# Learning to Speak and Act in a Fantasy Text Adventure Game

## TL;DR

This paper presents a large-scale crowdsourced text adventure platform for studying grounded dialogue, demonstrating that models grounded in environmental details improve their ability to predict dialogue and actions effectively.

## Contribution

It introduces a novel research platform and evaluates models that incorporate environmental grounding, advancing understanding of grounded dialogue in interactive environments.

## Key findings

- Grounded models outperform ungrounded ones in predicting dialogue and actions.
- Environmental context, including location and objects, enhances model performance.
- Grounding ingredients are crucial for successful dialogue and action prediction.

## Abstract

We introduce a large scale crowdsourced text adventure game as a research platform for studying grounded dialogue. In it, agents can perceive, emote, and act whilst conducting dialogue with other agents. Models and humans can both act as characters within the game. We describe the results of training state-of-the-art generative and retrieval models in this setting. We show that in addition to using past dialogue, these models are able to effectively use the state of the underlying world to condition their predictions. In particular, we show that grounding on the details of the local environment, including location descriptions, and the objects (and their affordances) and characters (and their previous actions) present within it allows better predictions of agent behavior and dialogue. We analyze the ingredients necessary for successful grounding in this setting, and how each of these factors relate to agents that can talk and act successfully.