Definitions, methods, and applications in interpretable machine learning
University of California, Berkeley · Allen Institute for Brain Science · +2 more institutions
Abstract
Machine-learning models have demonstrated great success in learning complex patterns that enable them to make predictions about unobserved data. In addition to using models for prediction, the ability to interpret what a model has learned is receiving an increasing amount of attention. However, this increased focus has led to considerable confusion about the notion of interpretability. In particular, it is unclear how the wide array of proposed interpretation methods are related and what common concepts can be used to evaluate them. We aim to address these concerns by defining interpretability in the context of machine learning and introducing the predictive, descriptive, relevant (PDR) framework for…
Citation impact
- FWCI
- 116.65
- Percentile
- 100%
- References
- 137
Authors
5- WJWilliam J. Murdoch
University of California, Berkeley
- CSChandan Singh
University of California, Berkeley
- KKKarl Kumbier
University of California, Berkeley
- RAReza Abbasi-Asl
Allen Institute for Brain Science, University of California, San Francisco, Allen Institute, University of California, Berkeley
- BYBin YuCorresponding
University of California, Berkeley
Topics & keywords
- Interpretability
- Computer science
- Artificial intelligence
- Categorization
- Machine learning
- Context (archaeology)
- Interpretation (philosophy)
- Modularity (biology)
- Quality Education