Our research focuses on developing the understanding and tools that enable real-world deployment of machine learning models in a reliable and responsible way. In particular, much of our thinking revolves around pinpointing the exact role of data in ML-driven decision making and how to tackle the challenge posed by the omnipresent distribution shift. We also are interested in thinking through the societal and policy implications of the usage of modern ML tools.
People
Alumni
Research
Software
Attribute (or cite) statements generated by LLMs back to in-context information with ContextCite!
Compute data attribution scores at scale in PyTorch with just a few lines of code using TRAK!
Train models at a fraction of the cost with accelerated data loading! FFCV is a drop-in data loading system that dramatically increases data throughput in model training.
3DB: a framework for debugging models using 3D rendering. Reproduce your favorite robustness analyses or design your own analyses/experiments in just a few lines of code!
