About
I am a Shanahan Foundation Fellow at the Allen Institute and the University of Washington. My research investigates how structured representations emerge in neural systems. Combining computational neuroscience, cognitive science, and AI, I develop theories and methods to uncover the recurring motifs that organize perception, memory, prediction, and abstraction across brains and machines. At UW and the Allen Institute, I work with Rajesh Rao, Stefan Mihalas, and Carl Schoonover . Prior to this, I completed my PhD at the Max Planck Institute for Biological Cybernetics under the mentorship of Eric Schulz, Peter Dayan, and Felix Wichmann. I received my master’s degree from the Institute of Neuroinformatics at the University of Zurich and ETH Zurich, and my bachelors from the University of Rochester on Physics, Applied Math, and Computer Science.
Contact
shuchen.wu at alleninstitute.org
Publications
Wu, S., Alaniz, S., Karthik, S., Dayan, P., Schulz, E., & Akata, Z. (2025). Concept-Guided Interpretability via Neural Chunking. The Thirty-ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025).Wu, S., Thalmann, M., & Schulz, E. (2025). Two types of motifs enhance human recall and generalization of long sequences. Communications Psychology, 3(1), 3
Wu, S., Thalmann, M., Dayan, P., Akata, Z., & Schulz, E. (2025). Building, Reusing, and Generalizing Abstract Representations from Concrete Sequences. ICLR 2025
Binz, M., Akata, E., Bethge, M., ... Wu, S., ... Schulz, E. (2024). Centaur: a foundation model of human cognition. arXiv preprint arXiv:2410.20268
Schreiber, A., Wu, S. C., Wu, C. X., Indiveri, G., Schulz, E. (2023). Biologically-plausible hierarchical chunking on mixed-signal neuromorphic hardware. NeurIPS 2023 Workshop on Machine Learning with New Compute Paradigms
Wu, S. C., Élteto, N.,Dasgupta, I., & Schulz, E. (2023). Chunking as a rational solution to the speed–accuracy trade-off in a serial reaction time task. Scientific Reports
Wu, S. C., Élteto, N.,Dasgupta, I., & Schulz, E. (2022). Learning Structure from the Ground-up—Hierarchical Representation Learning by Chunking. 36th Conference on Neural Information Processing Systems (NeurIPS 2022).