arXiv:2502.01384v3 Announce Type: replace-cross
Abstract: Discrete diffusion models have recently gained significant attention due to their ability to process complex discrete structures for language modeling. However, fine-tuning these models with policy gradient methods, as is commonly done in Reinforcement Learning from Human Feedback (RLHF), remains a challenging task. We propose an efficient, broadly applicable, and theoretically justified policy gradient algorithm, called Score Entropy Policy Optimization (SEPO), for fine-tuning discrete diffusion models over non-differentiable rewards. Our numerical experiments across several discrete generative tasks demonstrate the scalability and efficiency of our method. Our code is available at https://github.com/ozekri/SEPO.
The Impact of Patient-Generated Health Data From Mobile Health Technologies on Health Care Management and Clinical Decision-Making: Narrative Scoping Review
Background: Long-term health conditions and multimorbidity are increasing globally placing an unsustainable pressure on healthcare systems. Mobile health technologies, or mHealth, enable the collection of