Program Summary
The Understanding Group Biases (UGB) program seeks to develop and prove out capabilities that can radically enhance the scale, speed, and scope of automated, ethnographic-like methods for capturing group biases and cultural models from increasingly available large digital datasets. DARPA hypothesizes that there may be new opportunities for overcoming current methodological trade-offs between capturing qualitative “thick” data or capturing quantitative “big” data, in part, by turning a common bug of machine-learning (a tendency to pick up non-obvious and implicit biases from the datasets on which it is trained) into a useful feature. Hence, UGB aims to develop systems that can identify and characterize these biases at new speeds and scales in order to provide deeper insight into the diversity and complexity of human cultural models, as well as lead to better understanding of when, why, and how groups often interpret the same world differently.