GARL formalizes strategic prioritisation as a two-stage game, converts the resulting utilities into RL signals, and reports improved ranking performance that lets small open-source LLMs compete with closed-source models on legal-issue prioritisation.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
GARL: Game-Theoretic Reinforcement Learning for Multi-Agent Strategic Prioritisation
GARL formalizes strategic prioritisation as a two-stage game, converts the resulting utilities into RL signals, and reports improved ranking performance that lets small open-source LLMs compete with closed-source models on legal-issue prioritisation.