Knowledge-enhanced Agents for Interactive Text Games

Prateek Chhikara, Jiarui Zhang, Filip Ilievski, Jonathan Francis, Kaixin Ma, Daniel Garijo (Editor)

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

Communication via natural language is a key aspect of machine intelligence, and it requires computational models to learn and reason about world concepts, with varying levels of supervision. Significant progress has been made on fully-supervised non-interactive tasks, such as question-answering and procedural text understanding. Yet, various sequential interactive tasks, as in text-based games, have revealed limitations of existing approaches in terms of coherence, contextual awareness, and their ability to learn effectively from the environment. In this paper, we propose a knowledge-injection framework for improved functional grounding of agents in text-based games. Specifically, we consider two forms of domain knowledge that we inject into learning-based agents: memory of previous correct actions and affordances of relevant objects in the environment. Our framework supports two representative model classes: reinforcement learning agents and language model agents. Furthermore, we devise multiple injection strategies for the above domain knowledge types and agent architectures, including injection via knowledge graphs and augmentation of the existing input encoding strategies. We experiment with four models on the 10 tasks in the ScienceWorld text-based game environment, to illustrate the impact of knowledge injection on various model configurations and challenging task settings. Our findings provide crucial insights into the interplay between task properties, model architectures, and domain knowledge for interactive contexts.
Original languageEnglish
Title of host publicationK-CAP '23
Subtitle of host publicationProceedings of the 12th Knowledge Capture Conference 2023
EditorsBrent Venable, Daniel Garijo, Brian Jalaian
PublisherAssociation for Computing Machinery, Inc
Pages157-165
Number of pages9
ISBN (Electronic)9798400701412
DOIs
Publication statusPublished - 2023
EventProceedings of the 12th Knowledge Capture Conference 2023 - Pensacola, United States
Duration: 5 Sept 20237 Sept 2023
https://doi.org/10.1145/3587259.3627561

Conference

ConferenceProceedings of the 12th Knowledge Capture Conference 2023
Abbreviated title K-CAP 2023
Country/TerritoryUnited States
CityPensacola
Period5/09/237/09/23
Internet address

Fingerprint

Dive into the research topics of 'Knowledge-enhanced Agents for Interactive Text Games'. Together they form a unique fingerprint.

Cite this