Thesis Title: Measuring Intrinsic Quality of Human Decisions
Uncertainty and risk often contribute or even inhibit our ability to make optimal decisions. Although various requirements and frameworks are prevalent to aid in producing optimal decisions or to ensure acceptable quality of decision, any standardized measurement techniques to quantify the quality of these decisions is not developed. With the advancement of AI and computation technologies, it is possible to generate optimal solutions to decision problems if enough time is provided.
In our proposed model, we have suggested an adaptation of techniques used in the measurement theory to measure or quantify the quality of decisions with respect to (sub )optimal decisions produced by AI agents of supreme strength. We have tried to correlate the depth of cognition by humans to depth of search and evaluation performed by an AI agent and use this information to rank the quality of the decisions made. The current setting of our model uses data from real chess tournaments and evaluation from chess engines but the model can be applied to any fields of decision making where artificial agents can be found to generate the optimal choice.