Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Emergent Communication in a Multi-Modal, Multi-Step Referential Game
Katrina Evtimova, Andrew Drozdov, Douwe Kiela, Kyunghyun Cho
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Increase message dimensionality to 32 (ablation) | Adaptive multimodal referential game · Mammal out-of-domain test | accuracy@7: up to 45% | Improved |
| Vary message dimensionality (ablation) | Adaptive multimodal referential game · In-domain test | accuracy@6 | No clear effect |
| Increase message dimensionality (ablation) | Adaptive multimodal referential game · Mammal out-of-domain test | accuracy@7 | Improved |
| Vary message dimensionality (ablation) | Adaptive multimodal referential game · Transfer test | accuracy@K | Mixed |
| Add attention to the receiver (ablation) | Adaptive multimodal referential game · Insect transfer test | accuracy@K: 16.9% up to 27.4% | Improved |
| Add attention mechanism (ablation) | Adaptive multimodal referential game · In-domain test | accuracy@6 | No clear effect |
| Train the receiver while keeping the sender fixed (ablation) | Adaptive multimodal referential game · Training performance | overall performance | Hurt |
| Train the sender and receiver jointly rather than keeping the sender fixed (ablation) | Adaptive multimodal referential game · Training performance | overall performance | Improved |
| Use an adaptive conversation length | Multimodal referential game · In-domain test, across classes | Mean conversation length versus classifier F1: $-0.81$ | Mixed |
| Continue the conversation and exchange more information | Adaptive multimodal referential game · In-domain test | Prediction entropy | Improved |
| Continue the conversation and exchange more information | Adaptive multimodal referential game · Out-of-domain test | Prediction entropy | Mixed |
Closest papers by meaning
- Semantics and Spatiality of Emergent Communication (NeurIPS 2024 · similarity 0.90)
- Learning Multi-Object Positional Relationships via Emergent Communication (AAAI 2024 · similarity 0.89)
- Lewis's Signaling Game as beta-VAE For Natural Word Lengths and Segments (ICLR 2024 · similarity 0.88)
- Speaking Your Language: Spatial Relationships in Interpretable Emergent Communication (NeurIPS 2024 · similarity 0.88)
- Cost-Effective Communication: An Auction-based Method for Language Agent Interaction (AAAI 2026 · similarity 0.87)
- Bridging semantics and pragmatics in information-theoretic emergent communication (NeurIPS 2024 · similarity 0.87)
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.
