Ruipeng Liu

Ruipeng Liu

PhD Candidate @ Syracuse University

Publications

You can also find my articles on my Google Scholar profile.

Conference Papers

Lipschitz-Regularized Critic Leads to Policy Robustness Against Transition Dynamics Uncertainty

Published in FLAIRS 2026, 2026

This paper studies how Lipschitz regularization of the critic can improve policy robustness under transition dynamics uncertainty.

Recommended citation: Chen, X., Liu, R., Gan, Z., & Katz, G. E. (2026). Lipschitz-regularized critic leads to policy robustness against transition dynamics uncertainty. In Proceedings of the 39th Florida Artificial Intelligence Research Society Conference. Preprint available at arXiv:2404.13879.
Download Paper

Linearithmic Clean-up for Vector-Symbolic Key-Value Memory with Kroneker Rotation Products

Published in NeSy, 2025

This paper introduces a new Vector-Symbolic Architecture (VSA) codebook representation based on Kroneker Rotation Products (krop), enabling clean-up operations with linearithmic O(N log N) time complexity instead of the quadratic complexity typical in existing approaches. The method preserves comparable memory capacity while significantly improving scalability for vector-symbolic key-value memory systems.

Recommended citation: Liu, R., Qiu, Q., Khan, S., & Katz, G. E. (2025). Linearithmic clean-up for vector-symbolic key-value memory with Kroneker Rotation Products. In Proceedings of the 19th International Conference on Neurosymbolic Learning and Reasoning (Proceedings of Machine Learning Research, Vol. 284, pp. 1107–1118). PMLR.
Download Paper | Download Bibtex

On the Feasibility of Single-Pass Full-Capacity Learning in Linear Threshold Neurons with Binary Input Vectors

Published in ICML 2024, 2024

This paper proves that single-pass full-capacity learning rules satisfying the span rule constraint do not exist for linear threshold neurons with binary input vectors. The result establishes a fundamental limitation on simultaneously achieving single-pass learning and maximal memory capacity in this setting.

Recommended citation: Liu, R., He, B., Tahir, N., & Katz, G. E. (2024). On the feasibility of single-pass full-capacity learning in linear threshold neurons with binary input vectors. In Proceedings of the 41st International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 235, pp. 31119–31130). PMLR.
Download Paper | Download Bibtex