Yinsong Yan

My name is Yinsong Yan (闫银松). I am currently a PhD student in the Department of Data Science and Artificial Intelligence (DSAI) at The Hong Kong Polytechnic University, Hong Kong, under the supervision of Prof. Jibin Wu and Prof. Kay Chen Tan.

I received my bachelor’s degree in Applied Mathematics from the University of Electronic Science and Technology of China in 2020, and my master’s degree in Applied Mathematics from Xi’an Jiaotong University in 2023.

My research focuses on multimodal foundation models, omni agents, and neuromorphic computing. I work on multimodal perception and online interaction, while also exploring brain-inspired, energy-efficient algorithms.

By drawing insights from the brain’s energy-efficient information processing, I aim to enhance algorithmic performance for complex tasks while reducing computational overhead. My interdisciplinary approach combines mathematics, neuroscience, and computing.

I am always open to research discussions and collaborations. Please feel free to reach out via email.

Email:

RESEARCH INTEREST

  1. Multimodal Foundation Model
    • Multimodal perception
    • Online interaction
  2. Omni Agent
    • Multimodal reasoning and tool use
    • Real-time, long-horizon agentic interaction
  3. Neuromorphic Computing
    • Event-driven computation and sparse neural activity
    • Energy-efficient algorithms for resource-constrained systems

NEWS

  • [2026-09-22] The Qwen3.8-Omni technical report was released.
  • [2026-09-18] Qwen3.8-Omni-Flash was officially released.
  • [2026-01] Our paper on spatiotemporal representations in spiking neural networks was accepted by ICLR 2026.
  • [2025-09] Our paper on efficient and robust temporal processing with neural oscillations modulated spiking neural networks was published in Nature Communications.

INTERNSHIPS

Alibaba Qwen
2026-02-28–Present — Research Intern, Alibaba Tongyi Lab. Working on online interaction in Qwen-Omni, with a focus on multimodal perception and real-time interaction. Mentored by Dr. Jin Xu and Mr. Yuxuan Wang.

SELECTED PUBLICATIONS

You can find the full list on my Google Scholar.

† denotes the corresponding author(s).

Technical Reports

  1. Figure 1 from Qwen3.8-Omni: model and agent overview Qwen3.8-Omni: Towards Native Omni-Modal Agents

First-Author / Project Leader

  1. Figure 1 from Rhythm-SNN: neural oscillations and temporal processing Efficient and robust temporal processing with neural oscillations modulated spiking neural networks
    • Yinsong Yan, Qu Yang, Yujie Wu, Hanwen Liu, Malu Zhang, Haizhou Li, Kay Chen Tan, Jibin Wu
    • Nature Communications 2025
    • We introduce Rhythm-SNN, a neural modulation framework inspired by brain oscillations for robust and energy-efficient temporal processing. Paper, Code
  2. Figure 1 from the PIT paper: parametric invertible transformation workflow Advancing Spatiotemporal Representations in Spiking Neural Networks via Parametric Invertible Transformation
    • Yinsong Yan, Yujie Wu, Jibin Wu
    • International Conference on Learning Representations (ICLR) 2026
    • We propose a parametric invertible transformation to expand spatiotemporal representations and improve learning dynamics in spiking neural networks. Paper, Code
  3. Main architecture figure from MD3Net: model-driven and data-driven pansharpening MD3Net: Integrating Model-Driven and Data-Driven Approaches for Pansharpening
    • Yinsong Yan, Junmin Liu, Shuang Xu, Yicheng Wang, Xiangyong Cao
    • IEEE Transactions on Geoscience and Remote Sensing 2022
    • We combine model-driven and data-driven approaches for the ill-posed multimodal image-fusion task of pansharpening. Paper, Code

Collaborative Works

  1. Figure 1 from PMSN: comparison of spiking neuron structures PMSN: A Parallel Multi-Compartment Spiking Neuron for Multiscale Temporal Processing
    • X. Chen, Jibin Wu, C. Ma, Yinsong Yan, H. Liu, Y. Wu, Kay Chen Tan
    • IEEE Transactions on Neural Networks and Learning Systems, 2026
  2. KoopSTD overview: state-space observations, time-frequency embedding, and Koopman spectrum comparison Koopstd: Reliable similarity analysis between dynamical systems via approximating Koopman spectrum with timescale decoupling
    • S. Zhang, Z. Ye, Yinsong Yan, Z. Song, Y. Wu, Jibin Wu
    • Forty-second International Conference on Machine Learning (ICML 2025)

EDUCATION

  • 2023–Present — PhD in Data Science and Artificial Intelligence, The Hong Kong Polytechnic University, Hong Kong.
  • 2020–2023 — M.S. in Applied Mathematics, Xi’an Jiaotong University, Xi’an, China.
  • 2016–2020 — B.S. in Applied Mathematics, University of Electronic Science and Technology of China, Chengdu, China.

SERVICES

Conference Reviewer

  • IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
  • International Conference on Learning Representations (ICLR)
  • International Conference on Machine Learning (ICML)
  • Conference on Neural Information Processing Systems (NeurIPS)
  • European Conference on Computer Vision (ECCV)

Journal Reviewer

  • IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)
  • Neural Networks
  • IEEE Transactions on Neural Networks and Learning Systems (TNNLS)
  • IEEE Transactions on Cognitive and Developmental Systems (TCDS)