Generative Max-Mahalanobis Classifiers for Image Classification, Generation and More

• URL: http://arxiv.org/abs/2101.00122v3
• Date: Fri, 2 Apr 2021 22:30:49 GMT
• Title: Generative Max-Mahalanobis Classifiers for Image Classification, Generation and More
• Authors: Xiulong Yang, Hui Ye, Yang Ye, Xiang Li, Shihao Ji
• Abstract summary: Max-Mahalanobis (MMC) can be trained discriminatively, generatively, or jointly for image classification and generation. We show that our Generative MMC (GMMC) can be trained discriminatively, generatively, or jointly for image classification and generation.
• Score: 6.89001867562902
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Joint Energy-based Model (JEM) of Grathwohl et al. shows that a standard softmax classifier can be reinterpreted as an energy-based model (EBM) for the joint distribution p(x,y); the resulting model can be optimized to improve calibration, robustness, and out-of-distribution detection, while generating samples rivaling the quality of recent GAN-based approaches. However, the softmax classifier that JEM exploits is inherently discriminative and its latent feature space is not well formulated as probabilistic distributions, which may hinder its potential for image generation and incur training instability. We hypothesize that generative classifiers, such as Linear Discriminant Analysis (LDA), might be more suitable for image generation since generative classifiers model the data generation process explicitly. This paper therefore investigates an LDA classifier for image classification and generation. In particular, the Max-Mahalanobis Classifier (MMC), a special case of LDA, fits our goal very well. We show that our Generative MMC (GMMC) can be trained discriminatively, generatively, or jointly for image classification and generation. Extensive experiments on multiple datasets show that GMMC achieves state-of-the-art discriminative and generative performances, while outperforming JEM in calibration, adversarial robustness, and out-of-distribution detection by a significant margin. Our source code is available at https://github.com/sndnyang/GMMC.

Related papers

• Exploring Beyond Logits: Hierarchical Dynamic Labeling Based on Embeddings for Semi-Supervised Classification [49.09505771145326]
  We propose a Hierarchical Dynamic Labeling (HDL) algorithm that does not depend on model predictions and utilizes image embeddings to generate sample labels. Our approach has the potential to change the paradigm of pseudo-label generation in semi-supervised learning.
  arXiv  Detail & Related papers  (2024-04-26T06:00:27Z)
• Latent Space Energy-based Model for Fine-grained Open Set Recognition [46.0388856095674]
  Fine-grained open-set recognition (FineOSR) aims to recognize images belonging to classes with subtle appearance differences while rejecting images of unknown classes. As a type of generative model, energy-based models (EBM) are the potential for hybrid modeling of generative and discriminative tasks. In this paper, we explore the low-dimensional latent space with energy-based prior distribution for OSR in a fine-grained visual world.
  arXiv  Detail & Related papers  (2023-09-19T16:00:09Z)
• One-Line-of-Code Data Mollification Improves Optimization of Likelihood-based Generative Models [17.47235124122244]
  Likelihood-based GMs are attractive due to the possibility to generate new data by a single model evaluation. They typically achieve lower sample quality compared to state-of-the-art score-based diffusion models (DMs)
  arXiv  Detail & Related papers  (2023-05-30T09:58:47Z)
• A Hybrid of Generative and Discriminative Models Based on the Gaussian-coupled Softmax Layer [5.33024001730262]
  We propose a method to train a hybrid of discriminative and generative models in a single neural network. We demonstrate that the proposed hybrid model can be applied to semi-supervised learning and confidence calibration.
  arXiv  Detail & Related papers  (2023-05-10T05:48:22Z)
• Traditional Classification Neural Networks are Good Generators: They are Competitive with DDPMs and GANs [104.72108627191041]
  We show that conventional neural network classifiers can generate high-quality images comparable to state-of-the-art generative models. We propose a mask-based reconstruction module to make semantic gradients-aware to synthesize plausible images. We show that our method is also applicable to text-to-image generation by regarding image-text foundation models.
  arXiv  Detail & Related papers  (2022-11-27T11:25:35Z)
• GMMSeg: Gaussian Mixture based Generative Semantic Segmentation Models [74.0430727476634]
  We propose a new family of segmentation models that rely on a dense generative classifier for the joint distribution p(pixel feature,class) With a variety of segmentation architectures and backbones, GMMSeg outperforms the discriminative counterparts on closed-set datasets. GMMSeg even performs well on open-world datasets.
  arXiv  Detail & Related papers  (2022-10-05T05:20:49Z)
• Towards Bridging the Performance Gaps of Joint Energy-based Models [1.933681537640272]
  Joint Energy-based Model (JEM) achieves high classification accuracy and image generation quality simultaneously. We introduce a variety of training techniques to bridge the accuracy gap and the generation quality gap of JEM. Our SADA-JEM achieves state-of-the-art performances and outperforms JEM in image classification, image generation, calibration, out-of-distribution detection and adversarial robustness by a notable margin.
  arXiv  Detail & Related papers  (2022-09-16T14:19:48Z)
• Continual Learning with Fully Probabilistic Models [70.3497683558609]
  We present an approach for continual learning based on fully probabilistic (or generative) models of machine learning. We propose a pseudo-rehearsal approach using a Gaussian Mixture Model (GMM) instance for both generator and classifier functionalities. We show that GMR achieves state-of-the-art performance on common class-incremental learning problems at very competitive time and memory complexity.
  arXiv  Detail & Related papers  (2021-04-19T12:26:26Z)
• Cauchy-Schwarz Regularized Autoencoder [68.80569889599434]
  Variational autoencoders (VAE) are a powerful and widely-used class of generative models. We introduce a new constrained objective based on the Cauchy-Schwarz divergence, which can be computed analytically for GMMs. Our objective improves upon variational auto-encoding models in density estimation, unsupervised clustering, semi-supervised learning, and face analysis.
  arXiv  Detail & Related papers  (2021-01-06T17:36:26Z)
• Shaping Deep Feature Space towards Gaussian Mixture for Visual Classification [74.48695037007306]
  We propose a Gaussian mixture (GM) loss function for deep neural networks for visual classification. With a classification margin and a likelihood regularization, the GM loss facilitates both high classification performance and accurate modeling of the feature distribution. The proposed model can be implemented easily and efficiently without using extra trainable parameters.
  arXiv  Detail & Related papers  (2020-11-18T03:32:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.