Learning to Speak and Act in a Fantasy Text Adventure Game

By Jack Urbanek, Angela Fan, Siddharth Karamcheti, Saachi Jain, Samuel Humeau, Emily Dinan, Tim Rocktaschel, Douwe Kiela, Arthur Szlam, Jason Aaron Edward Weston. 2019

226 citations recorded by the source corpus.

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.

Subfields: Natural Language Processing

How researchers see this paper

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

In Natural Language Processing

Citations in the corpus

Open-Domain Conversational Agents: Current Progress, Open Problems, and Future Directions (2020, methodology): “This is a continuation of the approach taken in multiple papers using role-played situations (Rashkin et al., 2019), assigned traits (Shuster et al., 2018, 2019), assigned movie preferences (Kang et al., 2019), or even an entire fantasy universe (Urbanek et al., 2019; Prabhumoye et al., 2020).”
Language Models that Seek for Knowledge: Modular Search & Generation for Dialogue and Prompt Completion (2022, methodology): “This was applied successfully to Wizard of Wikipedia (Dinan et al., 2019), QA (Lee et al., 2019a) and LIGHT tasks (Urbanek et al., 2019).”
Queens Are Powerful Too: Mitigating Gender Bias in Dialogue Generation (2019, methodology): “For example, LIGHT (Urbanek et al., 2019) was created by crowdworkers in stages: crowd-workers were ﬁrst assigned a character (with previously crowdsourced names such as “farmer” or “witch”), as well as a previously crowdsourced persona , or short textual description of the char-# acter.”
BlenderBot 3: a deployed conversational agent that continually learns to responsibly engage (2022, background): “and the LIGHT (Urbanek et al., 2019) and LIGHT WILD (Shuster et al.”
Keep CALM and Explore: Language Models for Action Generation in Text-based Games (2020, background): “In a slightly different setting, Urbanek et al. (2019) trained BERT (Devlin et al., 2018) to generate contextually relevant dialogue utterances and actions in fantasy settings.”
Recipes for Building an Open-Domain Chatbot (2020, background): “See Urbanek et al. (2019); Prabhumoye et al. (2020) for other work by some of the authors connecting dialogue models to rich environments.”
Multi-Dimensional Gender Bias Classification (2020, neutral): “, 2018), OpenSubtitles (dialogue from movies) (Lison and Tiedemann, 2016), and LIGHT (chit-chat fantasy dialogue) (Urbanek et al., 2019).”
Beyond Goldfish Memory: Long-Term Open-Domain Conversation (2021, background): “from fantasy role-playing (Urbanek et al., 2019; Rameshkumar and Bailey, 2020) and TV shows as well (Poria et al.”
Towards Debiasing Sentence Representations (2020, background): “More recently, sentence representations such as ELMo [20], BERT [8], and GPT [21] have become the preferred choice for encoding text [30, 17, 31, 23, 26].”
The Dialogue Dodecathlon: Open-Domain Knowledge and Image Grounded Conversational Agents (2019, background): “It must be able to handle situated conversations and demonstrate empathy (Empa-thetic Dialog, LIGHT) .”
Interactive Fiction Games: A Colossal Adventure (2019, background): “Second, recent work on transformer-based neural architectures has yielded impressive gains in many NLP tasks (Devlin et al. 2018), including text-adventure game dialogues (Urbanek et al. 2019).”
Generating Fact Checking Briefs (2020, methodology): “BLINK combines a bi-encoder (Urbanek et al., 2019; Humeau et al., 2019) that identifies candidates with a cross-encoder that models the interaction between mention context and en-”
Deploying Lifelong Open-Domain Dialogue Learning (2020, methodology): “LIGHT MTurk (Urbanek et al., 2019) 8,538 110,877 110,877 33,789 18.”
Help, Anna! Visual Navigation with Natural Multimodal Assistance via Retrospective Curiosity-Encouraging Imitation Learning (2019, background): “There are efforts in building complex interactive text-based worlds (Cˆot´e et al., 2018; Urbanek et al., 2019) but the lack of a graphical component makes them not suitable for visually grounded learning.”
Reason first, then respond: Modular Generation for Knowledge-infused Dialogue (2021, background): “For example, in LIGHT, a player might ask another character where to ﬁnd a certain key to complete their quest.”
Graph Constrained Reinforcement Learning for Natural Language Action Spaces (2020, background): “Urbanek et al. (2019) introduce Light, a dataset of crowdsourced text-adventure game dialogs focusing on giving collaborative agents the ability to generate contextually relevant dialog and emotes.”
RedditBias: A Real-World Resource for Bias Evaluation and Debiasing of Conversational Language Models (2021, background): “for gender bias and extend LIGHT (Urbanek et al., 2019), a resource for grounded dialog, with crowdsourced gender-balanced utterances.”
AgentBench: Evaluating LLMs as Agents (2023, methodology): “Historically, text-based game environments [57; 17; 30; 82] have been employed for language agent evaluation.”