Anticipating Social Consequences with Distributional World Models

arXiv cs.AIOctober 23, 2026en

arXiv:2610.30657v1 Announce Type: new Abstract: Social world models predict how interactions develop after an action, helping agents negotiate, cooperate and mediate. In these situations, one action can lead to several plausible futures, hence predicting only one future can overlook consequential alternatives, while exploring them through repeated simulations increases inference cost. We propose DSWM, a Distributional Social World Model that predicts alternative responses, their probabilities and immediate and final outcome values in one output per candidate action. We train a 4B model on repeated simulations using supervised fine-tuning and GRPO. At inference time, the actor uses these poss

This is a short summary published by AI Global Wire. The full article is owned and hosted by arXiv cs.AI — open it there to read it in full.

• Forskning
• Agenter