Resume
Jian Song
Download PDFRESEARCH INTERESTS
- Robotics physical interaction modeling and sim-to-real dynamics.
- Physically reliable world models and vision-language-action systems.
- Event-driven 3D perception with event cameras and point clouds.
- Spiking neural networks for efficient perception and control.
EDUCATION
University of Colorado Denver 2025 – Present
PhD Student, Computer Science and Information Systems Denver, CO, USA
◦ PhD Advisor: Professor Zhengxiong Li
The University of Edinburgh Sept. 2021 – Oct. 2022
Master of Science, Drug Discovery and Translational Biology Edinburgh, UK
Central China Normal University Sept. 2017 – June 2021
Bachelor of Science, Biotechnology Wuhan, China
PUBLICATIONS
- Jian Song, Xiangfei Yang, Shangke Lyu, Donglin Wang (2025). Activation-wise Propagation: A One-Timestep Strategy for Spiking Neural Networks. AAAI Conference on Artificial Intelligence, 2026.
- Xiangfei Yang, Jian Song, Xuetao Zhang, Donglin Wang (2024). Adaptive Spiking TD3+BC for Offline-to-Online Spiking Reinforcement Learning. International Joint Conference on Neural Networks, pp. 1–6. IEEE. DOI: 10.1109/IJCNN60899.2024.10650965.
- Jian Song, Xiangfei Yang, Donglin Wang (2024). Calibrating the Converted Spiking Reinforcement Learning. International Conference on Intelligent Computing, pp. 419–431. Springer Nature Singapore. DOI: 10.1007/978-981-97-5581-3_34.
RESEARCH EXPERIENCE
Machine Intelligence Lab, Westlake University Jan. 2023 – Nov. 2025
Research Assistant Hangzhou, China
◦ Conducted research at the intersection of neuromorphic computing, event-driven perception, and RL.
◦ Developed spiking models for point-cloud recognition, imitation learning, and ANN-SNN calibration.
◦ Contributed to efficient neural computation methods for real-time sensing and decision-making.
◦ Developed spiking models for point-cloud recognition, imitation learning, and ANN-SNN calibration.
◦ Contributed to efficient neural computation methods for real-time sensing and decision-making.
PROJECTS
Activation-wise Propagation (AMP2): A One-Timestep Strategy for SNNs 2024 – 2025
Spiking neural networks
◦ One-timestep SNN training with activation-wise propagation across MLP, CNN, and Transformer backbones.
◦ Applied to point-cloud and event-stream recognition with better ModelNet40 and ScanObjectNN results.
◦ One-timestep SNN training with activation-wise propagation across MLP, CNN, and Transformer backbones.
◦ Applied to point-cloud and event-stream recognition with better ModelNet40 and ScanObjectNN results.
SpikeCEIL: Spiking Contextual Imitation Learning for Continuous Control 2023 – 2025
Imitation learning
◦ Spiking CEIL for online/offline imitation learning in continuous control.
◦ Added lfd/lfo support, source/target evaluation, and MuJoCo training scripts.
◦ Spiking CEIL for online/offline imitation learning in continuous control.
◦ Added lfd/lfo support, source/target evaluation, and MuJoCo training scripts.
Adaptive Spiking TD3+BC for Offline-to-Online Fine-Tuning 2024 – 2025
Offline-to-online RL
◦ Adaptive offline-to-online fine-tuning Spiking RL.
◦ Combined D4RL distillation, REDQ-style critics, and behavior cloning for stable MuJoCo experiments.
◦ Adaptive offline-to-online fine-tuning Spiking RL.
◦ Combined D4RL distillation, REDQ-style critics, and behavior cloning for stable MuJoCo experiments.
Spiking TD3 / TD3+BC Baselines 2023 – 2025
Continuous control
◦ Spiking actor-critic baselines for TD3, DDPG, and TD3+BC.
◦ Added membrane-state handling and replay-buffer pipelines for stable SNN training.
◦ Spiking actor-critic baselines for TD3, DDPG, and TD3+BC.
◦ Added membrane-state handling and replay-buffer pipelines for stable SNN training.
SKILLS
- Programming: Python, PyTorch, Bash
- Learning & Modeling: Deep Reinforcement Learning, Offline RL, Imitation Learning, SNNs, ANN-to-SNN Calibration
- Robotics & Perception: Event Cameras, Point Clouds, VLA, World Models, Physical Interaction Modeling
- Frameworks: SpikingJelly, Gym, D4RL, TensorBoardX, Weights & Biases
- Languages: Mandarin (native), English (fluent)
HONORS AND AWARDS
CCNU Eighty-Two Academic Excellence Scholarship — School of Life Science Mar. 2019
CCNU Mathematics Modeling Summer Camp Competition, 2nd Prize — School of Mathematics Jan. 2020
National Excellent Project — Undergraduate Innovation Training Program; PHR1 gene senescence project Oct. 2019
Biological Experiment 3rd Prize — Hubei Province Experimental Skills Competition Oct. 2018