ACL Findings2023
So many design choices: Improving and interpreting neural agent communication in signaling games
Timothée Bernard, Timothee Mickus
Abstract
Emergent language games are experimental protocols designed to model how communication may arise among a group of agents. In this paper, we focus on how to improve performances of neural agents playing a signaling game: a sender is exposed to an image and generates a sequence of symbols that is transmitted to a receiver, which uses it to distinguish between two images, one that is semantically related to the original image, and one that is not. We consider multiple design choices, such as pretraining the visual components of the agents, introducing regularization terms, how to sample training items from the dataset, and we study how these different choices impact the behavior and performances of the agents. To that end, we introduce a number of automatic metrics to measure the properties of the emergent languages. We find that some implementation choices are always beneficial, and that the information that is conveyed by the agents' messages is shaped not only by the game, but also by the overall design of the agents as well as seemingly unrelated implementation choices.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
1 top-tier paper has cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- Emergent Communication: Generalization and Overfitting in Lewis Games cites Compositional languages emerge in a neural iterated learning model
- Emergent Communication: Generalization and Overfitting in Lewis Games cites Compositionality and Generalization In Emergent Languages
- Emergent Communication: Generalization and Overfitting in Lewis Games cites Emergent Communication of Generalizations
- Learning to Draw: Emergent Communication through Sketching cites Compositional languages emerge in a neural iterated learning model
- Learning to Draw: Emergent Communication through Sketching cites Compositionality and Generalization In Emergent Languages
- One-to-Many Communication and Compositionality in Emergent Communication cites Compositional languages emerge in a neural iterated learning model
- One-to-Many Communication and Compositionality in Emergent Communication cites Compositionality and Generalization In Emergent Languages
- One-to-Many Communication and Compositionality in Emergent Communication cites Learning to Draw: Emergent Communication through Sketching
- One-to-Many Communication and Compositionality in Emergent Communication cites Emergent Communication of Generalizations
- Emergent Communication of Generalizations cites Compositional languages emerge in a neural iterated learning model
- Emergent Communication of Generalizations cites Compositionality and Generalization In Emergent Languages
Cited by top-tier papers1
Builds on6
- Compositional languages emerge in a neural iterated learning modelYi Ren, Shangmin Guo, Matthieu Labeau, Shay B. Cohen et al.ICLR 2020 · 111 citations
- Emergent Communication of GeneralizationsJesse Mu, Noah D. GoodmanNeurIPS 2021 · 60 citations
- Emergent Communication: Generalization and Overfitting in Lewis GamesMathieu Rita, Corentin Tallec, Paul Michel, Jean-Bastien Grill et al.NeurIPS 2022 · 41 citations
- Compositionality and Generalization In Emergent LanguagesRahma Chaabouni, Eugene Kharitonov, Diane Bouchacourt, Emmanuel Dupoux et al.ACL 2020 · 40 citations
- Learning to Draw: Emergent Communication through SketchingDaniela Mihai, Jonathon S. HareNeurIPS 2021 · 34 citations
Related papers
- Searching for the Most Human-like Emergent LanguageBrendon Boldt, David R. MortensenEMNLP 2025
- On the Word Boundaries of Emergent Languages Based on Harris's Articulation SchemeRyo Ueda, Taiga Ishii, Yusuke MiyaoICLR 2023
- Learning Multi-Object Positional Relationships via Emergent CommunicationYicheng Feng, Boshi An, Zongqing LuAAAI 2024 · 4 citations
- Emergent Graphical Conventions in a Visual Communication GameShuwen Qiu, Sirui Xie, Lifeng Fan, Tao Gao et al.NeurIPS 2022 · 18 citations
- On the interaction between supervision and self-play in emergent communicationRyan Lowe, Abhinav Gupta, Jakob N. Foerster, Douwe Kiela et al.ICLR 2020 · 30 citations
