ASCENT: Online Test-Time Training of Long-Horizon Agents via Self-Distillation of Verified Experience
Research
Research News
New papers, breakthroughs and academic AI research
Pythia: Toward Foundation World Models for Multimodal Time Series
Research
What Does a Harness Repair? A Preregistered Study of Visibility, Baseline Adequacy and Evaluation Defects
Research
Expanding LLM Reasoning
Research
Voltic: Distinguishing Volatility from Stochasticity in Recurrent Memory
Research
AdaSpark: Adaptive DSpark with Online Learning for Tree Verification and N-gram Fill
Research
AEGIS: Differentiable Mars Climate Model with Neural Closures
Research
CurveCodec 2: Skeleton-agnostic animation compression with a learned entropy model
Research
Erased, Rerouted, or Rescaled? Post-Training and the Causal Quotient of a Language Model's Belief State
Research
Selecting Long-Horizon Trajectories for Reliable and Efficient Terminal-Agent Training
Research
Off-Policy Merging Beats On-Policy Self-Distillation for Continual Learning
Research
Reward-DAgger: Robot-Gated Interactive Imitation Learning with General-Purpose Progress-Based Reward Models
Research