15 - Hidden Chinese Restaurant Game: Learning from Actions

Summary

The effectiveness of a decision may be uncertain due to the unknown system state. This uncertainty can be eliminated through learning from information sources, such as user-generated content or revealed actions. Nevertheless, user-generated content could be untrustworthy, since other agents may maliciously create misleading content for their selfish interests. Passively revealed actions are potentially more trustworthy and also easier to gather through simple observation. In this chapter, we introduce a game-theoretic framework – the hidden Chinese restaurant game (H-CRG) – to utilize the passively revealed actions in social learning process. We design grand information extraction, a novel Bayesian belief extraction process, to extract beliefs on hidden information directly from observed actions. The optimal policy is then analyzed in both centralized and game-theoretic approaches. We demonstrate how the H-CRG can be applied to the channel access problem in cognitive radio networks. The simulation results show that the equilibrium strategy derived in the H-CRG provides greater expected utilities for new users and maintains reasonably high social welfare.