arXiv:2608.20271ترجمه شده
User Safety: safe
لطفاً متن مورد نظر را ارسال کنید.
arXiv:2608.19197
SPADE: Self-Play in Adaptive Synthetic Executable Environments
Continuous self-improvement requires an ever-expanding pool of self-generated, diverse, adaptive goals. For language agents, existing training environment pools (hand-curated, statically synthesized, or frozen-verifier)…
arXiv:2608.19182
ADEPT: Accelerating Dexterity via Pre-Training and Post-Training using Reinforcement Learning
We introduce Accelerating Dexterity via Pre-Training (ADEPT), a large-scale reinforcement learning (RL) framework for learning sim-to-real transferable dexterity across high degree-of-freedom (DoF) robot embodiments that…
arXiv:2608.19181
Beyond Teacher Likelihood: Group-Calibrated On-Policy Distillation for Long-Context Reasoning
On-policy distillation (OPD) trains a student on its own responses using dense token-level guidance from a stronger teacher. In long-context tasks, however, token-level teacher support can favor locally plausible respons…
arXiv:2608.19174
Finetuning Strategies for Querying Sounds by Vocal Imitation
This technical report describes our winning submission to the AES AIMLA 2025 Challenge on querying sound effects by vocal imitation. We investigate two complementary fine-tuning strategies: contrastive learning with a fr…