AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

Generating Interactive Worlds with Text

Angela Fan, Jack Urbanek, Pratik Ringshia, Emily Dinan, Emma Qian, Siddharth Karamcheti, Shrimai Prabhumoye, Douwe Kiela, Tim Rocktaschel, Arthur Szlam, Jason Weston

AAAI 2020 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Emergent communication in referential games · 20 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use Starspace instead of Information Retrieval for location predictionLIGHT location prediction · LIGHT test setHits at 1: 30.0 -> 45.5Improved
Provide location descriptions in addition to names (ablation)Starspace location prediction · LIGHT test setHits at 1: 44.5 -> 45.5Improved
Use Starspace instead of Information Retrieval for object predictionLIGHT object prediction · LIGHT test setHits at 1: 21.9 -> 47.3Improved
Provide location descriptions in addition to names (ablation)Starspace object prediction · LIGHT test setHits at 1: 13.3 -> 47.3Improved
Use BERT Cross-Encoder instead of StarspaceLIGHT character prediction · LIGHT test setHits at 1: 35.7 -> 36.1Improved
Pretrain the generative Transformer on Reddit comments (ablation)LIGHT location generation · LIGHT test setF1: 23.3 -> 26.3Improved
Pretrain the generative Transformer on Reddit comments (ablation)LIGHT character generation · LIGHT test setF1: 28.4 -> 29.3Improved
Pretrain the generative Transformer on Reddit comments (ablation)LIGHT object generation · LIGHT test setF1: 32.5 -> 38.8Improved
Pretrain the generative Transformer on Reddit comments (ablation)LIGHT affordance generation · LIGHT test setF1: 57.0 -> 62.6Improved
Use Starspace-predicted location paths instead of data-created pathsLIGHT world construction · Human pairwise evaluationpreference: over 60%Improved
Add model suggestions to the world-building interfaceInteractive world creation · Nine-location game environment designtime spent: 10-20 minutes -> 10 minutes or lessImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy