Latent Preserving Generative Adversarial Network for Imbalance Classification

Abstract

Many real-world classification problems have imbalanced frequency of class labels; a well-known issue known as the "class imbalance" problem. Classic classification algorithms tend to be biased towards the majority class, leaving the classifier vulnerable to misclassification of the minority class. While the literature is rich with methods to fix this problem, as the dimensionality of the problem increases, many of these methods do not scale-up and the cost of running them become prohibitive. In this paper, we present an end-to-end deep generative classifier. We propose a domain-constraint autoencoder to preserve the latent-space as prior for a generator, which is then used to play an adversarial game with two other deep networks, a discriminator and a classifier. Extensive experiments are carried out on three different multi-class imbalanced problems and a comparison with state-of-the-art methods. Experimental results confirmed the superiority of our method over popular algorithms in handling high-dimensional imbalanced classification problems. Our code is available on https://github.com/TanmDL/SLPPL-GAN

X-Fuzz: An Evolving and Interpretable Neuro-Fuzzy Learner for Data Streams

“…These gains over state-of-the-art techniques are attributed to X-Fuzz's adaptive architecture and lower dependence on user-defined parameters. The empirical results thus highlight X-Fuzz's capabilities in classifying challenging non-stationary and imbalanced data streams [50], [51], demonstrating its viability for real-world applications.…”

Ten years of generative adversarial nets (GANs): a survey of the state-of-the-art

“…A prevalent yet intricate issue encountered in pattern recognition is referred to as 'class imbalance' , signifying disparities in the frequencies of class labels [96]. To address this challenge, GANs can be used to generate synthetic data for the minority class of various imbalanced datasets as a method of intelligent oversampling [97]. Pioneering approaches such as balancing GAN [98] and classification enhancement GAN (CEGAN) [99] have been developed to restore balance in the distributions of imbalanced datasets and enhance the precision of the data-driven models.…”

Adaptive Terrain Perception and Decision-Making Systems for Agile Mobile Robots in Dynamic Search and Rescue Scenarios

“…Despite the fact that these robots are designed for autonomous navigation, their ability to traverse a variety of terrains efficiently is crucial to their overall performance [1][2][3]. Terrain surface classification methods for autonomous ground vehicles (AGVs) have seen significant advancements in recent years [4][5][6][7]. Identifying the type of terrain, a complex aspect of environmental perception, has a significant impact on both the motion [8] planning and the control [9] of mobile robotics.…”

“…There are two main ways to identify the type of ground a robot is moving over: active and passive methods [7,13]. Active methods use external sensors like cameras and radars to look at the ground, but this can be unreliable and computationally heavy, so often a combination of sensors is used, which can be expensive and complex [14].…”