Molei Tao

Associate Professor@Georgia Institute of Technology·美国·OpenReview

研究方向

optimization · sampling · computational mathematics · dynamical systems · generative modeling · AI for sciences · diffusion model · deep learning theory

Molei Tao

What Exactly Does Guidance Do in Masked Discrete Diffusion Models

Proximal Diffusion Neural Sampler

AlignFlow: Improving Flow-based Generative Models with Semi-Discrete Optimal Transport

Provable Separations between Memorization and Generalization in Diffusion Models

Complexity Analysis of Normalizing Constant Estimation: from Jarzynski Equality to Annealed Importance Sampling and beyond

Improving Reasoning for Diffusion Language Models via Group Diffusion Policy Optimization

Generalization of Diffusion Models Arises with a Balanced Representation Space

Improving Classifier-Free Guidance in Masked Diffusion: Low-Dim Theoretical Insights with High-Dim Impact

Efficient Diffusion Models under Nonconvex Equality and Inequality constraints via Landing

Convergence of Near-Linear Width ReLU Networks with Unbalanced Initialization

Enhancing Reasoning for Diffusion LLMs via Distribution Matching Policy Optimization

TR2-D2: Tree Search Guided Trajectory-Aware Fine-Tuning for Discrete Diffusion

Mimicking or Reasoning: Rethinking Multi-Modal In-Context Learning in Vision-Language Models

A Closer Look at Model Collapse: From a Generalization-to-Memorization Perspective

MDNS: Masked Diffusion Neural Sampler via Stochastic Optimal Control

Variational Learning Finds Flatter Solutions at the Edge of Stability

Fast Non-Log-Concave Sampling under Nonconvex Equality and Inequality Constraints with Landing

Provable Benefit of Annealed Langevin Monte Carlo for Non-log-concave Sampling

Trivialized Momentum Facilitates Diffusion Generative Modeling on Lie Groups

Diffusion Generative Modeling for Spatially Resolved Gene Expression Inference from Histology Images

Non-equilibrium Annealed Adjoint Sampler

Fast Solvers for Discrete Diffusion Models: Theory and Applications of High-Order Algorithms

Diffuse Everything: Multimodal Diffusion Models on Arbitrary State Spaces

Plug-and-Play Controllable Generation for Discrete Masked Models