An Inductive Bias for Emergent Communication in a Continuous Setting

John Isak Fjellvang Villanger, Troels Arnfred Bojesen
Proceedings of the 5th Northern Lights Deep Learning Conference ({NLDL}), PMLR 233:235-243, 2024.

Abstract

We study emergent communication in a multi-agent reinforcement learning setting, where the agents solve cooperative tasks and have access to a communication channel. The communication channel may consist of either discrete symbols or continuous variables. We introduce an inductive bias to aid with the emergence of good communication protocols for continuous messages, and we look at the effect this type of inductive bias has for continuous and discrete messages in itself or when used in combination with reinforcement learning. We demonstrate that this type of inductive bias has a beneficial effect on the communication protocols learnt in two toy environments, Negotiation and Sequence Guess.

Cite this Paper

BibTeX
@InProceedings{pmlr-v233-villanger24a, title = {An Inductive Bias for Emergent Communication in a Continuous Setting}, author = {Villanger, John Isak Fjellvang and Bojesen, Troels Arnfred}, booktitle = {Proceedings of the 5th Northern Lights Deep Learning Conference ({NLDL})}, pages = {235--243}, year = {2024}, editor = {Lutchyn, Tetiana and Ramírez Rivera, Adín and Ricaud, Benjamin}, volume = {233}, series = {Proceedings of Machine Learning Research}, month = {09--11 Jan}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v233/villanger24a/villanger24a.pdf}, url = {https://proceedings.mlr.press/v233/villanger24a.html}, abstract = {We study emergent communication in a multi-agent reinforcement learning setting, where the agents solve cooperative tasks and have access to a communication channel. The communication channel may consist of either discrete symbols or continuous variables. We introduce an inductive bias to aid with the emergence of good communication protocols for continuous messages, and we look at the effect this type of inductive bias has for continuous and discrete messages in itself or when used in combination with reinforcement learning. We demonstrate that this type of inductive bias has a beneficial effect on the communication protocols learnt in two toy environments, Negotiation and Sequence Guess.} }
Endnote
%0 Conference Paper %T An Inductive Bias for Emergent Communication in a Continuous Setting %A John Isak Fjellvang Villanger %A Troels Arnfred Bojesen %B Proceedings of the 5th Northern Lights Deep Learning Conference ({NLDL}) %C Proceedings of Machine Learning Research %D 2024 %E Tetiana Lutchyn %E Adín Ramírez Rivera %E Benjamin Ricaud %F pmlr-v233-villanger24a %I PMLR %P 235--243 %U https://proceedings.mlr.press/v233/villanger24a.html %V 233 %X We study emergent communication in a multi-agent reinforcement learning setting, where the agents solve cooperative tasks and have access to a communication channel. The communication channel may consist of either discrete symbols or continuous variables. We introduce an inductive bias to aid with the emergence of good communication protocols for continuous messages, and we look at the effect this type of inductive bias has for continuous and discrete messages in itself or when used in combination with reinforcement learning. We demonstrate that this type of inductive bias has a beneficial effect on the communication protocols learnt in two toy environments, Negotiation and Sequence Guess.
APA
Villanger, J.I.F. & Bojesen, T.A.. (2024). An Inductive Bias for Emergent Communication in a Continuous Setting. Proceedings of the 5th Northern Lights Deep Learning Conference ({NLDL}), in Proceedings of Machine Learning Research 233:235-243 Available from https://proceedings.mlr.press/v233/villanger24a.html.

Related Material