A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Aniri, Jinhe Bi, Peng Liao, et al.
OPD-V introduces a visual on-policy self-distillation paradigm that uses modality-balance trust regions to improve MLLM reasoning while reducing training cost.
Changle Qu, Sunhao Dai, Hengyi Cai, et al.
TurnSight is a turn-level hindsight self-distillation framework for tool-integrated reasoning that derives supervision from execution-conditioned hindsight and cross-horizon agreement to improve credit assignment.
Jiawei Xu, Minghui Liu, Juzheng Zhang, et al.
β-OPSD generalizes on-policy self-distillation by introducing a controllable KL penalty weight β, enabling efficient distillation to approximate policy optimization and improving reasoning performance.
Unknown
Opsdl introduces an on-policy self-distillation method to extend the effective context length of LLMs through post-training.
Jonas Hubotter, Frederike Lubeck, L. Behric, et al.
SDPO converts tokenized feedback from verifiable environments into a dense learning signal via self-distillation, improving sample efficiency and accuracy over RLVR baselines.