Jiang Liu

Senior Research Scientist · Open Foundation Models · Agentic AI

I’m a Senior Research Scientist at AMD GenAI, where I lead the development of Instella, a series of fully open language models at AMD.

I build models end-to-end—from data curation and large-scale pre-training to post-training, evaluation, systems optimization, and open release.

I received my Ph.D. in Electrical and Computer Engineering from Johns Hopkins University in 2024, advised by Prof. Rama Chellappa, and my B.S.E. from Tsinghua University in 2019.

Email Google Scholar LinkedIn

Research

My research interests have evolved from computer vision to multimodal learning and foundation models. I’ve published at leading AI venues, including CVPR, ICCV, ACL, EMNLP, ICLR, NeurIPS, TPAMI, and Nature Machine Intelligence.

Looking ahead, I’m interested in agentic AI systems that integrate perception, memory, planning, and interaction to solve complex, long-horizon, real-world tasks. I believe impactful AI research should be open, reproducible, and broadly accessible.

We are hiring full-time research scientists and research interns across generative AI. Research collaborations are also welcome—feel free to email me with your CV.

Open Models

Instella and AMD OLMo

Latest · 2026 Project Lead

Instella-MoE

State-of-the-art fully open MoE · 64K context · End-to-end open recipe

Blog Models Code

Project Lead

Instella

Fully open training data, code, checkpoints, and recipes

Blog Models Code

Project Lead

AMD OLMo

AMD’s first fully open language model series

Blog Models

Core Contributor

Instella-Math

Fully open mathematical reasoning with reinforcement learning

Blog Model Code

Core Contributor

Instella-Long

Fully open long-context language modeling

Blog Model Code

Contributor

Instella-VL

AMD’s first fully open vision-language model

Blog Model Code

Contributor

Instella-T2I

Open text-to-image generation with 1D discrete latents

Blog Model Code

News

July 2026: We released Instella-MoE, a state-of-the-art fully open Mixture-of-Experts language model with 16B total parameters and 2.8B active parameters.
April 2026: Three papers accepted to ACL 2026 on LLM and MLLM reasoning. Check out DRIFT, RULES, and Step-Level Advantage Selection (SAS).
February 2026: Two papers accepted to CVPR 2026. Check out VideoSeek and CounterVQA.
January 2026: Three papers accepted to ICLR 2026. Check out ImageDoctor, XModBench, and Latent Visual Reasoning.
January 2026: One paper accepted to ICASSP 2026 as an oral presentation. Check out MOVi.
September 2025: One paper accepted to NeurIPS 2025 as a spotlight. Check out VideoMarathon.
August 2025: Two papers accepted to EMNLP 2025. Check out TTT-Bench and Agent Laboratory.
July 2025: Promoted to Senior Applied Research Scientist at AMD.

Selected Works

Representative and recent work

Introducing Instella-MoE: A State-of-the-Art Fully Open Mixture-of-Experts Language Model Project LeadFully Open
Jiang Liu, Sudhanshu Ranjan, Prakamya Mishra, Yonatan Dukler, Gowtham Ramesh, Zicheng Liu, Jialian Wu, Ximeng Sun, Wen Xie, Chaojun Hou, Vikram Appia, Zhenyu Gu, Emad Barsoum
Tech Report, 2026
Blog / Code / Hugging Face

VideoSeek: Long-Horizon Video Agent with Tool-Guided Seeking
Jingyang Lin, Jialian Wu, Jiang Liu, Ximeng Sun, Ze Wang, Xiaodong Yu, Jiebo Luo, Zicheng Liu, Emad Barsoum
CVPR, 2026
Paper / arXiv / Code

DRIFT: Directional Reasoning Injection for Fine-Tuning MLLMs
Chao Huang, Zeliang Zhang, Jiang Liu, Ximeng Sun, Jialian Wu, Xiaodong Yu, Ze Wang, Chenliang Xu, Emad Barsoum, Zicheng Liu
ACL Findings, 2026
Paper / Project Page / arXiv

XModBench: Benchmarking Cross-Modal Capabilities and Consistency in Omni-Language Models
Xingrui Wang , Jiang Liu , Chao Huang , Xiaodong Yu , Ze Wang , Ximeng Sun , Jialian Wu , Alan Yuille , Emad Barsoum , Zicheng Liu
ICLR, 2026
Paper / Project Page / arXiv

