Filters








2,377 Hits in 7.0 sec

Combining Deep Universal Features, Semantic Attributes, and Hierarchical Classification for Zero-Shot Learning [article]

Jared Markowitz, Aurora C. Schmidt, Philippe M. Burlina, I-Jeng Wang
2017 arXiv   pre-print
We address zero-shot (ZS) learning, building upon prior work in hierarchical classification by combining it with approaches based on semantic attribute estimation.  ...  For both non-novel and novel image classes we compare multiple formulations of the problem, starting with deep universal features in each case.  ...  Acknowledgments We thank the authors of [ ] for providing the class to attribute mapping of their data set.  ... 
arXiv:1712.03151v1 fatcat:maz4josz7baehms6ixbscowf2e

Recent Advances in Zero-shot Recognition [article]

Yanwei Fu, Tao Xiang, Yu-Gang Jiang, Xiangyang Xue, Leonid Sigal, and Shaogang Gong
2017 arXiv   pre-print
One approach to scaling up the recognition is to develop models capable of recognizing unseen categories without any training instances, or zero-shot recognition/ learning.  ...  This article provides a comprehensive review of existing zero-shot recognition techniques covering various aspects ranging from representations of models, and from datasets and evaluation settings.  ...  Yanwei Fu is supported by The Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning.  ... 
arXiv:1710.04837v1 fatcat:u3mp6dgj2rgqrarjm4dcywegmy

A Survey on Visual Transfer Learning using Knowledge Graphs [article]

Sebastian Monka, Lavdim Halilaj, Achim Rettinger
2022 arXiv   pre-print
We explain the notion of feature extractor, while specifically referring to visual and semantic features.  ...  Recent approaches of computer vision utilize deep learning methods as they perform quite well if training and testing domains follow the same underlying data distribution.  ...  [92] propose SAE a semantic autoencoder for zero-shot learning.  ... 
arXiv:2201.11794v1 fatcat:tapql5h4j5dvrnxjkaxek2cquu

Integrating Propositional and Relational Label Side Information for Hierarchical Zero-Shot Image Classification [article]

Colin Samplawski, Heesung Kwon, Erik Learned-Miller, Benjamin M. Marlin
2019 arXiv   pre-print
Zero-shot learning (ZSL) is one of the most extreme forms of learning from scarce labeled data.  ...  In this paper, we present a new ZSL framework that leverages both label attribute side information and a semantic label hierarchy.  ...  For example, the deep visual-semantic embedding (De-ViSE) model [1] learns a linear compatibility function between image feature vectors and label attribute vectors a y .  ... 
arXiv:1902.05492v1 fatcat:q2e2bushafdxpnjgsn2xqlq7da

A survey on visual transfer learning using knowledge graphs

Sebastian Monka, Lavdim Halilaj, Achim Rettinger, Mehwish Alam, Davide Buscaldi, Michael Cochez, Francesco Osborne, Diego Reforgiato Recupero, Harald Sack
2022 Semantic Web Journal  
We explain the notion of feature extractor, while specifically referring to visual and semantic features.  ...  Recent approaches of CV utilize deep learning (DL) methods as they perform quite well if training and testing domains follow the same underlying data distribution.  ...  Acknowledgements This publication was created as part of the research project "KI Delta Learning" (project number: 19A19013D) funded by the Federal Ministry for Economic Affairs and Energy (BMWi) on the  ... 
doi:10.3233/sw-212959 fatcat:f4s43if3nbcxxfvrbtpdrrs2ry

Zero-Shot Action Recognition in Videos: A Survey [article]

Valter Estevam, Helio Pedrini, David Menotti
2020 arXiv   pre-print
Therefore, we present a survey of the methods that comprise techniques to perform visual feature extraction and semantic feature extraction as well to learn the mapping between these features considering  ...  Zero-Shot Action Recognition has attracted attention in the last years and many approaches have been proposed for recognition of objects, events and actions in images and videos.  ...  Method Used in appoaches A Annotated Wang and Chen [90] Mishra et al. [  ... 
arXiv:1909.06423v2 fatcat:w5eh7wjdmnaktnbsqczsdmhane

Learning Robust Visual-Semantic Embeddings [article]

Yao-Hung Hubert Tsai and Liang-Kang Huang and Ruslan Salakhutdinov
2017 arXiv   pre-print
We evaluate our method on Animals with Attributes and Caltech-UCSD Birds 200-2011 dataset with a wide range of applications, including zero and few-shot image recognition and retrieval, from inductive  ...  The proposed method combines representation learning models (i.e., auto-encoders) together with cross-domain learning criteria (i.e., Maximum Mean Discrepancy loss) to learn joint embeddings for semantic  ...  For example, Socher et al. [41] used deep architectures [13] to learn representations for both images and text, and then used a Bayesian framework to perform classification. Norouzi et al.  ... 
arXiv:1703.05908v2 fatcat:bmvr3bbvavepbg7k7gwgks7gne

Tell me what you see: A zero-shot action recognition method based on natural language descriptions [article]

Valter Estevam and Rayson Laroca and David Menotti and Helio Pedrini
2021 arXiv   pre-print
Recently, several approaches have explored the detection and classification of objects in videos to perform Zero-Shot Action Recognition with remarkable results.  ...  The projection onto this space is straightforward for both types of information, visual and semantic, because they are sentences, enabling the classification with nearest neighbour rule in this shared  ...  Alternative semantic represen- tations for zero-shot human action recognition, in: Machine Learning and Knowledge Discovery in Databases, pp. 87–102. [52] Wang, Q., Chen, K., 2017b.  ... 
arXiv:2112.09976v1 fatcat:5bvci2dyjnbyjbvo7qagnuzbpy

