About Me

I am currently a Distinguished Associate Researcher at Peking University (since Jul. 2026). From Jul. 2024 to Jun. 2026, I was a Boya Postdoctoral Researcher at Peking University, working with Prof. Yonghong Tian.

I received my Ph.D. in Pattern Recognition and Intelligent Systems from the Shenyang Institute of Automation, Chinese Academy of Sciences (State Key Laboratory of Robotics) in Jun. 2024, under the supervision of Prof. Yang Cong. During Sep. 2022 - Sep. 2023, I was a visiting Ph.D. student supported by CSC at Computer Vision Center (CVC), Universitat Autònoma de Barcelona, working with Prof. Joost van de Weijer.

My research focuses on building adaptive AI systems for scientific discovery, spanning continual learning, scientific agents and multi-agent systems, and embodied intelligence for autonomous experimentation.

Research Interests

Continual Learning, Scientific Agents & Multi-Agent Systems, AI for Science, Embodied AI & Autonomous Experimentation

Academic Services

Area Chair for ICLR; Reviewer for TPAMI, TRO, TMM, TCSVT, TVCJ, TCDS, CVPR, ICCV, ICLR, NeurIPS, ICML, AAAI, ECCV

Research Themes

Three connected research directions

Continual Learning & Adaptive Intelligence

From lifelong visual perception to continual adaptation of foundation and multimodal models.

Scientific Agents & Multi-Agent Systems

Reliable reasoning, tool use, neuro-symbolic planning, and collaboration for scientific tasks.

Embodied AI & Autonomous Experimentation

Grounding perception, planning, safety, and robotic execution in real experimental environments.

Join Us

We welcome students interested in continual learning, scientific agents and multi-agent systems, AI for Science, embodied AI, and autonomous experimentation to join us for research internships.

I particularly enjoy working with students who are passionate, self-motivated, and curious, and who are willing to explore open-ended research problems. I care more about your initiative, independent thinking, and willingness to turn ideas into working systems than about having a perfectly matched background. If you are interested, please email me with your CV, research interests, and representative projects, code, or publications (if available).

Contact Me

News

2026.09 Paper RipplePLM accepted to NeurIPS 2026!
2026.08 Serving as a subproject PI for the National Science and Technology Major Project on Next-Generation AI.
2026.07 LabEvolver released on arXiv.
2026.05 BioProBench accepted to ICML 2026!
2026.04 Oral BioProAgent accepted to ACL 2026 as an Oral!
2025.12 🎉 Awarded the 2025 Shenzhen Outstanding Postdoctoral Fellow (50 recipients/year)!
2025.11 Paper on Open-set Active Learning accepted to IJCV!
2025.11 Oral Paper on LLM Safety accepted to AAAI 2026!
2025.11 Paper on MLLM Reasoning accepted to AAAI 2026!
2024.12 Paper accepted to ICLR 2025!
2024.12 Oral Paper accepted to COLING 2025!
2024.11 Granted China Postdoctoral Science Foundation (76th batch)!
2024.07 Paper accepted to ECCV 2024!
2024.06 🎉 Awarded Postdoctoral Innovative Talents Support Program (A)!

Featured Research

Selected public platforms, preprints, and recent peer-reviewed work

Platform Demo

AI4S LAB: A One-Stop Digital Intelligent Life Science Research Platform

AI4S LAB integrates computing, data, models, and automated experiments, supporting a closed research loop from computational prediction and protocol design to robotic execution and data analysis.

  • 15+ professional datasets and 22 sets of intelligent experimental equipment
  • Natural-language-driven experimental design and execution
  • Integrated dry-lab / wet-lab workflow for life-science research
Yuyang Liu, et al. · 2025–2026
Project
arXiv

LabEvolver: Training-Free Experience Evolution for Safe and Grounded Wet-Lab Agents

A training-free framework that lets robotic wet-lab agents improve from grounded execution experience, coupling state-aware closed-loop operation with reusable skill, strategy, and safety experience.

