CICERO: intent-grounded dialogue with strategic planning

What changed

Integrated human-data-trained intent-conditioned dialogue, human-regularized strategic planning and message filtering, achieving strong human-level performance in the reported online Diplomacy blitz league.

Scientific scope

Integrated human-data-trained intent-conditioned dialogue, human-regularized strategic planning and message filtering, achieving strong human-level performance in the reported online Diplomacy blitz league. Forty-game five-minute-negotiation-turn evaluation, not unrestricted long-form Diplomacy or general negotiation competence. Human demonstrations and dialogue data are used; it is not learning solely from self-play. Dialogue grounding/filtering is imperfect and the planning model does not model long-term relationship effects; no never-deceives or universal trustworthiness guarantee is implied. R2C2/BART and earlier regularized search retain predecessor credit.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Forty-game five-minute-negotiation-turn evaluation, not unrestricted long-form Diplomacy or general negotiation competence. Human demonstrations and dialogue data are used; it is not learning solely from self-play. Dialogue grounding/filtering is imperfect and the planning model does not model long-term relationship effects; no never-deceives or universal trustworthiness guarantee is implied. R2C2/BART and earlier regularized search retain predecessor credit.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Human-level play in the game of Diplomacy by combining language models with strategic reasoning

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction