Zhengbao He

PhD student@Shanghai Jiaotong University·中国·OpenReview

研究方向

sharpness-aware minimization · optimization methods · dataset distillation · dataset condesation · adversarial attack · adversarial training · security

6.5

RAIN-Merging: A Gradient-Free Method to Enhance Instruction Following in Large Reasoning Models with Preserved Thinking Format

ICLR 2026Oral

5.5

T2I-ConBench: Text-to-Image Benchmark for Continual Post-training

ICLR 2026Rejected

4.0

Learn the Ropes, Then Trust the Wins: Self-imitation with Progressive Exploration for Agentic Reinforcement Learning

ICLR 2026Poster

三作

4.0

Remaining-data-free Machine Unlearning by Suppressing Sample Contribution

ICLR 2026Poster

6.8

Pursuing Feature Separation based on Neural Collapse for Out-of-Distribution Detection

ICLR 2025Poster

6.6

Flat-LoRA: Low-Rank Adaptation over a Flat Loss Landscape

ICML 2025Poster

二作

6.3

Simulating Training Dynamics to Reconstruct Training Data from Deep Neural Networks

ICLR 2025Poster

5.8

Flat-LoRA: Low-Rank Adaption over a Flat Loss Landscape

合作者 (20)