Scholar
Pengxiang LI
Google Scholar ID: rUp_4RgAAAAJ
The Hong Kong Polytechnic University
Citations & Impact
Citations
242
H-index
9
i10-index
7
Publications
17
Co-authors
20
list available
Publications
Browse publications on Google Scholar (top-right) ↗
Resume (English only)
Academic Achievements
- 1. Three papers accepted at NeurIPS 2025: 'Adaptive Classifier-Free Guidance via Dynamic Low-Confidence Masking', 'The Curse of Depth in Large Language Models', and 'GPAS: Accelerating Convergence of LLM Pretraining via Gradient-Preserving Activation Scaling'
- 2. One paper accepted to ICLR 2025: 'Mix-LN: Unleashing the Power of Deeper Layers by Combining Pre-LN and Post-LN'
- 3. One paper accepted to ACL 2025 Findings: 'Outlier-weighed Layerwise Sampling for LLM Fine-tuning'
- 4. Two papers accepted to WACV 2025: 'Automated Evaluation of Large Vision-Language Models on Self-driving Corner Cases' and 'TrackDiffusion'
- 5. Two papers published on arXiv 2025: 'InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection' and 'InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners'
Research Experience
- Currently a Ph.D. student at The Hong Kong Polytechnic University, focusing on AI research.
Education
- 1. Ph.D. Student at The Hong Kong Polytechnic University, supervised by Prof. Hongxia Yang
- 2. MSc from Dalian University of Technology, supervised by Prof. Huchuan Lu
Background
- Research interests include Large Language Models, Multimodal GUI Agents, and Diffusion Models for Video Generation.
