Your Classifier is Secretly an Energy Based Model and You Should Treat it Like One

TL;DR

This paper reinterprets standard classifiers as energy-based models for joint distribution, enabling improved calibration, robustness, and generative capabilities while maintaining training efficiency.

Contribution

It introduces a novel framework that treats classifiers as energy-based models for joint distribution, unifying discriminative and generative learning.

Findings

Improves calibration, robustness, and out-of-distribution detection.

Generates high-quality samples comparable to GANs.

Achieves state-of-the-art performance in both discriminative and generative tasks.

Abstract

We propose to reinterpret a standard discriminative classifier of p(y|x) as an energy based model for the joint distribution p(x,y). In this setting, the standard class probabilities can be easily computed as well as unnormalized values of p(x) and p(x|y). Within this framework, standard discriminative architectures may beused and the model can also be trained on unlabeled data. We demonstrate that energy based training of the joint distribution improves calibration, robustness, andout-of-distribution detection while also enabling our models to generate samplesrivaling the quality of recent GAN approaches. We improve upon recently proposed techniques for scaling up the training of energy based models and presentan approach which adds little overhead compared to standard classification training. Our approach is the first to achieve performance rivaling the state-of-the-artin both generative and discriminative learning within one hybrid model.