"inference-time reward alignment" Papers

$\Psi$-Sampler: Initial Particle Sampling for SMC-Based Inference-Time Reward Alignment in Score Models

Taehoon Yoon, Yunhong Min, Kyeongmin Yeo et al.