Generative modeling within constrained sets is essential for scientific and engineering applications involving physical, geometric, or safety requirements (e.g., molecular generation, robotics). We present a unified framework for constrained diffusion models on generic nonconvex feasible sets $Σ$ that simultaneously enforces equality and inequality constraints throughout the diffusion process. Our framework incorporates both overdamped and underdamped dynamics for forward and backward sampling. A key algorithmic innovation is a computationally efficient landing mechanism that replaces costly and often ill-defined projections onto $Σ$, ensuring feasibility without iterative Newton solves or projection failures. By leveraging underdamped dynamics, we accelerate mixing toward the prior distribution, effectively alleviating the high simulation costs typically associated with constrained diffusion. Empirically, this approach reduces function evaluations and memory usage during both training and inference while preserving sample quality. On benchmarks featuring equality and mixed constraints, our method achieves comparable sample quality to state-of-the-art baselines while significantly reducing computational cost, providing a practical and scalable solution for diffusion on nonconvex feasible sets.

Howevb isnon-zero Similarlyr(s, ) are features: r(s, a)= (s, a)>w. Assumption A3(Featureexcitation)Forapolicy withstationarydistributiond (s, a), define =E(s,a) d [ (s, a) (s, a)>].

In this paper, we introduce a new class of structured matrices, which unifies and generalizes structured classes from previous works.

Neural Information Processing Systems http://nips.cc/

One of the core problems of modern statistics and machine learning is to approximate difficultto-compute probability distributions.

Reinforcement learning (RL) algorithms often require a large number of data samples to learn a control policy. As a result, training them directly on the real-world systems is expensive and potentially dangerous.

( , 0)

This workproposesMaskTune, a masking strategy that prevents over-reliance on spurious (or a limited number of) features. MaskTuneforces the trained model to explore new features during asingleepochfinetuning bymasking previously discoveredfeatures.MaskTune, unlike earlier approaches for mitigating shortcut learning, does not require any supervision, suchasannotating spurious features orlabels forsubgroup samples in a dataset.