ALFWorld: Aligning Text and Embodied Environments for Interactive Learning

Abstract

Given a simple request like Put a washed apple in the kitchen fridge, humans can reason in purely abstract terms by imagining action sequences and scoring their likelihood of success, prototypicality, and efficiency, all without moving a muscle. Once we see the kitchen in question, we can update our abstract plans to fit the scene. Embodied agents require the same abilities, but existing work does not yet provide the infrastructure necessary for both reasoning abstractly and executing concretely. We address this limitation by introducing ALFWorld, a simulator that enables agents to learn abstract, text based policies in TextWorld (Côté et al., 2018) and then execute goals from the ALFRED benchmark (Shridhar et al., 2020) in a rich visual environment. ALFWorld enables the creation of a new BUTLER agent whose abstract knowledge, learned in TextWorld, corresponds directly to concrete, visually grounded actions. In turn, as we demonstrate empirically, this fosters better agent generalization than training only in the visually grounded environment. BUTLER's simple, modular design factors the problem to allow researchers to focus on models for improving every piece of the pipeline (language understanding, planning, navigation, and visual scene understanding).

arXiv commentICLR 2021; Data, code, and videos are available at alfworld.github.io

References in corpus (17)

• Adam: A Method for Stochastic Optimization
• VQA: Visual Question Answering
• Deep Convolutional Inverse Graphics Network
• Deep Recurrent Q-Learning for Partially Observable MDPs
• Layer Normalization
• Recurrent World Models Facilitate Policy Evolution
• AI2-THOR: An Interactive 3D Environment for Visual AI
• Relevance of Unsupervised Metrics in Task-Oriented Dialogue for Evaluating Natural Language Generation
• Language as an Abstraction for Hierarchical Deep Reinforcement Learning
• TextWorld: A Learning Environment for Text-based Games
• Counting to Explore and Generalize in Text-based Games
• The NetHack Learning Environment
• Graph Constrained Reinforcement Learning for Natural Language Action Spaces
• Hierarchical Decision Making by Generating and Following Natural Language Instructions
• Unnatural Language Processing: Bridging the Gap Between Synthetic and Natural Language Data
• RTFM: Generalising to Novel Environment Dynamics via Reading
• Language is Power: Representing States Using Natural Language in Reinforcement Learning

Cited by in corpus (7)

• TidyBot: Personalized Robot Assistance with Large Language Models
• On Grounded Planning for Embodied Tasks with Language Models
• FILM: Following Instructions in Language with Modular Methods
• Hierarchical Task Learning from Language Instructions with Unified Transformers and Self-Monitoring
• Modeling Worlds in Text
• SocialAI: Benchmarking Socio-Cognitive Abilities in Deep Reinforcement Learning Agents
• Moral Stories: Situated Reasoning about Norms, Intents, Actions, and their Consequences