1 citations · 1 across the 9 of their papers we have counts for
15 papers
When Should a World Model Move? Loss-Conditioned State Execution
Jintao Xu, Zhengyu Chen, Ben Zhang +2
We introduce loss-conditioned state execution, a model-agnostic method that decides whether to execute a world model's fixed feasible proposal or retain the current state. Predicti…
SVI-Bench: A Dynamic Microworld for Strategic Video Intelligence
Yulu Pan, Han Yi, Seongsu Ha +4
True video intelligence demands more than recognizing what is visible: it requires reasoning about why events unfold, predicting what would change under different conditions, and d…
ISOMORPH: A Supply Chain Digital Twin for Simulation, Dataset Generation, and Forecasting Benchmarks
Zhizhen Zhang, Hyemin Gu, Benjamin J. Zhang +6
Open time-series forecasting (TSF) benchmarks cover retail, energy, weather, and traffic, but supply-chain logistics remains underserved. We introduce ISOMORPH, the first public di…
Vocabulary-size-independent Convergence of Discrete Diffusion Models: adjoint equations induce the right space
Kelvin Kan, Xingjian Li, Benjamin J. Zhang +3
Discrete diffusion has become a leading framework for generative modeling in various applications including language, vision, and biology. Existing convergence theory, however, exh…
Stability of Transformers under Layer Normalization
Kelvin Kan, Xingjian Li, Benjamin J. Zhang +4
Despite their widespread use, training deep Transformers can be unstable. Layer normalization, a standard component, improves training stability, but its placement has often been a…
Probabilistic operator learning: generative modeling and uncertainty quantification for foundation models of differential equations
Benjamin J. Zhang, Siting Liu, Stanley J. Osher +1
In-context operator networks (ICON) are a class of operator learning methods based on the novel architectures of foundation models. Trained on a diverse set of datasets of initial…
ExAct: A Video-Language Benchmark for Expert Action Analysis
Han Yi, Yulu Pan, Feihong He +4
We present ExAct, a new video-language benchmark for expert-level understanding of skilled physical human activities. Our new benchmark contains 3521 expert-curated video question-…
Particle exchange Monte Carlo methods for eigenfunction and related nonlinear problems
Paul Dupuis, Benjamin J. Zhang
We introduce and develop a novel particle exchange Monte Carlo method. Whereas existing methods apply to eigenfunction problems where the eigenvalue is known (e.g., integrals with…
Optimal Control for Transformer Architectures: Enhancing Generalization, Robustness and Efficiency
Kelvin Kan, Xingjian Li, Benjamin J. Zhang +3
We study Transformers through the perspective of optimal control theory, using tools from continuous-time formulations to derive actionable insights into training and architecture…
Proximal optimal transport divergences
Ricardo Baptista, Panagiota Birmpa, Markos A. Katsoulakis +2
We introduce the proximal optimal transport divergence, a novel discrepancy measure that interpolates between information divergences and optimal transport distances via an infimal…
Robustness and Structure Preservation in Flow-Based Generative Models via Wasserstein Path-Space Divergences
Ziyu Chen, Markos A. Katsoulakis, Benjamin J. Zhang
We introduce a novel Wasserstein-1 () path-space divergence for stochastic and deterministic dynamics and establish a Wasserstein Uncertainty Propagation (WUP) theorem that bo…
Combining Wasserstein-1 and Wasserstein-2 proximals: robust manifold learning via well-posed generative flows
Hyemin Gu, Markos A. Katsoulakis, Luc Rey-Bellet +1
We formulate well-posed continuous-time generative flows for learning distributions that are supported on low-dimensional manifolds through Wasserstein proximal regularizations of…
Score-based generative models are provably robust: an uncertainty quantification perspective
Nikiforos Mimikos-Stamatopoulos, Benjamin J. Zhang, Markos A. Katsoulakis
Through an uncertainty quantification (UQ) perspective, we show that score-based generative models (SGMs) are provably robust to the multiple sources of error in practical implemen…
Nonlinear denoising score matching for enhanced learning of structured distributions
Jeremiah Birrell, Markos A. Katsoulakis, Luc Rey-Bellet +2
We present a novel method for training score-based generative models which uses nonlinear noising dynamics to improve learning of structured distributions. Generalizing to a nonlin…
Wasserstein proximal operators describe score-based generative models and resolve memorization
Benjamin J. Zhang, Siting Liu, Wuchen Li +2
We focus on the fundamental mathematical structure of score-based generative models (SGMs). We first formulate SGMs in terms of the Wasserstein proximal operator (WPO) and demonstr…