Autonomous Empirical Research Group

Mission

Our research group strives to enhance and accelerate scientific discovery by automating each step of the empirical research process, from constructing a scientific hypothesis to conducting novel experiments. To this end, we devise an open-source programming language for autonomous empirical research. We seek to co-create and share this system with empirical scientists across disciplines who desire to advance their research without sacrificing scientific standards, such as reproducibility and transparency. Finally, we seek to accelerate open science efforts by developing methods for the automated documentation and dissemination of steps taken in the empirical research process.

Approach

Autonomous Empirical Research (AER) consists of two artificial agents that compete and together implement a form of adversarial curiosity: (a) an autonomous theorist whose objective is to construct computational models from existing data by linking experiment conditions to dependent measures, and (b) an autonomous experimentalist whose objective is to invalidate the theorist by deriving predictions from its models and designing novel experiments to collect unexplained data.