English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

While it is known that communication facilitates cooperation in multi-agentsettings, it is unclear how to design artificial agents that can learn toeffectively and efficiently communicate with each other. Much research oncommunication emergence uses reinforcement learning (RL) and exploresunsituated communication in one-step referential tasks -- the tasks are nottemporally interactive and lack time pressures typically present in naturalcommunication. In these settings, agents may successfully learn to communicate,but they do not learn to exchange information concisely -- they tend towardsover-communication and an inefficient encoding. Here, we explore situatedcommunication in a multi-step task, where the acting agent has to forgo anenvironmental action to communicate. Thus, we impose an opportunity cost oncommunication and mimic the real-world pressure of passing time. We comparecommunication emergence under this pressure against learning to communicatewith a cost on articulation effort, implemented as a per-message penalty (fixedand progressively increasing). We find that while all tested pressures candisincentivise over-communication, situated communication does it mosteffectively and, unlike the cost on effort, does not negatively impactemergence. Implementing an opportunity cost on communication in a temporallyextended environment is a step towards embodiment, and might be a pre-conditionfor incentivising efficient, human-like communication.

Over-communicate no more: Situated RL agents learn concise communication protocols