a first-year Ph.D. student in the Artificial Intelligence Thrust, Information Hub, at The Hong Kong University of Science and Technology (Guangzhou), under the supervision of Prof. Yao Shu. Prior to joining HKUST(GZ), I received my master's and bachelor's degree in Physics from Sun Yat-sen University, where I was advised by Prof. Haiping Huang. My research focuses on developing theoretically grounded and practically efficient algorithms for modern AI systems, with an emphasis on zeroth-order optimization, efficient post-training of large language models, post-training quantization, and on-device AI. My recent work has been published in venues including ICML, Physical Review Research, Physical Review E.
Research Interests
- Variance-reduced zeroth-order optimization for black-box objectives and on-device fine-tuning of large language models.
- Post-training quantization for deploying large models on edge devices.
- Efficient post-training methods for LLMs in both online and offline settings.
- On-device AI algorithms that balance performance, efficiency, and privacy.
Manuscripts Under Review
- APEX: Accuracy Projection from Verifier-Labeled Experience for Offline RLVR Junbin Qiu, Yao Shu Submitted to NeurIPS 2026
- Predicting Quantization Price for Selecting PTQ Configurations Before Deployment Junbin Qiu, Jian Mu, Weitong Zhang, Yao Shu Submitted to NeurIPS 2026
- The Cost of Mismatch: Noise Amplification in Zeroth-Order Reinforcement Learning Lianming Chen, Junbin Qiu, Chenxing Wei, Yao Shu, Kai He Submitted to NeurIPS 2026
Selected Publications
- Revisiting Zeroth-Order Hessian Approximation: A Single-Step Policy Optimization Lens Junbin Qiu, Zhaowei Hong, Renzhe Xu, Yao Shu ICML 2026
- Zeroth-Order Optimization is Secretly Single-Step Policy Optimization Junbin Qiu, Xiangda Yan, Yongjie Yang, Yao Shu ICML 2025 Workshop
- An optimization-based equilibrium measure describes non-equilibrium steady state dynamics: application to edge of chaos Junbin Qiu, Haiping Huang Communications in Theoretical Physics, 2024
- Meta predictive learning model of languages in neural circuits Chan Li, Junbin Qiu, Haiping Huang Physical Review E, 2024
- Equivalence between belief propagation instability and transition to replica symmetry breaking in perceptron learning systems Yang Zhao, Junbin Qiu, Mingshan Xie, Haiping Huang Physical Review Research, 2022
Education
- Ph.D. student in Artificial Intelligence (2025.9 - now) The Hong Kong University of Science and Technology (Guangzhou)
- Master's degree in Statistical Physics (2021.9 - 2024.6) Sun Yat-sen University
- Bachelor's degree in Physics (2017.9 - 2021.6) Sun Yat-sen University
Activities
- Talk: A language model based on free energy minimization Swarma pattern study group, 2024.6.15
Contact
I am open to research discussions and collaboration opportunities around theoretical machine learning, zeroth-order optimization, efficient LLM post-training, and model deployment.
Email: jqiu236 [at] connect.hkust-gz.edu.cn