Invertible Zero-Shot Recognition Flows

Document Type

Conference Proceeding

Source of Publication

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Publication Date

1-1-2020

Abstract

© 2020, Springer Nature Switzerland AG. Deep generative models have been successfully applied to Zero-Shot Learning (ZSL) recently. However, the underlying drawbacks of GANs and VAEs (e.g., the hardness of training with ZSL-oriented regularizers and the limited generation quality) hinder the existing generative ZSL models from fully bypassing the seen-unseen bias. To tackle the above limitations, for the first time, this work incorporates a new family of generative models (i.e., flow-based models) into ZSL. The proposed Invertible Zero-shot Flow (IZF) learns factorized data embeddings (i.e., the semantic factors and the non-semantic ones) with the forward pass of an invertible flow network, while the reverse pass generates data samples. This procedure theoretically extends conventional generative flows to a factorized conditional scheme. To explicitly solve the bias problem, our model enlarges the seen-unseen distributional discrepancy based on a negative sample-based distance measurement. Notably, IZF works flexibly with either a naive Bayesian classifier or a held-out trainable one for zero-shot recognition. Experiments on widely-adopted ZSL benchmarks demonstrate the significant performance gain of IZF over existing methods, in both classic and generalized settings.

ISBN

9783030585167

ISSN

0302-9743

Publisher

Springer International Publishing

Volume

12361 LNCS

First Page

614

Last Page

631

Disciplines

Computer Sciences

Keywords

Generative flows, Invertible networks, Zero-Shot Learning

Indexed in Scopus

no

Open Access

yes

Open Access Type

Green: A manuscript of this publication is openly available in a repository

Link to Full Text

Share

COinS