Ontologically Faithful Generation of Non-Player Character Dialogues

Anonymous

Paper Type: long - non archival

Keywords: NPC dialogue, knowledge grounding, knowledge conditioning, dialogue trees, in-context learning, ontological constraints, quests

TL;DR: We introduce a new dataset of complex, professionally written NPC dialogue trees from a real video game, plus a new task definition for knowledge-conditioned NPC dialogue generation and a series of strong baseline generation models using GPT-3.

Abstract: We introduce a language generation task grounded in a popular video game environment. KNUDGE (KNowledge Constrained User-NPC Dialogue GEneration) requires models to produce trees of dialogue between video game characters that accurately reflect quest and entity specifications stated in natural language. KNUDGE is constructed from side quest dialogues drawn directly from game data of Obsidian Entertainment’s The Outer Worlds, leading to real-world complexities in generation: (1) dialogues are branching trees as opposed to linear chains of utterances; (2) utterances must remain faithful to the game lore– character personas, backstories, and entity relationships; and (3) a dialogue must accurately reveal new quest details to the human player. We report results for a set of neural generation models using supervised and in-context learning techniques; we find competent performance but room for future work addressing the challenges of creating realistic, game-quality dialogues.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2212.10618/code)