ALFRED

The 'ALFRED' dataset is a benchmark that contains a collection of tasks designed to evaluate embodied agents' ability to perform complex actions and reasoning in interactive environments.

Papers using ALFRED (3)

When Should a Robot Think? Resource-Aware Reasoning via Reinforcement Learning for Embodied Robotic Decision-Making2026Recurrent Reasoning with Vision-Language Models for Estimating Long-Horizon Embodied Task Progress2026EPO: Hierarchical LLM Agents with Environment Preference Optimization2024 · 3 cites