Jiajun Fan

I am a Computer Science Ph.D. student at UIUC. My research focuses on autonomous RL post-training for large generative models — making diffusion/flow models and multi-modal reasoning LLMs continuously self-improve with less and less human intervention. Previously, I pushed RL to superhuman performance: breaking 24 Atari world records and outperforming Agent57 with 500× less data.

• Jan 2026 Accept2 papers at ICLR 2026 — CESAR & SP-VLA. See you in Rio 🇧🇷
• Sep 2025 Accept2 papers at NeurIPS 2025 — ADRPO & VarCon. See you in San Diego 🌊
• Jun 2025 AcceptPaper accepted at IEEE TPAMI: PRANCE.
• Feb 2025 AcceptPaper accepted at ICLR 2025: ORW-CFM-W2 (Flow Matching self-evolution).
• Jan 2025 ServiceReviewer: ICLR 2025, NeurIPS 2024, CVPR 2026, AAAI 2025, AISTATS 2025.
• Aug 2024 🎓 Started Ph.D. at UIUC CS (GPA 4.0/4.0).
• Jan 2023 Oral · Top 5%LBC at ICLR 2023, ranked 5/4176 — broke 24 Atari world records.

▼ Show more news

ICLR 2026

CESAR: Process Rewards for Audio LLM Reasoning

First to show test-time inverse scaling in Audio LLMs — and fix it with process rewards via online RL.

🏆 SOTA MMAU · Beats Gemini 2.5 Pro

NeurIPS 2025

ADRPO: Adaptive Regularization for Generative Model RLHF

Sample-level adaptive divergence regularization — 2B model outperforms 12B competitors.

🚀 2B SD3 > FLUX 12B & SANA 4.8B

ICLR 2025

ORW-CFM-W2: Online RLHF for Flow Matching

First collapse-free online RL framework for flow matching models with W2 regularization.

✨ First online RLHF for flow matching

* = first/co-first author  ·  Full list on Google Scholar  /  Publications page

ICLR 20262026

Incentivizing Consistent, Effective and Scalable Reasoning Capability in Audio LLMs via Reasoning Process Rewards Project Page

CESAR resolves test-time inverse scaling in Audio LLMs by rewarding the reasoning process via GRPO, achieving SOTA on MMAU — outperforming Gemini 2.5 Pro and GPT-4o Audio.

J. Fan*, R. Ren, J. Li, R. Pandey, P.G. Shivakumar, A. Gandhe, G. Liu, Y. Gu, I. Bulyko

CESAR: process-reward RL (GRPO) resolving test-time inverse scaling in Audio LLMs — models produce hallucinatory reasoning without proper guidance; CESAR rewrites that.

🏆 SOTA on MMAU Test-mini · Outperforms Gemini 2.5 Pro & GPT-4o Audio

📄 Paper 🔗 Project 📎 arXiv

ICLR 20262026

SP-VLA: A Joint Model Scheduling and Token Pruning Approach for VLA Model Acceleration

SP-VLA introduces action-aware model scheduling and spatio-semantic token pruning for VLA model acceleration, achieving 1.5× lossless speedup on LIBERO and 2.4× speedup on SimplerEnv.

Y. Li, Y. Meng, Z. Sun, K. Ji, C. Tang, J. Fan, X. Ma, S.-T. Xia, Z. Wang, W. Zhu

Action-aware model scheduling + spatio-semantic token pruning for VLA acceleration.

⚡ 1.5× lossless speedup (LIBERO) · 2.4× speedup (SimplerEnv)

NeurIPS 20252025

Adaptive Divergence Regularized Policy Optimization for Fine-tuning Generative Models Project Page

ADRPO introduces sample-level adaptive divergence regularization for RLHF — high-value samples get more freedom, poor samples get stronger constraints. Plug-and-play on any RL method.

