Why Build an Assistant in Minecraft?
arXiv:1907.09273
Abstract
In this document we describe a rationale for a research program aimed at building an open "assistant" in the game Minecraft, in order to make progress on the problems of natural language understanding and learning from dialogue.
References in corpus (34)
- Playing Atari with Deep Reinforcement Learning
- Progressive Growing of GANs for Improved Quality, Stability, and Variation
- Exploring the Limits of Language Modeling
- Learning to Map Sentences to Logical Form: Structured Classification with Probabilistic Categorial Grammars
- Seq2SQL: Generating Structured Queries from Natural Language using Reinforcement Learning
- StarCraft II: A New Challenge for Reinforcement Learning
- Neural-Symbolic VQA: Disentangling Reasoning from Vision and Language Understanding
- Matterport3D: Learning from RGB-D Data in Indoor Environments
- AI2-THOR: An Interactive 3D Environment for Visual AI
- Building Generalizable Agents with a Realistic and Rich 3D Environment
- Neural Programmer-Interpreters
- Control of Memory, Active Perception, and Action in Minecraft
- MINOS: Multimodal Indoor Simulator for Navigation in Complex Environments
- Towards Deep Conversational Recommendations
- Language to Logical Form with Neural Attention
- TextWorld: A Learning Environment for Text-based Games
- Programming with a Differentiable Forth Interpreter
- Learning Dependency-Based Compositional Semantics
- TorchCraft: a Library for Machine Learning Research on Real-Time Strategy Games
- Dialog-based Language Learning
- TerpreT: A Probabilistic Programming Language for Program Induction
- HoME: a Household Multimodal Environment
- Learning through Dialogue Interactions by Asking Questions
- Hierarchical and Interpretable Skill Acquisition in Multi-task Reinforcement Learning
- MazeBase: A Sandbox for Learning from Games
- Neural Symbolic Machines: Learning Semantic Parsers on Freebase with Weak Supervision
- Teaching Machines to Describe Images via Natural Language Feedback
- End-to-end optimization of goal-driven and visually grounded dialogue systems
- Differentiable Programs with Neural Libraries
- Explore, Exploit or Listen: Combining Human Feedback and Policy Model to Speed up Deep Reinforcement Learning in 3D Worlds
- Data Recombination for Neural Semantic Parsing
- CraftAssist: A Framework for Dialogue-enabled Interactive Agents
- Deep Reinforcement Learning with Model Learning and Monte Carlo Tree Search in Minecraft
- Learning to Ask Questions in Open-domain Conversational Systems with Typed Decoders