Tag
ConicFlow 将流匹配模型的约束采样建模为推理时的最优控制问题，采用滚动时域原对偶算法在满足硬终端约束的同时高奖励生成并少偏离预训练轨迹，理论上给出了次优性保证并在实验中优于现有基线。
The paper introduces MintFlow, a training-free constrained sampling framework for flow matching models that enforces constraints via minimal, closed-form perturbations of intermediate flow states, preserving the pretrained generative distribution better than state-of-the-art constrained methods.
Needle 2 is an open, 45M-parameter AI model for tool calling and structured extraction, optimized to run in browsers at 14MB with guaranteed JSON output via constrained sampling.
Discusses methods for generating structured outputs from large language models using constrained sampling techniques.
This paper proposes primal-dual inference for constrained diffusion models, jointly inferring the optimal distribution and its dual variable via a dual-conditioned score network, with convergence guarantees and applications in wireless resource allocation and portfolio management.
DiRecT introduces a training-free algorithm for safe diffusion-based planning that enforces constraints only on final clean trajectories using receding-horizon denoising, improving safety and performance over existing methods.
Proposes Constraint-Aware Flow Matching, a novel end-to-end framework that aligns the model's learning dynamics with constrained sampling procedure, mitigating distributional shift from projection corrections for high-quality constrained generation.