Qiyu Chen

Je pense donc je suis.

This is Qiyu Chen’s personal website, where he will share his life and studies periodically.

About Me

Curriculum Vitae (PDF)

Education

  • B.Sc. in Physics, School of Physics, Peking University, Sep 2023 - Present.
    • Overall average: 90/100.
    • Selected coursework: Introduction to Computation (Python), Linear Algebra, Advanced Mathematics, Methods of Mathematical Physics, Classical Mechanics, Statistical Physics, Foundations of Neural Modeling, Machine Learning.
  • Chengdu No.7 High School, Sep 2021 - Jun 2023.

Research Interests

  • Computational and theoretical neuroscience; dynamical systems approaches to neural and artificial intelligence.
  • Dynamical foundations of learning, representation, and cognition in neural models.
  • Brain-inspired computation and neuromorphic computing.

Publications

  • Qiyu Chen and Guozhang Chen (2026).
    Aligning Inductive Bias for Data-Efficient Generalization in State Space Models.
    Accepted to NeurIPS 2026 (Poster). arXiv:2509.20789.
    [Paper]
  • Wang, G., Liu, C., Jin, L., Sun, Y., Chen, X., Huang, M., Yuan, M., Chen, Q., Yadkori, Y., Chen, G., & Song, S. (2026).
    Brain-inspired dual-timescale processing unlocks deep search in neural networks.
    Under review.

Research Experience

A Preliminary Study on the Generalization of Recurrent Neural Networks Based on Spatiotemporal Inductive Bias

Sep 2026 - Present
Final-year thesis, Peking University with Prof. Guozhang Chen.

  • Aim to extend the theoretical framework for SSM generalization from temporal to spatiotemporal inductive biases.
  • Aim to generalize the framework from linear SSMs to nonlinear recurrent neural networks.

Dynamical Analysis of Hierarchical Reasoning Models

Feb 2026 - Present
Peking University with Prof. Guozhang Chen, Prof. Sen Song (Tsinghua University), and Sapient Inc.

  • Developed a framework based on DMDc and MSD to analyze the internal dynamics of Hierarchical Reasoning Models.
  • Analyzed energy landscapes, representations, dual-timescale dynamics, and generalization in Hierarchical Reasoning Models.
  • Assisted in running additional experiments on Hierarchical Reasoning Models.

Compute-In-Memory Implementation of Energy-Shaped Continuous-Time SSM Blocks for Physical Transient Sequence Processing

Jun 2026 - Aug 2026
The University of Hong Kong with Prof. Can Li.

  • Investigated continuous-time state space models (SSMs) as physically implementable dynamical systems for temporal computation.
  • Developed energy-shaped SSM blocks for neuromorphic compute-in-memory applications and simulated their circuit dynamics in LTSpice.
  • Explored how physical transient dynamics can serve as a computational substrate for brain-inspired temporal sequence processing.

Aligning Inductive Bias for Data-Efficient Generalization in State Space Models

Nov 2024 - Jun 2026
Peking University with Prof. Guozhang Chen. [Paper]

  • Independently carried out the theoretical derivations, algorithm design, and experiments as the first author.
  • Developed a spectral perspective on how state space models impose inductive biases over temporal signals.
  • Established a connection between SSM frequency response, kernel eigenstructure, and finite-sample generalization, and proposed task-dependent initialization for data-efficient sequence learning.

Academic Engagement

  • Summer Research Programme 2026 (participant), Jun 2026 - Aug 2026.
    The University of Hong Kong, Hong Kong SAR, China.
  • The 3rd International Symposium on Cognitive Neuroscience and Intelligent Application (attendee), Oct 2025.
    Hangzhou Dianzi University, Hangzhou, China.
  • Workshop on Physics of Learning (attendee), Nov 2024.
    Institute of Theoretical Physics, Chinese Academy of Sciences (ITP), Beijing, China.

Honors and Awards

  • Award for Scientific Research, Peking University, 2026.
  • Jinde Elite Scholarship, 2026.
  • Outstanding Completion Award & HKU-PS Conditional Selection, HKU Summer Research Programme, 2026.
  • Merit Student, Peking University, 2024.
  • CMC Scholarship, 2024.
  • Excellence Plan, School of Physics, Peking University, 2023.
  • The First Prize of the 39th CPhO (Second Round), 2022.

Skills

  • Languages: Mandarin (native), English (TOEFL 111).
  • Programming: Python, Mathematica, PyTorch, LTSpice.

Extracurricular Activities

Peer Mentor

Sep 2026 - Present
Office of Student Financial Aid, Peking University.

  • Provide peer support and transition guidance to incoming students with financial need.

Student Assistant

Feb 2025 - Jun 2026
Student Career Center, Peking University.

“Challenge Cup” Competition of Students’ Extracurricular Academic and Scientific Achievements

Nov 2025 - Mar 2026
Peking University.

  • Built the website and organized text and data for Trans-Temporal Interpretation and Visualization of the Zhuangzi, supporting sentence-level segmentation, commentary comparison, and concept visualization. [Project]

CUPT Competition

Oct 2023 - Mar 2024
Peking University.

  • Investigated nonlinear physical phenomena, including wet scroll dynamics, the ruler trick, and quantum light dimmer, through mathematical modeling and experiment-driven problem solving.
  • Awarded the Encouragement Prize in the CUPT.

Hobbies

  • Football ⚽ My favorite player is Messi!!!
  • Formula One 🏎️ Leclerc 😍
  • Musical Theater 🎵 Have you ever watched Mozart L’opéra Rock?
  • Art 🎨