Zichen Liu
- Papers
- 25
Cite
Notes
Only stored in your browser.
Authored papers
25MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research
arXiv 2026
From P(y|x) to P(y): Investigating Reinforcement Learning in Pre-train Space
arXiv 2026
Rethinking the Trust Region in LLM Reinforcement Learning
arXiv 2026
RetroAgent: From Solving to Evolving via Retrospective Dual Intrinsic Feedback
arXiv 2026
Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Library
arXiv 2025
Understanding R1-Zero-Like Training: A Critical Perspective
arXiv 2025
Optimizing Anytime Reasoning via Budget Relative Policy Optimization
arXiv 2025
Token Coordinated Prompt Attention is Needed for Visual Prompting
arXiv 2025
MagicQuillV2: Precise and Interactive Image Editing with Layered Visual Cues
arXiv 2025
SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning
arXiv 2025
Defeating the Training-Inference Mismatch via FP16
arXiv 2025
Language Models Can Learn from Verbal Feedback Without Scalar Rewards
arXiv 2025
Reinforcing General Reasoning without Verifiers
arXiv 2025
Hallucination at a Glance: Controlled Visual Edits and Fine-Grained Multimodal Learning
arXiv 2025
Sailor2: Sailing in South-East Asia with Inclusive Multilingual LLMs
arXiv 2025
Variational Reasoning for Language Models
arXiv 2025
Calligrapher: Freestyle Text Image Customization
arXiv 2025
Efficient Process Reward Model Training via Active Learning
arXiv 2025
Scaling Instruction-Based Video Editing with a High-Quality Synthetic Dataset
arXiv 2025
GEM: A Gym for Agentic LLMs
arXiv 2025
MagicQuill: An Intelligent Interactive Image Editing System
CVPR 2025 1
Sample-Efficient Alignment for LLMs
arXiv 2024
Dynamic Typography: Bringing Text to Life via Video Diffusion Prior
ICCV 2025
Bootstrapping Language Models with DPO Implicit Rewards
arXiv 2024
EnvPool: A Highly Parallel Reinforcement Learning Environment Execution Engine
arXiv 2022
Affiliations
Frequent co-authors
10from 25 papers