Research Materials
Slides and posters by Jiequn Han, Research Scientist at the Flatiron Institute.
Slides
- Driftlite: Lightweight drift control for inference-time scaling of diffusion models (2026)
- Generative modeling from black-box corruptions via self-consistent stochastic interpolants (2026)
- Test-time generalization for physics through neural operator splitting (2026)
- Data Manifolds as Priors for Inverse Problems: From Regularization to Representation (2025)
- Progressive optimal path sampling for closed-loop optimal control design with deep neural networks (2025)
- Solving High-Dimensional PDEs Using Deep Learning: Original Insights and Recent Progress (2025)
- Provable posterior sampling with denoising oracles via tilted transport (2024)
- Solving High-Dimensional Partial Differential Equations Using Deep Learning (2024)
- Enjoy the Best of Both Worlds: a Neural-Network Warm-Start Approach for PDE Problems (2023)
- Deep Learning Algorithms for High-Dimensional Partial Differential Equations (2022)
- DeepHAM: A global solution method for heterogeneous agent models with aggregate shocks (2022)
- Developing Reduced-Order PDEs With Machine Learning-Based Closure Models (2022)
- Perturbational complexity by distribution mismatch: A systematic analysis of reinforcement learning in reproducing kernel Hilbert space (2022)
- On the Curse of Memory in Recurrent Neural Networks: Approximation and Optimization Analysis (2021)
- Symmetry-Preserving Neural Networks (2021)
Posters
- Driftlite: Lightweight drift control for inference-time scaling of diffusion models (2026)
- Generative modeling from black-box corruptions via self-consistent stochastic interpolants (2026)
- Learning free terminal time optimal closed-loop control of manipulators (2025)
- Provable posterior sampling with denoising oracles via tilted transport (2024)
- Differentiable physics simulations with contacts: Do they have correct gradients w.r.t. position, velocity and control? (2022)