Bridging the Projection Gap: Overcoming Projection Bias Through Parameterized Distance Learning

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhang, Chong, Jin, Mingyu, Yu, Qinkai, Xue, Haochen, Gowda, Shreyank N, Jin, Xiaobo
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910612895825920
author Zhang, Chong
Jin, Mingyu
Yu, Qinkai
Xue, Haochen
Gowda, Shreyank N
Jin, Xiaobo
author_facet Zhang, Chong
Jin, Mingyu
Yu, Qinkai
Xue, Haochen
Gowda, Shreyank N
Jin, Xiaobo
contents Generalized zero-shot learning (GZSL) aims to recognize samples from both seen and unseen classes using only seen class samples for training. However, GZSL methods are prone to bias towards seen classes during inference due to the projection function being learned from seen classes. Most methods focus on learning an accurate projection, but bias in the projection is inevitable. We address this projection bias by proposing to learn a parameterized Mahalanobis distance metric for robust inference. Our key insight is that the distance computation during inference is critical, even with a biased projection. We make two main contributions - (1) We extend the VAEGAN (Variational Autoencoder \& Generative Adversarial Networks) architecture with two branches to separately output the projection of samples from seen and unseen classes, enabling more robust distance learning. (2) We introduce a novel loss function to optimize the Mahalanobis distance representation and reduce projection bias. Extensive experiments on four datasets show that our approach outperforms state-of-the-art GZSL techniques with improvements of up to 3.5 \% on the harmonic mean metric.
format Preprint
id arxiv_https___arxiv_org_abs_2309_01390
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Bridging the Projection Gap: Overcoming Projection Bias Through Parameterized Distance Learning
Zhang, Chong
Jin, Mingyu
Yu, Qinkai
Xue, Haochen
Gowda, Shreyank N
Jin, Xiaobo
Computer Vision and Pattern Recognition
Artificial Intelligence
Generalized zero-shot learning (GZSL) aims to recognize samples from both seen and unseen classes using only seen class samples for training. However, GZSL methods are prone to bias towards seen classes during inference due to the projection function being learned from seen classes. Most methods focus on learning an accurate projection, but bias in the projection is inevitable. We address this projection bias by proposing to learn a parameterized Mahalanobis distance metric for robust inference. Our key insight is that the distance computation during inference is critical, even with a biased projection. We make two main contributions - (1) We extend the VAEGAN (Variational Autoencoder \& Generative Adversarial Networks) architecture with two branches to separately output the projection of samples from seen and unseen classes, enabling more robust distance learning. (2) We introduce a novel loss function to optimize the Mahalanobis distance representation and reduce projection bias. Extensive experiments on four datasets show that our approach outperforms state-of-the-art GZSL techniques with improvements of up to 3.5 \% on the harmonic mean metric.
title Bridging the Projection Gap: Overcoming Projection Bias Through Parameterized Distance Learning
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2309.01390