  • State-grounded inner loop for perception, planning, execution, and safety validation
  • Outer loop distills completed trajectories into reusable experience without model-weight updates
  • Validated on real robotic solution-preparation tasks and continual embodied benchmarks
Jingya Wang, Yuyang Gao, Liuzhenghao Lv, Yuyang Liu · 2026
arXiv Project
Continual Learning for Vision-Language Models Survey · arXiv

Continual Learning for Vision-Language Models: A Survey and Taxonomy

A systematic study of continual learning for vision-language models, focusing on cross-modal representation drift, interference in shared modules, and degradation of zero-shot capabilities, with a unified problem-driven taxonomy.

  • Identifies key challenges in continual vision-language learning
  • Proposes a problem-driven taxonomy
  • Reviews datasets, protocols, metrics, and future directions
Yuyang Liu, et al. · 2025
arXiv GitHub

Selected Publications

Full list on Google Scholar →
* Equal contribution  ·  † Corresponding author

2026
ACL Oral

BioProAgent: Neuro-Symbolic Grounding for Constrained Scientific Planning

Y Liu, J Wang, L Lv, Y Tian

Annual Meeting of the Association for Computational Linguistics (ACL), 2026

ICML

BioProBench: Comprehensive Dataset and Benchmark in Biological Protocol Understanding and Reasoning

Y Liu, L Lv, X Zhang, J Wang, L Yuan, Y Tian

International Conference on Machine Learning (ICML), 2026

arXiv

LabEvolver: Training-Free Experience Evolution for Safe and Grounded Wet-Lab Agents

J Wang, Y Gao, L Lv, Y Liu

arXiv preprint, 2026

NeurIPS

RipplePLM: Structural and Property Decoupling for Protein Mutation Effect Generation

L Lv, Y Liu†, Y Gao, L Yuan, Y Tian†

Advances in Neural Information Processing Systems (NeurIPS), 2026

Nature Comm.

Navigating Chemical-Linguistic Sharing Space with Heterogeneous Molecular Encoding

L Lv, H Li, Y Wang, Z Yan, Z Chen, Z Lin, Y Liu, L Yuan, Y Tian

Nature Communications, 2026

IJCV

Sparse Orthogonal Parameters Tuning for Continual Learning

K Ning, H Ke, Y Liu, J Yao, Y Tian, L Yuan

International Journal of Computer Vision (IJCV), 2026

IJCV

Evidence Conflict Sampling for Open-set Active Learning

K Ning*, H Ke*, J Yao, Y Liu†, Y Tian, L Yuan†

International Journal of Computer Vision (IJCV), 2026

AAAI Oral

AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin

S Yang, Q Zhang, Y Liu†, Y Huang, X Jia, K Ning, J Yao, J Wang, H Dai, Y Song, L Yuan†

AAAI Conference on Artificial Intelligence (AAAI), 2026

AAAI

Look-Back: Implicit Visual Re-focusing in MLLM Reasoning

S Yang, Y Niu, Y Liu†, Y Ye, B Lin, L Yuan†

AAAI Conference on Artificial Intelligence (AAAI), 2026

2025
ICLR

PiCO: Peer Review in LLMs based on the Consistency Optimization

K Ning, S Yang, Y Liu†, J Yao, Z Liu, Y Tian, Y Song, L Yuan†

International Conference on Learning Representations (ICLR), 2025

COLING Oral

Is Parameter Collision Hindering Continual Learning in LLMs?

S Yang, K Ning, Y Liu†, J Yao, Y Tian, Y Song, L Yuan†

International Conference on Computational Linguistics (COLING), 2025

ICME

CASA: Class-Agnostic Shared Attributes in Vision-Language Models for Efficient Incremental Object Detection

M Guo*, Y Liu*†, Z Lin, P Peng, Y Tian†

IEEE International Conference on Multimedia and Expo (ICME), 2025

IEEE TAI

ProLLaMA: A Protein Language Model for Multi-Task Protein Language Processing

L Lv, Z Lin, H Li, Y Liu, J Cui, C Chen, L Yuan†, Y Tian†

IEEE Transactions on Artificial Intelligence, 2025

2024
ECCV

Cs²K: Class-specific and Class-shared Knowledge Guidance for Incremental Semantic Segmentation

