Text-based adventures of the golovin AI agent

By Bartosz Kostka, Jaroslaw Kwiecieli, Jakub Kowalski, Paweł Rychlikowski. 2017

31 citations recorded by the source corpus.

The domain of text-based adventure games has been recently established as a new challenge of creating the agent that is both able to understand natural language, and acts intelligently in text-described environments. In this paper, we present our approach to tackle the problem. Our agent, named Golovin, takes advantage of the limited game domain. We use genre-related corpora (including fantasy books and decompiled games) to create language models suitable to this domain. Moreover, we embed mechanisms that allow us to specify, and separately handle, important tasks as fighting opponents, managing inventory, and navigating on the game map. We validated usefulness of these mechanisms, measuring agent's performance on the set of 50 interactive fiction games. Finally, we show that our agent plays on a level comparable to the winner of the last year Text-Based Adventure AI Competition.

Subfields: Artificial Intelligence, Sociology and Political Science, LLM Agents & Tool Use, Reinforcement Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Sociology and Political Science

In LLM Agents & Tool Use

In Reinforcement Learning

Citations in the corpus

Keep CALM and Explore: Language Models for Action Generation in Text-based Games (2020, background): “Other approaches (Kostka et al., 2017; Hausknecht et al., 2019b) trained simple n-gram language models to learn affordances for action generation.”
A Survey of Reinforcement Learning Informed by Natural Language (2019, background): “They note that existing work on word-level embedding models for text games (e.g. [Kostka et al. , 2017; Fulda et al. , 2017]) achieve good performance only on easy tasks.”
NAIL: A General Interactive Fiction Agent (2019, background): “Golovin (Kostka et al., 2017) generates actions using a large set of predeﬁned command patterns extracted through walkthroughs, tutorials, and decompilation of games.”
Learn What Not to Learn: Action Elimination with Deep Reinforcement Learning (2018, methodology): “Previous work on TBG used pre-trained embeddings directly for control (Kostka et al., 2017; Fulda et al., 2017).”