Data-driven nonlinear model reduction
Learning low-dimensional dynamics from high-dimensional, multiscale stochastic systems while preserving the geometry and structure that make long-time prediction trustworthy.
My research lies at the intersection of scientific computing and machine learning, with mathematical structure as the unifying theme. I use machine learning to discover and preserve structure in stochastic scientific models, and ideas from scientific computing, probability, and dynamical systems to make modern AI more scalable and reliable.
Learning low-dimensional dynamics from high-dimensional, multiscale stochastic systems while preserving the geometry and structure that make long-time prediction trustworthy.
Connecting entropic optimal transport with attention to develop analytic sensitivity methods and IO-aware GPU kernels for large-scale Sinkhorn computation.
Building probabilistic and dynamical foundations for memory, stability, discretization, and quantization in selective state space models and long-context inference.
Collaborative work on fine-grained audio-visual reasoning, video moment retrieval, efficient token selection, and multimodal generative modeling.
Awarded support for research in nonlinear model reduction, optimal transport, and state space models.
$200,000 · September 2026–August 2029
PI: Felix X.-F. Ye · Co-PI: Barbara Giunti
$170,265 total · January 2025–December 2026
Single PI · Includes an approved extension through December 31, 2026
$42,000 · 2023–2028
Single PI · Travel Support for Mathematicians
$10,000 in cloud compute · December 2025–December 2026
Single PI
Prior support: AMS–Simons Travel Grant, $5,000, 2020–2023.
Forty-third International Conference on Machine Learning (ICML 2026) · Code
SIAM Journal on Scientific Computing, 47(3), C630–C654
Journal of Nonlinear Science, 34, Article 22
* Corresponding author.
Listed in reverse chronological order. See Google Scholar for citation records.
RatSeizure: A Benchmark and Saliency-Context Transformer for Rat Seizure Localization
Ting Yu Tsai, An Yu, Linda Lee, Felix X.-F. Ye, Daniel S. Shin, Tsang-Jiun Kao, Xingjie Li, and Ming-Ching Chang · MICCAI 2026, accepted.
SMART: Shot-Aware Multimodal Video Moment Retrieval with Audio-Enhanced MLLM
An Yu, Weiyu Lu, Jiazhi Li, Zhaorun Zhang, Yifan Shen, Felix X.-F. Ye, and Ming-Ching Chang · IEEE Transactions on Multimedia, accepted.
Memory Along the Chain: Disentangling Dynamics from Discretization in Mamba
Felix X.-F. Ye, Ting Yu Tsai, Ming-Ching Chang, and Davis Wertheimer · Submitted.
DropKV: Decoupling Residual-Output Perturbation for Near-Optimal KV-Cache Eviction
Aozhong Zhang, Selcuk Gurses, Yanxia Deng, Naigang Wang, Chi-Chun Liu, Davis Wertheimer, Derrick Liu, Xin Li, Zi Yang, Felix X.-F. Ye, and Penghang Yin · Submitted.
FAVE: A Structured Benchmark for Fine-Grained Audio-Visual Temporal Evaluation in Multimodal LLMs
Weiyu Lu, An Yu, Jiazhi Li, Zhaorun Zhang, Felix X.-F. Ye, and Ming-Ching Chang · CVPR 2026, Poster.
Minimal Intersection Radius for n Growing, Non-Homogeneous Ellipsoids in Rd
Barbara Giunti, Sean Hill, and Felix X.-F. Ye · Submitted.
Geometric Regularization of Autoencoders via Observed Stochastic Dynamics
Sean Hill and Felix X.-F. Ye · Submitted.
ReDiPrune: Relevance-Diversity Pre-Projection Token Pruning for Efficient Multimodal LLMs
An Yu, Ting Yu Tsai, Zhaorun Zhang, Weiyu Lu, Felix X.-F. Ye, and Ming-Ching Chang · Submitted.
FlashSinkhorn: IO-Aware Entropic Optimal Transport on GPU
Felix X.-F. Ye, Xingjie Li, An Yu, Ming-Ching Chang, Linsong Chu, and Davis Wertheimer · ICML 2026, Oral (Top 0.7%).
Diffuse Everything: Multimodal Diffusion Models on Arbitrary State Spaces
Kevin Rojas, Yuchen Zhu, Sichen Zhu, Felix X.-F. Ye, and Molei Tao · ICML 2025, Poster.
Robust First- and Second-Order Differentiation for Regularized Optimal Transport
Xingjie Li, Fei Lu, Molei Tao, and Felix X.-F. Ye · SIAM Journal on Scientific Computing, 47(3), C630–C654.
Nonlinear Model Reduction for Slow-Fast Stochastic Systems Near Unknown Invariant Manifolds
Felix X.-F. Ye, Sichen Yang, and Mauro Maggioni · Journal of Nonlinear Science, 34, Article 22.
Estimate Exponential Memory Decay in Hidden Markov Model and Its Applications to Inference
Felix X.-F. Ye, Yi-an Ma, and Hong Qian · Physica D: Nonlinear Phenomena, 460, 134053.
NySALT: Nyström-type Inference-Based Schemes Adaptive to Large Time-Stepping
Xingjie Li, Fei Lu, Molei Tao, and Felix X.-F. Ye · Journal of Computational Physics, 477, 111952.
ISALT: Inference-Based Schemes Adaptive to Large Time-Stepping for Locally Lipschitz Ergodic Systems
Xingjie Li, Fei Lu, and Felix X.-F. Ye · Discrete and Continuous Dynamical Systems - S, 15(4), 747–771.
Quantifying Information Accumulation Encoded in the Dynamics of Biochemical Signaling
Ying Tang, Adewunmi Adelaja, Felix X.-F. Ye, Eric J. Deeds, Roy Wollman, and Alexander Hoffmann · Nature Communications, 12(1), 1–10.
Synchronization in Discrete-Time, Discrete-State Random Dynamical Systems
Wen Huang, Hong Qian, Shirou Wang, Felix X.-F. Ye, and Yingfei Yi · SIAM Journal on Applied Dynamical Systems, 19(1), 233–251.
Stochastic Dynamics II: Finite Random Dynamical Systems, Linear Representation, and Entropy Production
Felix X.-F. Ye and Hong Qian · Discrete and Continuous Dynamical Systems - B, 24(8), 4341–4366.
Time-Dependent Saddle–Node Bifurcation: Breaking Time and the Point of No Return in a Non-Autonomous Model of Critical Transitions
Jeremiah H. Li, Felix X.-F. Ye, Hong Qian, and Sui Huang · Physica D: Nonlinear Phenomena, 395, 7–14.
Dynamic Looping of a Free-Draining Polymer
Felix X.-F. Ye, Panos Stinis, and Hong Qian · SIAM Journal on Applied Mathematics, 78(1), 104–123.
Stochastic Dynamics: Models for Intrinsic and Extrinsic Noises and Their Applications
Yi-an Ma, Hong Qian, and Felix X.-F. Ye · SCIENTIA SINICA Mathematica, 47(12), 1693–1702.
Stochastic Dynamics: Markov Chains and Random Transformations
Felix X.-F. Ye, Yue Wang, and Hong Qian · Discrete and Continuous Dynamical Systems - B, 21(7), 2337–2361.
Evolution of Recombination Rates in a Multi-Locus, Haploid-Selection, Symmetric-Viability Model
J. R. Chasnov and Felix X.-F. Ye · Theoretical Population Biology, 83, 155–165.