Selective Agent Guidance via Entropy: Learning Autonomous Policies from Imperfect VLM Teachers
2026-09-01 • Artificial Intelligence
Artificial IntelligenceComputation and LanguageMachine Learning
AI summaryⓘ
The authors study how to use big vision-language models (VLMs) to teach smaller agents without relying on them all the time, since VLMs are slow and not always perfect. They created a method called SAGE that only asks the VLM for help when the agent is unsure, then learns from those suggestions using reinforcement learning. Their agents eventually act on their own without needing the VLM and can sometimes even do better than the VLM. This approach works best when the VLM can guide the agent to good actions, but is less helpful if the agent can figure things out alone or the VLM’s advice isn’t useful.
Vision-Language ModelsReinforcement LearningPolicy DistillationSparse RewardsExplorationAgent GuidanceEntropyTeacher-Student LearningOnline LearningVisual Reasoning
Authors
Matteo Merler, Giovanni Bonetta, Davide Zago, Rossella Cancelliere, Bernardo Magnini
Abstract
Vision-Language Models (VLMs) provide useful priors for interactive decision-making, but using them directly as policies is expensive and brittle: they must be queried at every step, do not improve from environment interaction, and can repeat systematic errors. We study how to learn a cheap autonomous policy from an online, expensive, and imperfect but informative VLM teacher. We propose SAGE (Selective Agent Guidance via Entropy), a framework that queries a VLM only when the learner is uncertain, executes the suggested action during training, and distills guidance into a lightweight Reinforcement Learning (RL) policy. Because VLM advice is not always reliable, SAGE can weight teacher-action distillation using environment-derived advantages rather than treating all suggestions as equally useful. Across sparse-reward visual reasoning and navigation tasks, SAGE learns policies that act without VLM guidance at evaluation time and improves over unguided RL in several environments, including settings where the learned policy exceeds its VLM teacher. The results show that selective guidance is most beneficial when the VLM can help the agent discover high-reward trajectories, and less useful when unguided exploration already succeeds or teacher actions do not lead to informative experience. SAGE also reduces VLM usage by prompting the teacher only on a fraction of training steps and requiring no VLM calls at deployment. Overall, our results suggest that VLMs don't need to be used as fixed policies to be useful; they can instead act as temporary, imperfect sources of guidance whose value is tested and internalized through interaction.