Hierarchical Framework to Reduce Zero-Shot Learning Complexity

Hierarchical Framework to Reduce Zero-Shot Learning Complexity

  IJETT-book-cover           
  
© 2025 by IJETT Journal
Volume-73 Issue-7
Year of Publication : 2025
Author : Shaista Khanam, Poonam. N. Sonar
DOI : 10.14445/22315381/IJETT-V73I7P142

How to Cite?
Shaista Khanam, Poonam. N. Sonar, "Hierarchical Framework to Reduce Zero-Shot Learning Complexity," International Journal of Engineering Trends and Technology, vol. 73, no. 7, pp.543-561, 2025. Crossref, https://doi.org/10.14445/22315381/IJETT-V73I7P142

Abstract
Zero-Shot Learning (ZSL) is an emerging machine learning approach that enables the classification of images belonging to categories absent from the training data. By leveraging semantic information, ZSL facilitates classification with minimal or no training images. This paper presents a novel approach for ZSL employing a hierarchical framework, designed to enhance accuracy while significantly reducing complexity. The proposed framework employs a two-stage hierarchical classification structure with primary and secondary classifiers specific to each stage of the hierarchy. A Convolutional Neural Network (CNN) works as the principal component of the primary classifier, which uses a deep hierarchical clustering technique to classify images into two larger categories (Subclass-0 and Subclass-1). The secondary classifier integrates fastText for semantic feature extraction and ResNet-50 for visual feature extraction, enabling the classification of unseen (zero-shot) images. The usefulness of the proposed approach is validated on three standard datasets, viz. SUN, AWA2, and CUB. Accord-ing to experimental results, the hierarchical architecture achieves accuracy levels comparable to the best available methods while drastically reducing training complexity by almost 80%, training length by 25%, and testing time by 23%. The frame-work facilitates more effective learning by breaking the task up into smaller class subsets, which makes it ideal for large-scale ZSL applications.

Keywords
Hierarchical deep clustering, Hierarchical framework, Image classification, Model complexity, Zero-shotlearning.

