paper
EvolveNav: Proactive Preflection and Self-Evolving Memory for Zero-Shot Object Goal Navigation
paperactiveprovisional
evolvenav-proactive-preflection-and-self-evolving-memory-for-zero-shot-object-goal-navigation-1cf21696·1 events·first seen 9h agoAliases: EvolveNav: Proactive Preflection and Self-Evolving Memory for Zero-Shot Object Goal Navigation
Co-occurring entities
More like this (12)
Self-Refining Agentic Reinforcement Learning for Vision-Conditioned UAV NavigationEvolveNavVisual Verification Enables Inference-time Steering and Autonomous Policy ImprovementHumanoid-GPT: Scaling Data and Structure for Zero-Shot Motion TrackingTempoVLA: Learning Speed-Controllable Vision-Language-Action Policieszero-shot learningzero-shot evaluationLearning Red Agent Policy from Observations for Neurosymbolic Autonomous Cyber Agentserror-aware specialization objectiveSelf-Supervised PretrainingGradient-Guided Reward OptimizationLearning from the Self-future: On-policy Self-distillation for dLLMs
Recent events (1)
EvolveNav: Self-evolving memory and preflection for zero-shot object-goal navigation
EvolveNav is a new framework for Zero-Shot Object-Goal Navigation (ZS-OGN) that enables test-time improvement through a self-evolving agentic rule memory built from past trajectories. A retrieval strategy based on upper confidence bound balances semantic relevance and historical success when selecting rules, while a memory-guided preflection module forecasts action outcomes before execution to reduce inefficient exploration. The method achieves a 10.1% improvement in success rate over existing zero-shot baselines with fewer unnecessary steps.