Semantic Autoencoder for Zero-Shot Learning [article]

Elyor Kodirov, Tao Xiang, Shaogang Gong
2017 arXiv   pre-print
Existing zero-shot learning (ZSL) models typically learn a projection function from a feature space to a semantic embedding space (e.g.~attribute space).  ...  However, such a projection function is only concerned with predicting the training seen class semantic representation (e.g.~attribute prediction) or classification.  ...  Sparse feature learning for [26] D. Jayaraman and K. Grauman. Zero-shot recognition with deep belief networks. In Advances in neural information pro- unreliable attributes.  ... 
arXiv:1704.08345v1 fatcat:3tlzperd2nc3vmdmmigqavdonm

Class Representative Learning for Zero-shot Learning Using Purely Visual Data

Mayanka Chandrashekar, Yugyung Lee
2021 SN Computer Science  
Recent works, including zero-shot learning (ZSL) and generalized zero-short learning (G-ZSL), have attempted to overcome the apparent gap through transfer learning.  ...  However, most of these works are required to build a model using visual input with associated data like semantics, attributes, and textual information.  ...  [1] , zero-shot learning was categorized, based on feature requirements, into engineered semantic space for attribute, lexical, and keyword information and learned semantic space for label-embedding,  ... 
doi:10.1007/s42979-021-00648-y fatcat:urjq5mbr3jcmxjb6zn36u6m6ty

Large-scale zero-shot learning in the wild: Classifying zoological illustrations

Lise Stork, Andreas Weber, Jaap van den Herik, Aske Plaat, Fons Verbeek, Katherine Wolstencroft
2021 Ecological Informatics  
We train a prototypical network for zero-shot classification, and introduce fused prototypes (FP) and hierarchical prototype loss (HPL) to optimise the model.  ...  (BHL), can be used to share knowledge between classes for zero-shot learning.  ...  Nguyen and Sophia Ananiadou (National Centre for Text Mining) for providing the text embeddings from the Biodiversity Heritage Library.  ... 
doi:10.1016/j.ecoinf.2021.101222 fatcat:jxcutjpsvrfmdhxsr4egqnqjie

A Survey on Neural-symbolic Systems [article]

Dongran Yu, Bo Yang, Dayou Liu, Hui Wang
2021 arXiv   pre-print
Combining the fast computation ability of neural systems and the powerful expression ability of symbolic systems, neural-symbolic systems can perform effective learning and reasoning in multi-domain tasks  ...  This paper surveys the latest research in neural-symbolic systems along four dimensions: the necessity of combination, technical challenges, methods, and applications.  ...  manual rulemaking and realizing end-to-end knowledge graph reasoning. (3) Few-shot learning and zero-shot learning The main challenge for few-shot learning and zero-shot learning is the shortage of training  ... 
arXiv:2111.08164v1 fatcat:bc33afiitnb73bmjtrfbdgkwpy

Zero-Shot Learning – A Comprehensive Evaluation of the Good, the Bad and the Ugly [article]

Yongqin Xian, Christoph H. Lampert, Bernt Schiele, Zeynep Akata
2020 arXiv   pre-print
Moreover, we propose a new zero-shot learning dataset, the Animals with Attributes 2 (AWA2) dataset which we make publicly available both in terms of image features and the images themselves.  ...  First, given the fact that there is no agreed upon zero-shot learning benchmark, we first define a new benchmark by unifying both the evaluation protocols and data splits of publicly available datasets  ...  We point out that extracting image features via a pre-trained deep neural network Y tr Fig. 1 : Zero-shot learning (ZSL) vs generalized zero-shot learning (GZSL): At training time, for both cases the  ... 
arXiv:1707.00600v4 fatcat:62lsg45xdvco5kz72bprebctbm

Zero-Shot Learning - A Comprehensive Evaluation of the Good, the Bad and the Ugly

Yongqin Xian, Christoph H. Lampert, Bernt Schiele, Zeynep Akata
2018 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Moreover, we propose a new zero-shot learning dataset, the Animals with Attributes 2 (AWA2) dataset which we make publicly available both in terms of image features and the images themselves.  ...  First, given the fact that there is no agreed upon zero-shot learning benchmark, we first define a new benchmark by unifying both the evaluation protocols and data splits of publicly available datasets  ...  We point out that extracting image features via a pre-trained deep neural network Y tr Fig. 1 : Zero-shot learning (ZSL) vs generalized zero-shot learning (GZSL): At training time, for both cases the  ... 
doi:10.1109/tpami.2018.2857768 pmid:30028691 fatcat:24a6dwragrel5bnyjfvbrmabi4

Zero-Shot Object Recognition System Based on Topic Model

Wai Lam Hoo, Chee Seng Chan
2015 IEEE Transactions on Human-Machine Systems  
We propose a novel zero-shot learning strategy that utilizes the topic model and hierarchical class concept.  ...  In this paper, we study the problem of object recognition where the training samples are missing during the classifier learning stage, a task also known as zero-shot learning.  ...  Similar work that employed the hierarchical class strategy in zero-shot learning paradigm includes Rohrbach et al. [4] and Frome et al. [5] .  ... 
doi:10.1109/thms.2014.2358649 fatcat:byamfnuto5apviihl5rwzhxbgu
« Previous Showing results 1 — 15 out of 2,377 results