Preserving Semantic Relations for Zero-Shot Learning

03/08/2018
by   Yashas Annadani, et al.
0

Zero-shot learning has gained popularity due to its potential to scale recognition models without requiring additional training data. This is usually achieved by associating categories with their semantic information like attributes. However, we believe that the potential offered by this paradigm is not yet fully exploited. In this work, we propose to utilize the structure of the space spanned by the attributes using a set of relations. We devise objective functions to preserve these relations in the embedding space, thereby inducing semanticity to the embedding space. Through extensive experimental evaluation on five benchmark datasets, we demonstrate that inducing semanticity to the embedding space is beneficial for zero-shot learning. The proposed approach outperforms the state-of-the-art on the standard zero-shot setting as well as the more realistic generalized zero-shot setting. We also demonstrate how the proposed approach can be useful for making approximate semantic inferences about an image belonging to a category for which attribute information is not available.

READ FULL TEXT
research
06/30/2019

Visual Space Optimization for Zero-shot Learning

Zero-shot learning, which aims to recognize new categories that are not ...
research
07/27/2016

Improving Semantic Embedding Consistency by Metric Learning for Zero-Shot Classification

This paper addresses the task of zero-shot image classification. The key...
research
02/10/2020

From Anchor Generation to Distribution Alignment: Learning a Discriminative Embedding Space for Zero-Shot Recognition

In zero-shot learning (ZSL), the samples to be classified are usually pr...
research
06/11/2019

From Fully Supervised to Zero Shot Settings for Twitter Hashtag Recommendation

We propose a comprehensive end-to-end pipeline for Twitter hashtags reco...
research
06/13/2019

Semantics to Space(S2S): Embedding semantics into spatial space for zero-shot verb-object query inferencing

We present a novel deep zero-shot learning (ZSL) model for inferencing h...
research
05/03/2018

Multi-Context Label Embedding

Label embedding plays an important role in zero-shot learning. Side info...
research
02/05/2021

Metric Embedding Sub-discrimination Study

Deep metric learning is a technique used in a variety of discriminative ...

Please sign up or login with your details

Forgot password? Click here to reset