References
[1] Xiaohong Sun, Jinan Gu, and Hongying Sun, “Research Progress of Zero-Shot Learning,” Applied Intelligence, vol. 51, no. 6, pp. 3600-3614, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[2] Wei Wang et al., “A Survey of Zero-Shot Learning: Settings, Methods, and Applications,” ACM Transactions on Intelligent Systems and Technology (TIST), vol. 10, no. 2, pp. 1-37, 2019.
[CrossRef] [Google Scholar] [Publisher Link]
[3] Genevieve Patterson, Xu Chen, and James Hays, Sun Attribute Database: Discovering, Annotating, and Recognizing Scene Attributes, SUN Attribute Dataset, 2011. [Online]. Available: https://cs.brown.edu/~gmpatter/sunattributes.html
[4] Animals with Attributes 2: A Free Dataset for Attribute-Based Classification and Zero-Shot Learning, Institute of Science and Technology Austria (ISTA), 2016. [Online]. Available :https://cvml.ista.ac.at/AwA2/ [5] Catherine Wah et al., “The Caltech-Ucsd Birds-200-2011 Dataset,” California Institute of Technology, 2011.
[Google Scholar] [Publisher Link]
[6] Christoph H. Lampert, Hannes Nickisch, and Stefan Harmeling, “Learning to Detect Unseen Object Classes BY Between-Class Attribute Transfer,” 2009 IEEE Conference on Computer Vision and Pattern Recognition, Miami, FL, USA, pp. 951-958, 2009.
[CrossRef] [Google Scholar] [Publisher Link]
[7] Christoph H. Lampert, Hannes Nickisch, and Stefan Harmeling, “Attribute-Based Classification for Zero-Shot Visual Object Categori-zation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 36, no. 3, pp. 453-465, 2013.
[CrossRef] [Google Scholar] [Publisher Link]
[8] Zeynep Akata et al., “Label-Embedding for Image Classification,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 38, no. 7, pp. 1425-1438, 2015.
[CrossRef] [Google Scholar] [Publisher Link]
[9] Xuesong Wang et al., “Zero-Shot Learning Based on Deep Weighted Attribute Prediction,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 50, no. 8, pp. 2948-2957, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[10] Andrea Frome et al., “Devise: A Deep Visual-Semantic Embedding Model,” Advances in Neural Information Processing Systems (NIPS 2013), vol. 26, pp. 2121-2129, 2013.
[Google Scholar] [Publisher Link]
[11] Ansari Shaista Khanam, and Poonam. N. Sonar, “Enhanced Zero-shot Learning using Deep Neural Network ResNet50,” 2023 4th Inter-national Conference for Emerging Technology (INCET), Belgaum, India, 2023, pp. 1-6, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[12] Yongqin Xian et al., “Feature Generating Networks for Zero-Shot Learning,” 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, pp. 5542-5551, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[13] Minwan Zhang et al., “Zero-Shot Learning with Joint Generative Adversarial Networks,” Electronics, vol. 12, no. 10, pp. 1-18, 2023.
[CrossRef] [Google Scholar] [Publisher Link]
[14] Varun Khare et al., “A Generative Framework for ZSL with Adversarial Domain Adaptation,” 2020 IEEE Winter Conference on Appli-cations of Computer Vision (WACV), Snowmass, CO, USA, pp. 3090-3099, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[15] Fadi Al Machot, Mohib Ullah, and Habib Ullah, “HFM: A Hybrid Feature Model Based on Conditional Auto Encoders for Zero-Shot Learning,” Journal of Imaging, vol. 8, no. 6, pp. 1-12, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[16] Rich Lee, and Ing-Yi Chen, “The Time Complexity Analysis of Neural Network Model Configurations,” 2020 International Conference on Mathematics and Computers in Science and Engineering (MACISE), Madrid, Spain, pp. 178-183, 2020.
[CrossRef] [Google Scholar] [Publisher Link]
[17] Pedro Freire et al., “Computational Complexity Evaluation of Neural Network Applications in Signal Processing,” arXiv Preprint, pp. 1-25, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[18] Bhoomi Shah, and Hetal Bhavsar, “Time Complexity in Deep Learning Models,” Procedia Computer Science, vol. 215, pp. 202-210, 2022.
[CrossRef] [Google Scholar] [Publisher Link]
[19] Saad Albawi, Tareq Abed Mohammed, and Saad Al-Zawi, “Understanding of a Convolutional Neural Network,” 2017 International Conference on Engineering and Technology (ICET), Antalya, Turkey, pp. 1-6, 2017.
[CrossRef] [Publisher Link]
[20] Sheldon Mascarenhas, and Mukul Agarwal, “A Comparison between VGG16, VGG19 and ResNet50 Architecture Frameworks for Image Classification,” 2021 International Conference on Disruptive Technologies for Multi-Disciplinary Research and Applications (CENTCON), Bengaluru, India, pp. 96-99, 2021.
[CrossRef] [Google Scholar] [Publisher Link]
[21] Armand Joulin et al., “Bag of Tricks for Efficient Text Classification,” Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistic, Valencia, Spain, vol. 2, pp. 427-431, 2017.
[Google Scholar] [Publisher Link]
[22] Lecture 41: Space and Computational Complexity in DNN, Deep Learning For Visual Computing - IITKGP, YouTube, 2018. [Online] Available at: https://www.youtube.com/@deeplearningforvisualcompu3823
[23] Kaiming He et al., “Deep Residual Learning for Image Recognition,” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, pp. 770-778, 2016.
[CrossRef] [Google Scholar] [Publisher Link]
[24] Elyor Kodirov, Tao Xiang, and Shaogang Gong, “Semantic Autoencoder for Zero-Shot Learning,” 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, pp. 4447-4456, 2017.
[CrossRef] [Google Scholar] [Publisher Link]
[25] Bernardino Romera-Paredes, and Philip Torr, “An Embarrassingly Simple Approach to Zero-Shot Learning,” Proceedings of the 32nd International Conference on Machine Learning, PMLR, vol. 37, pp. 2152-2161, 2015.
[Google Scholar] [Publisher Link]
[26] Yongqin Xian et al., “Latent Embeddings for Zero-Shot Classification,” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, pp. 69-77, 2016.
[CrossRef] [Google Scholar] [Publisher Link]
[27] Yongqin Xian et al., “Zero-Shot Learning-A Comprehensive Evaluation of the Good, the Bad and the Ugly,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 41, no. 9, pp. 2251-2265, 2018.
[CrossRef] [Google Scholar] [Publisher Link]
[28] Soravit Changpinyo et al., “Synthesized Classifiers for Zero-Shot Learning,” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, pp. 5327-5336, 2016.
[CrossRef] [Google Scholar] [Publisher Link]