AI Paper Atlas

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

Emergent Communication under Varying Sizes and Connectivities

Jooyeon Kim, Alice Oh

NeurIPS 2021 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Emergent communication in referential games · 4 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
optimize link structure with CMA-ES instead of random linkscommunication graph · Hard constraint, N=16, 3 outgoing linkssuccess rate: 0.578 -> 0.894Improved
optimize links instead of neighbor-to-neighbor linkscommunication graph · Hard constraint, N=16, 3 outgoing linkssuccess rate: 0.454 -> 0.894Improved
optimize links instead of all-to-one linkscommunication graph · Hard constraint, N=16, 3 outgoing linkssuccess rate: 0.384 -> 0.894Improved
restrict to 3 outgoing links instead of all-to-allcommunication graph · Hard constraint, N=16, 3 outgoing linkssuccess rate: 0.950 -> 0.894Mixed
add any communication links versus nonecommunication graph · Hard constraint, N=16, 3 outgoing linkssuccess rate: 0.236 -> 0.894Improved
optimize with distance penalty instead of random links at similar penaltycommunication graph · Soft constraint, N=16, k=0.0012success rate: 0.513 -> 0.877Improved
optimize with distance penalty instead of neighbor-to-neighborcommunication graph · Soft constraint, N=16, k=0.0012success rate: 0.454 -> 0.877Improved
optimize instead of 1-to-All graphcommunication graph · Soft constraint, N=16, k=0.0012success rate: 0.251 -> 0.877Improved
increase group size N up to hundredsemergent language · bipartite referential gamemessage agreement (cosine similarity)Hurt
increase group size N up to hundredsemergent language · bipartite referential gamecommunication success rateNo clear effect
increase max communication range j in neighbor-to-neighbor graphemergent language · N=64success rate between non-neighborsImproved
increase max communication range j from 1 to 16emergent language · N=64dialect structure by locality: j = 1 -> j = 16Mixed

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