ImageDoctor: Diagnosing Text-to-Image Generation via Grounded Image Reasoning
Yuxiang Guo*, Jiang Liu*, Ze Wang, Hao Chen, Ximeng Sun, Yang Zhao, Jialian Wu, Xiaodong Yu, Zicheng Liu, Emad Barsoum
ICLR, 2026
Paper / Project Page / arXiv

Latent Visual Reasoning
Bangzheng Li, Ximeng Sun, Jiang Liu, Ze Wang, Jialian Wu, Xiaodong Yu, Hao Chen, Emad Barsoum, Muhao Chen, Zicheng Liu
ICLR, 2026
Paper / Project Page / arXiv

Learning from Online Videos at Inference Time for Computer-Use Agents
Yujian Liu, Ze Wang, Hao Chen, Ximeng Sun, Xiaodong Yu, Jialian Wu, Jiang Liu, Emad Barsoum, Zicheng Liu, Shiyu Chang
TMLR, 2026
Paper / arXiv / Code

MOVi: Training-free Text-conditioned Multi-Object Video Generation Oral
Aimon Rahman, Jiang Liu, Ze Wang, Ximeng Sun, Jialian Wu, Xiaodong Yu, Yusheng Su, Vishal M. Patel, Zicheng Liu, Emad Barsoum
ICASSP (Oral), 2026
Paper / arXiv / Code

APRIL: Active Partial Rollouts in Reinforcement Learning to Tame Long-tail Generation
Yuzhen Zhou, Jiajun Li, Yusheng Su, Gowtham Ramesh, Zilin Zhu, Xiang Long, Chenyang Zhao, Jin Pan, Xiaodong Yu, Ze Wang, Kangrui Du, Jialian Wu, Ximeng Sun, Jiang Liu, Qiaolin Yu, Hao Chen, Zicheng Liu, Emad Barsoum
arXiv, 2025
arXiv / Code

TTT-Bench: A Benchmark for Evaluating Reasoning Ability with Simple and Novel Tic-Tac-Toe-style Games
Prakamya Mishra, Jiang Liu, Jialian Wu, Xiaodong Yu, Zicheng Liu, Emad Barsoum
EMNLP Main Conference, 2025
Paper / Project Page / arXiv / Data

Agent Laboratory: Using LLM Agents as Research Assistants
Samuel Schmidgall, Yusheng Su, Ze Wang, Ximeng Sun, Jialian Wu, Xiaodong Yu, Jiang Liu, Michael Moor, Zicheng Liu, and Emad Barsoum
EMNLP Findings, 2025
Paper / Project Page / arXiv / Code

Instella-T2I: Pushing the Limits of 1D Discrete Latent Space Image Generation
Ze Wang, Hao Chen, Benran Hu, Jiang Liu, Ximeng Sun, Jialian Wu, Yusheng Su, Xiaodong Yu, Emad Barsoum, Zicheng Liu
Tech Report, 2025
Blog / arXiv / Code / Hugging Face

Instella-Math: A Fully Open Language Model with Reasoning Capability
Xiaodong Yu, Jiang Liu, Yusheng Su, Gowtham Ramesh, Zicheng Liu et al.
Tech Report, 2025
Blog / Code / Hugging Face

Instella-Long: A Fully Open Language Model with Long-Context Capability
Jialian Wu, Jiang Liu, Sudhanshu Ranjan, Xiaodong Yu, Gowtham Ramesh, Prakamya Mishra, Zicheng Liu, et al.
Tech Report, 2025
Blog / Code / Hugging Face

Instella: Fully open language models with stellar performance Project LeadFully Open
Jiang Liu, Jialian Wu, Xiaodong Yu, Prakamya Mishra, Sudhanshu Ranjan, Zicheng Liu, et al.
Tech Report, 2025
Blog / arXiv / Code / Hugging Face

