Humans use unique conventions beyond AI to succeed at Hanabi
What and Where in Identifying and Quantifying Conventions in Humans and AI Agents
Human-Computer InteractionArtificial IntelligenceMultiagent Systems
Summary
Success in cooperative games like Hanabi depends on shared, unspoken rules called conventions that help players understand each other's actions better than the basic game rules predict. The authors studied how humans and AI agents use these conventions and found humans rely on a special type of hint that AI agents do not learn. They also showed that these conventions vary depending on what players already know and when in the game they are used. Importantly, the study turned conventions from just a guess based on score drops into measurable behaviors that can be tested by modifying agents.
What this means in practice
- •For ai developers: Improve cooperative AI agents by incorporating human-like unconventional hinting strategies in Hanabi and similar teamwork settings.
- •For game designers: Design better cooperative games by understanding when and where players use conventions to communicate beyond explicit game rules.
Authors
Makoto Fukushima, Tim Puphal
Abstract
Cooperative success rests in part on implicit communication. Partners converge on conventions: regularities that give an action meaning beyond its direct effect. Yet conventions are usually observed only indirectly: a score drop when partners are swapped shows that a convention existed, but not what it was or where it was used. In the card game Hanabi, the exact probability that a play fails can be computed from the hints received and the cards in view. Players who use conventions succeed more often than this prediction allows; the resulting margin is the convention gap. Off-belief learning provides a hierarchy of agents (levels 1-5) whose first level uses no conventions by construction and whose later levels use progressively more. For humans and these agents on one scale, we ask what the conventions are and where they are used. On what: in two independently collected corpora of human games, humans show a larger gap than every agent level. On the conventions both populations use, however, humans do no better than level 2. The human advantage is concentrated in a situation that no agent level learns to use: hints that point to a card other than the ones they touch. On where: conventions depend on what the player already knows about the card, not on how recently it was hinted, and they are used most in the mid-game; humans use them earlier and later in the game than the agents. In interventions on the agents, blocking a single high-gap situation costs little, because meaning reroutes to other situations, but blocking several together costs far more than the sum of the single costs. Thus, the convention gap turns conventions from an inferred cause of score loss into a measurable, comparable, and causally testable object.