Songjun Tu
Scholar

Songjun Tu

Google Scholar ID: _5Ir0soAAAAJ
Institute of Automation, Chinese Academy of Sciences; Pengcheng Laboratory
Large Language ModelsReinforecement Learning
Google Scholar↗
Citations & Impact
All-time
Citations
56
 
H-index
5
 
i10-index
2
 
Publications
11
 
Co-authors
2
list available
Contact
No contact links provided.
Publications
8 items
Saliency-Guided Representation with Consistency Policy Learning for Visual Unsupervised Reinforcement Learning
2026
Cited
0
Dynamic Dual-Granularity Skill Bank for Agentic RL
2026
Cited
0
Perception-Consistency Multimodal Large Language Models Reasoning via Caption-Regularized Policy Optimization
2025
Cited
0
SRFT: A Single-Stage Method with Supervised and Reinforcement Fine-Tuning for Reasoning
2025
Cited
0
AlphaDecay:Module-wise Weight Decay for Heavy-Tailed Balancing in LLMs
2025
Cited
0
Learning When to Think: Shaping Adaptive Reasoning in R1-Style Models via Multi-Stage RL
2025
Cited
0
Enhancing LLM Reasoning with Iterative DPO: A Comprehensive Empirical Investigation
2025
Cited
0
Salience-Invariant Consistent Policy Learning for Generalization in Visual Reinforcement Learning
2025
Cited
0
Resume (English only)
Co-authors
2 total
Qichao Zhang
Qichao Zhang
中国科学院自动化研究所
Dongbin Zhao
Dongbin Zhao
Institute of Automation, Chinese Academy of Sciences