Unleashing Hour-Scale Video Training for Long Video-Language Understanding Spotlight
Jingyang Lin, Jialian Wu, Ximeng Sun, Ze Wang, Jiang Liu, Yusheng Su, Xiaodong Yu, Hao Chen, Jiebo Luo, Zicheng Liu, Emad Barsoum
NeurIPS, 2025 (Spotlight)
Paper / Project Page / arXiv / Code / Dataset

DIFFNAT: Improving Diffusion Image Quality Using Natural Image Statistics
Aniket Roy, Maiterya Suin, Anshul Shah, Ketul Shah, Jiang Liu, Rama Chellappa
TMLR, 2025
Paper / arXiv

KeyVID: Keyframe-Aware Video Diffusion for Audio-Synchronized Visual Animation
Xingrui Wang, Jiang Liu, Ze Wang, Xiaodong Yu, Jialian Wu, Ximeng Sun, Yusheng Su, Alan Yuille, Zicheng Liu, Emad Barsoum
ICCV Workshop Gen4AVC, 2025
Paper / Project Page / arXiv / Code

Self-Taught Agentic Long Context Understanding
Yufan Zhuang, Xiaodong Yu, Jialian Wu, Ximeng Sun, Ze Wang, Jiang Liu, Yusheng Su, Jingbo Shang, Zicheng Liu, Emad Barsoum
ACL, 2025
Paper / arXiv / Code

SoftVQ-VAE: Efficient 1-Dimensional Continuous Tokenizer
Hao Chen, Ze Wang, Xiang Li, Ximeng Sun, Fangyi Chen, Jiang Liu, Jindong Wang, Bhiksha Raj, Zicheng Liu, Emad Barsoum
CVPR, 2025
Paper / Code

AMD OLMo: Introducing the First AMD 1B Language Models
Jiang Liu, Jialian Wu, Prakamya Mishra, Zicheng Liu et al.
Tech Report, 2025
Blog / Hugging Face

Instruct2Attack: Language-Guided Semantic Adversarial Attacks
Jiang Liu, Chen Wei, Yuxiang Guo, Heng Yu, Alan Yuille, Soheil Feizi, Chun Pong Lau, Rama Chellappa
Preprint, 2023
arXiv

DiffProtect: Generate Adversarial Examples with Diffusion Models for Facial Privacy Protection
Jiang Liu, Chun Pong Lau, Zhongliang Guo, Yuxiang Guo, Zhaoyang Wang, Rama Chellappa
Pattern Recognition, 2025
Paper / arXiv / Code

Interpolated Joint Space Adversarial Training for Robust and Generalizable Defenses
Chun Pong Lau, Jiang Liu, Hossein Souri, Wei-An Lin, Soheil Feizi, Rama Chellappa
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2023
Paper / arXiv

One Model to Synthesize Them All: Multi-contrast Multi-scale Transformer for Missing Data Imputation
Jiang Liu*, Srivathsa Pasumarthi*, Ben Duffy, Enhao Gong, Keshav Datta, Greg Zaharchuk (*equal contribution)
IEEE Transactions on Medical Imaging (TMI), 2023
Paper / arXiv

PolyFormer: Referring Image Segmentation as Sequential Polygon Generation
Jiang Liu*, Hui Ding*, Zhaowei Cai, Yuting Zhang, Ravi Kumar Satzoda, Vijay Mahadevan, R. Manmatha (*equal contribution)
CVPR, 2023
Paper / Project Page / arXiv / Code

Segment and Complete: Defending Object Detectors Against Adversarial Patch Attacks With Robust Patch Detection
Jiang Liu, Alexander Levine, Chun Pong Lau, Rama Chellappa, Soheil Feizi
CVPR, 2022
Paper / arXiv / Code / Apricot-Mask Dataset / Supp

Mutual Adversarial Training: Learning together is better than going alone
Jiang Liu, Chun Pong Lau, Hossein Souri, Soheil Feizi, Rama Chellappa
IEEE Transactions on Information Forensics and Security (TIFS), 2022
Paper / arXiv

Professional Services

Conference Reviewer: CVPR, ICCV, ECCV, WACV, ACL, ICLR, ICML, ACM MM, FG, ACCV, AAAI, MICCAI
Journal Reviewer: TPAMI, TIP, TIFS, TAI, TOPS, TMI, TCSVT, IJCV, MEDIA