Abstract: A method for determining the validity of a computational model using a blockchain and zero knowledge principles includes: storing, in a memory of a first computing system, a computational model; receiving, by a receiver of the first computing system, a blockchain data value from one block of a plurality of blocks comprising a blockchain, wherein the blockchain data value includes a data set; receiving, by the receiver of the first computing system, an expected accuracy value; applying, by a processor of the first computing system, the data set to the computational model to generate a result value; and determining, by the processor of the first computing system, a validity measurement for the computational model based on a comparison of the generated result value and the expected accuracy value.

Abstract: Embodiments provide methods and systems for reducing bias in an artificial intelligence model. A method includes computing, by a processor, a reward value based at least in part on a similarity between model predictions from a pre-trained model and agent predictions from a Reinforcement Learning (RL) agent. The method includes performing each step of one or more steps of a rule of a plurality of rules. The rule is assigned a weight and the rule includes a protected attribute, a cumulative statistic value type, and a comparison threshold. The method includes sending a cumulative reward value generated using the reward value and each weighted punishment value computed based at least in part on applying each rule of the plurality of rules to the RL agent. The RL agent learns to biases from the agent predictions while maintaining similarity with model predictions by maximizing the cumulative reward value.