Gated-Attention Architectures for Task-Oriented Language Grounding

TL;DR

This paper introduces a neural architecture that combines visual and language inputs using Gated-Attention for task-oriented language grounding in 3D environments, enabling agents to follow natural language instructions without prior knowledge.

Contribution

It presents an end-to-end trainable model that effectively grounds language in visual environments using a novel Gated-Attention mechanism and reinforcement learning, tested on a new 3D environment.

Findings

Effective on unseen instructions and maps

Outperforms baseline models in grounding tasks

Demonstrates generalization in complex 3D environments

Abstract

To perform tasks specified by natural language instructions, autonomous agents need to extract semantically meaningful representations of language and map it to visual elements and actions in the environment. This problem is called task-oriented language grounding. We propose an end-to-end trainable neural architecture for task-oriented language grounding in 3D environments which assumes no prior linguistic or perceptual knowledge and requires only raw pixels from the environment and the natural language instruction as input. The proposed model combines the image and text representations using a Gated-Attention mechanism and learns a policy to execute the natural language instruction using standard reinforcement and imitation learning methods. We show the effectiveness of the proposed model on unseen instructions as well as unseen maps, both quantitatively and qualitatively. We also introduce a novel environment based on a 3D game engine to simulate the challenges of task-oriented language grounding over a rich set of instructions and environment states.