PAC Confidence Predictions for Deep Neural Network Classifiers
By Sangdon Park, Shuo Li, Osbert Bastani, Insup Lee. 2020
27 citations recorded by the source corpus.
A key challenge for deploying deep neural networks (DNNs) in safety critical settings is the need to provide rigorous ways to quantify their uncertainty. In this paper, we propose a novel algorithm for constructing predicted classification confidences for DNNs that comes with provable correctness guarantees. Our approach uses Clopper-Pearson confidence intervals for the Binomial distribution in conjunction with the histogram binning approach to calibrated prediction. In addition, we demonstrate how our predicted confidences can be used to enable downstream guarantees in two settings: (i) fast DNN inference, where we demonstrate how to compose a fast but inaccurate DNN with an accurate but slow DNN in a rigorous way to improve performance without sacrificing accuracy, and (ii) safe planning, where we guarantee safety when using a DNN to predict whether a given action is safe based on visual observations. In our experiments, we demonstrate that our approach can be used to provide guarantees for state-of-the-art DNNs.
Subfields: Artificial Intelligence, Model Architectures, Statistical Learning Theory
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Tommi S. Jaakkola
- Regina Barzilay
- Percy Liang
- Mohit Bansal
- Luke S. Zettlemoyer
- Tal Schuster
- Caiming Xiong
- Ruslan Salakhutdinov
In Model Architectures
- Tommi S. Jaakkola
- Regina Barzilay
- Jitendra Malik
- Michael I. Jordan
- Adam Fisch
- Tal Schuster
- Christoph Feichtenhofer
- Yoshua Bengio
In Statistical Learning Theory
- Regina Barzilay
- Adam Fisch
- Tal Schuster
- Tommi S. Jaakkola
- Michael I. Jordan
- Jitendra Malik
- Stephen Bates
- Anastasios Nikolas Angelopoulos
Citations in the corpus
- Consistent Accelerated Inference via Confident Adaptive Transformers (2021, methodology): “Additionally, many methods exist for bounding the true error of a classifier (Langford, 2005; Park et al., 2021), but do not give end-users opportunities to control it.”
- Distribution-Free, Risk-Controlling Prediction Sets (2021, background): “Recently, tolerance regions have been instantiated to form prediction sets for deep learning models [6, 7].”
- A Gentle Introduction to Conformal Prediction and Distribution-Free Uncertainty Quantification (2021, background): “Recent work also addresses the issue of calibrating a scalar notion of uncertainty to have probabilistic meaning via histogram binning [104, 105]—this is like a rigorous version of Platt scaling or isotonic regression.”