AI Paper Atlas

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

Learning Knowledge Graph-based World Models of Textual Environments

Prithviraj Ammanabrolu, Mark Riedl

NeurIPS 2021 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Language-grounded reinforcement agents · 2 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
add graph difference prediction (ablation)Worldformer · JerichoWorld KG prediction, overallGraph F1: 31.41 -> 41.06Improved
add graph difference prediction (ablation)Worldformer · JerichoWorld KG prediction, overallToken F1: 41.02 -> 52.45Improved
add multi-task training (ablation)Worldformer · JerichoWorld KG prediction, overallGraph F1: 34.65 -> 41.06Improved
add multi-task training (ablation)Worldformer · JerichoWorld KG prediction, overallToken EM: 42.74 -> 51.32Improved
add SOS loss (ablation)Worldformer · JerichoWorld KG prediction, overallGraph F1: 36.17 -> 41.06Improved
add SOS loss (ablation)Worldformer · JerichoWorld KG prediction, overallToken EM: 48.82 -> 51.32Improved
add all three components to Seq2Seq baseline (ablation)Worldformer · JerichoWorld KG prediction, overallGraph EM: 14.29 -> 39.15Improved
add multi-task training (ablation)Worldformer · JerichoWorld valid action prediction, overallEM: 20.78 -> 23.22Improved
add multi-task training (ablation)Worldformer · JerichoWorld valid action prediction, overallF1: 22.42 -> 25.54Improved
add SOS loss (ablation)Worldformer · JerichoWorld valid action prediction, overallEM: 20.12 -> 23.22Improved
add SOS loss (ablation)Worldformer · JerichoWorld valid action prediction, overallF1: 21.28 -> 25.54Improved
add all components to Seq2Seq baseline (ablation)Worldformer · JerichoWorld valid action prediction, overallEM: 18.10 -> 23.22Improved

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