W Cong, Y Cong, Y Liu, G Sun

European Conference on Computer Vision (ECCV), 2024

CVPR

Resurrecting Old Classes with New Data for Exemplar-Free Continual Learning

D Goswami, A Soutif-Cormerais, Y Liu†, S Kamath, B Twardowski, J van de Weijer

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024

2023
NeurIPS

FeCAM: Exploiting the Heterogeneity of Class Distributions in Exemplar-Free Continual Learning

D Goswami, Y Liu†, B Twardowski, J van de Weijer

Advances in Neural Information Processing Systems (NeurIPS), 2023

ICCV

Augmented Box Replay: Overcoming Foreground Shift for Incremental Object Detection

Y Liu, Y Cong, D Goswami, X Liu, J van de Weijer

IEEE/CVF International Conference on Computer Vision (ICCV), 2023

IEEE TCSVT

Self-paced Weight Consolidation for Continual Learning

W Cong, Y Cong, G Sun, Y Liu, J Dong

IEEE Transactions on Circuits and Systems for Video Technology, 2023

IEEE TNNLS

Open-ended Online Learning for Autonomous Visual Perception

H Yu, Y Cong, G Sun, D Hou, Y Liu, J Dong

IEEE Transactions on Neural Networks and Learning Systems, 2023

2022
IROS

Class-Incremental Gesture Recognition Learning with Out-of-Distribution Detection

M Li*, Y Cong, Y Liu*, G Sun

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2022

IEEE TCSVT

Lifelong Visual-Tactile Spectral Clustering for Robotic Object Perception

Y Liu, Y Cong, G Sun, Z Ding

IEEE Transactions on Circuits and Systems for Video Technology, 2022

2021 & Earlier
IEEE TPAMI

What and How: Generalized Lifelong Spectral Clustering via Dual Memory

G Sun, Y Cong, J Dong, Y Liu, Z Ding, H Yu

IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021

IEEE TIP

L3DOC: Lifelong 3D Object Classification

Y Liu, Y Cong, G Sun, T Zhang, J Dong, H Liu

IEEE Transactions on Image Processing, 2021

AAAI

Generative Partial Visual-Tactile Fused Object Clustering

T Zhang, Y Cong, G Sun, J Dong, Y Liu, Z Ding

AAAI Conference on Artificial Intelligence, 2021

ECCV

CSCL: Critical Semantic-Consistent Learning for Unsupervised Domain Adaptation

J Dong, Y Cong, G Sun, Y Liu, X Xu

European Conference on Computer Vision (ECCV), 2020

Research Projects

Subproject Principal Investigator

National Science and Technology Major Project on Next-Generation Artificial Intelligence (2026ZD0127201)

AI-Driven Dry-Wet Closed-Loop Data Production Line for Peptide Candidate Drug Discovery

Ministry of Science and Technology of China 2026.08 - Present

Principal Investigator

Young Scientists Fund (Category C), National Natural Science Foundation of China (62606021)

Cognition and Adaptive Evolution Methods for Embodied Scientific Agents in Dry-Wet Closed-Loop Experimentation

National Natural Science Foundation of China 2027.01 -

Principal Investigator

Postdoctoral Innovative Talents Support Program (Grade A)

Embodied Intelligence Learning Methods for Scientific Experiments Driven by Large Models

China Postdoctoral Science Foundation 2024.06 - 2026.06

Principal Investigator

China Postdoctoral Science Foundation (76th Batch)

Efficient Continual Learning Methods for Pre-trained Foundation Models

China Postdoctoral Science Foundation 2024.11 - 2026.06

Honors & Awards

2025

Shenzhen Outstanding Postdoctoral Fellow (50 recipients/year)

Shenzhen Postdoctoral Development Promotion Association

2024

Postdoctoral Innovative Talents Support Program (Grade A)

China Postdoctoral Science Foundation · Top national postdoc talent program

2023

Innovation Award & Third Place

ICCV 2023 Visual Continual Learning Workshop Challenge

2022

National Scholarship & CSC Scholarship

Ministry of Education & China Scholarship Council

2019-20

First-Class Academic Scholarship

University of Chinese Academy of Sciences

2015

National First Prize

National Undergraduate Electronics Design Contest

Visitor Statistics

Flag Counter