Understanding intelligence across human brains, bodies, and machines.
JINGYUAN SUN / MANCHESTER, UKI am an Assistant Professor (Lecturer) in Computer Science at The University of Manchester. I am a member of ELLIS and a Principal Investigator within the ELLIS network. I also serve as a founding editorial board member of the Journal of Terminology and Knowledge Engineering (JTKE), founded the AI4BCI Conference, and coordinate the HIT Webinar series. I am also a Fellow of Advance HE (FHEA).
Before joining Manchester, I was a postdoctoral researcher in the Department of Computer Science at KU Leuven, where I worked with Prof. Marie-Francine Moens from 2022 to 2024. My research there was supported by the ERC Advanced Grant project CALCULUS under Horizon 2020. Prior to that, I worked as a Senior R&D Engineer at Baidu in China.
I received my PhD and master's degree from the Institute of Automation, Chinese Academy of Sciences, under the supervision of Prof. Chengqing Zong. My doctoral work brought together computational linguistics and neurolinguistics, and established my long-standing interest in the relationship between biological and artificial intelligence.
Fully funded PhD opportunities are available. Get in touch ↗
My research is centred on the broader question of how human intelligence can be understood computationally, and how this understanding can contribute to the development of more capable artificial intelligence. My work currently develops along three main directions.
We investigate the computational principles underlying human intelligence by connecting neural activity with representations in AI. We develop models that decode human perception from brain signals and use AI as an experimental framework for understanding cognition, dysfunction, and recovery.
We build machine intelligence capable of memory, adaptation, planning, and robust reasoning. By studying intelligence under limited resources and uncertainty, we aim to develop AI systems that are efficient, resilient, and trustworthy.
We seek to uncover the organizing principles of living systems across molecular, cellular, and organismal scales. By combining foundation models with mechanistic and causal computation, we aim to move AI beyond pattern recognition toward scientific explanation, prediction, and intervention.
One Scholar entry titled “Association for Computational Linguistics” is a publisher-name metadata record rather than a paper, so it is omitted.
I am grateful to work alongside this talented group. Their curiosity, dedication and collaborative spirit make our research possible.
Co-supervisors: Hongpeng Zhou and Goran Nenadic. Develops foundation models for heterogeneous multi-omics data and disease-related tissue organization; enjoys tennis, travel and varied cuisines.
Co-supervisors: Chenghua Lin and Mustafa Mustafa. Studies LLMs, BCI and long-sequence representation learning for robust neural decoding from non-invasive signals.
Main supervisor: Riza Batista-Navarro; co-supervisor: Chenghua Lin. Studies efficient multimodal LLMs, token pruning, compression and optimization, and the broader impact of efficient AI.
Main supervisor: Goran Nenadic. Uses LLMs, computational lesions and spiking neural networks to study neural encoding and multilingual language comprehension.
Jimin Huang · Yuyan Wang · Yixiang Zheng
Supervised / co-promoted: Ali Togar Dincer — Decoding Linguistic Stimulus from Human Brain Activities with Language Models; Jonathan Swinnen — Decoding Visual Stimulus from Human Brain Activities with Diffusion Models. Both were listed as ongoing on the previous website.
Thesis assessor: David Debot — Approximating Volume Computations with Neural Networks; Arne Vermolen — Investigating Neural-Symbolic Methods and Tasks: A Comparative Evaluation; Anissa Faik — Bringing a New Perspective to the Classroom: Detecting and Explaining Student Outliers; Lara Roosens — Turning Competency Data into Actionable Insights for Teachers; Toon Sauvillers — Learning Analytics Dashboard: Visualising Students’ Performance Taking into Account Their Socioeconomic Background.
Principles, practices and beyond. Listed as tutorial TH09 in the official AAAI-26 programme.
IJCAI 2024Methods, applications and future directions. Sessions cover an introduction, language and speech, and vision.
ACL 2024Principles, practices and future directions. Four sessions: introduction (Shaonan Wang), language and speech I (Jingyuan Sun), language and speech II (Jixing Li), vision (Zijiao Chen).
Slides, recording and session resources: ACL 2024 ↗ · IJCAI 2024 ↗. For the ACL recording, the previous website also gives jingyuansun01@gmail.com ↗ as a contact.
Chair of the AAAI 2024 workshop Artificial Intelligence for Brain Encoding and Decoding; organizer of the ERC CALCULUS symposium; and organizing chair of AI for BCI 2026 ↗.
Session chair at IJCAI 2023; programme committee member for AAAI, ECAI, IJCAI, EMNLP and COLING; reviewer for IEEE TPAMI.
News dates are drawn from the previous Google Sites homepage. Some entries were combined when they described the same month.