AI Paper Atlas

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

Emergent Communication through Negotiation

Kris Cao, Angeliki Lazaridou, Marc Lanctot, Joel Z Leibo, Karl Tuyls, Stephen Clark

ICLR 2018 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Emergent communication in referential games · 43 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
Use binding, game-grounded proposals instead of linguistic cheap talkSelf-interested negotiating agents · Negotiation environment, random terminationfraction of joint reward: 0.75 ± 0.22 -> 0.87 ± 0.11Improved
Use linguistic cheap talk instead of the proposal channelProsocial negotiating agents · Negotiation environment, random terminationfraction of joint reward: 0.93 ± 0.10 -> 0.99 ± 0.02Improved
Add the proposal channel alongside the linguistic channelProsocial negotiating agents · Negotiation environment, random terminationfraction of joint reward: 0.99 ± 0.02 -> 0.92 ± 0.11Hurt
Use the proposal channel instead of no communicationProsocial negotiating agents · Negotiation environment, random terminationfraction of joint reward: 0.95 ± 0.11 -> 0.93 ± 0.10Hurt
Allow the full 10 turns rather than random terminationProsocial agents using the proposal channel · Negotiation environmentfraction of joint reward: 0.93 ± 0.10 -> 0.96 ± 0.07Improved
Reveal opponent IDs to the fixed selfish agentFixed selfish agent A, 1 prosocial opponent, proposal channel · Society negotiationpersonal reward: 0.81 ± 0.17 -> 0.96 ± 0.06Improved
Reveal opponent IDs to the fixed selfish agentFixed selfish agent A, 1 prosocial opponent, linguistic channel · Society negotiationpersonal reward: 0.97 ± 0.05 -> 0.97 ± 0.10No clear effect
Reveal opponent IDs to a fixed prosocial agentFixed prosocial agent A, 5 prosocial opponents · Society negotiation, linguistic channeljoint reward: 0.98* ± 0.06 -> 0.97 ± 0.06Hurt
Use linguistic cheap talk instead of no communicationProsocial agents · Negotiation environment, random terminationfraction of joint reward: 0.95 ± 0.11 -> 0.99 ± 0.02Improved
Use linguistic cheap talk instead of no communicationSelf-interested agents · Negotiation environment, random terminationfraction of joint reward: 0.77 ± 0.22 -> 0.75 ± 0.22No clear effect
Use linguistic messages to convey hidden utilitiesProsocial-prosocial agents · Linguistic message probingagent B value accuracy (%): 20 -> 81Improved

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