J. Fan*, T. Wei, C. Cheng, Y. Chen, G. Liu

ADRPO: sample-level adaptive divergence regularization — high-value samples get more freedom, poor samples get stronger constraint. Plug-and-play on top of any RLHF method.

🚀 2B SD3 surpasses 4.8B & 12B models · Generalizes to LLMs & audio reasoning

📄 Paper 🔗 Project 📎 arXiv

NeurIPS 20252025

Variational Supervised Contrastive Learning

VarCon reformulates supervised contrastive learning as variational inference, achieving SOTA 79.36% Top-1 accuracy on ImageNet-1K with ResNet-50.

Z. Wang, J. Fan, T. Nguyen, H. Ji, G. Liu

VarCon: supervised contrastive learning as variational inference — posterior-weighted ELBO replaces pairwise comparisons.

📊 SOTA 79.36% Top-1 on ImageNet-1K (ResNet-50)

ICLR 20252025

Online Reward-Weighted Fine-Tuning of Flow Matching with Wasserstein Regularization Project Page

ORW-CFM-W2 is the first online RLHF method for flow matching — no human data, no likelihood estimation. Wasserstein regularization maintains generation diversity.

J. Fan*, S. Shen, C. Cheng, Y. Chen, C. Liang, G. Liu

ORW-CFM-W2: first online RLHF for flow matching — no human data, no likelihood, no collapse. W2 regularization keeps generation diverse.

Preprint2025

Fine-tuning Flow Matching Generative Models with Intermediate Feedback Project Page

AC-Flow introduces actor-critic with intermediate feedback for flow matching — reward shaping + dual-stability mechanism + Wasserstein regularization enables robust SD3 fine-tuning without collapse.

J. Fan*, C. Cheng, S. Shen, X. Zhou, G. Liu  ·  Under Review

AC-Flow: actor-critic with intermediate feedback for flow matching — reward shaping + dual-stability + Wasserstein regularization. Robust fine-tuning on SD3 without collapse.

TPAMI 20262026

PRANCE: Joint Token-Optimization and Structural Channel-Pruning for Adaptive ViT Inference

PRANCE jointly optimizes token pruning and structural channel pruning for adaptive ViT inference, achieving significant speedup while maintaining accuracy.

Y. Li, C. Tang, Y. Meng, J. Fan, Z. Chai, X. Ma, Z. Wang, W. Zhu  ·  IEEE TPAMI

ICLR 2023
Oral2023

Learnable Behavior Control: Breaking Atari Human World Records via Sample-Efficient Behavior Selection Project Page

LBC introduces a learnable hybrid behavior mapping and bandit meta-controller for exploration control in deep RL, breaking 24 Atari human world records with 500× less data than prior SOTA.

J. Fan*, Y. Zhuang, Y. Liu, J. Hao, B. Wang, J. Zhu, H. Wang, S.-T. Xia

LBC: learnable hybrid behavior mapping + bandit meta-controller. Unified framework for exploration control in deep RL.

🏅 Ranked 5/4176 · 10,077% mean human score · 24 world records · 500× data efficiency

ICML 20222022

Generalized Data Distribution Iteration Project Page

GDI shows that optimizing the training data distribution is the key lever for superhuman RL efficiency. Provides a unified framework that subsumes diverse RL algorithms as special cases.

J. Fan*, C. Xiao

GDI: optimizing the data distribution is the key to superhuman RL efficiency. Unified framework for diverse RL algorithms.

📈 Agent57 beaten with 500× less data & 2× avg performance

Hover a node to highlight connections. Papers are grouped by research theme.

Making AI Systems That Improve Themselves

Today's AI is frozen after training. I work to change that: AI that never stops getting better, with progressively less human scaffolding.

Key AI/ML venue deadlines I track — for the full list see ccfddl.com.

Happy to discuss research, internships, or collaborations. Best reached by email.
📧 jiajunf3@illinois.edu  ·  🏛 Siebel Center for CS, UIUC  ·  CV (PDF)