Toll Free Helpline (India): 1800 1234 070
Rest of World: +91-9810852116
Free Publication Certificate
Submit Manuscript
Journal code(s)
- E-ISSN Number: 2277-7695
- P-ISSN Number: 2349-8242
- CODEN Code: PIHNBQ
- ZDB-Number: 2663038-2
- IC Journal ID: 7725
Main Menu
Special Issue
Copyright Form
Identifier
Vol. 8, Special Issue 4 (2019)
The role of semi-supervised learning in harnessing unlabeled data for model training
Author(s):
Dr. Sunil Kumar Mishra, AM Tripathi and Mukesh Chauhan
Dr. Sunil Kumar Mishra, AM Tripathi and Mukesh Chauhan
Abstract:
The ever-growing availability of vast amounts of unlabeled data poses a significant opportunity for enhancing machine learning models. Semi-supervised learning (SSL) has emerged as a powerful paradigm to harness the potential of unlabeled data by combining it with limited labeled samples. This review paper provides a comprehensive overview of the role of semi-supervised learning in model training, focusing on its applications, methodologies, and advancements.
The ever-growing availability of vast amounts of unlabeled data poses a significant opportunity for enhancing machine learning models. Semi-supervised learning (SSL) has emerged as a powerful paradigm to harness the potential of unlabeled data by combining it with limited labeled samples. This review paper provides a comprehensive overview of the role of semi-supervised learning in model training, focusing on its applications, methodologies, and advancements.
The paper begins by elucidating the challenges associated with traditional supervised learning paradigms, emphasizing the scarcity of labeled data in many real-world scenarios. It then delves into the principles of semi-supervised learning, elucidating how it enables models to learn from both labeled and unlabeled data, thereby capitalizing on the abundance of unannotated information.
A critical analysis of various SSL techniques is presented, ranging from traditional methods such as self-training and co-training to more recent advancements like consistency regularization and generative adversarial networks. The review also explores the effectiveness of SSL across different domains, including computer vision, natural language processing, and speech recognition, showcasing its versatility and widespread applicability.
Furthermore, the paper discusses challenges and open research questions in the field of semi-supervised learning, addressing issues such as model robustness, scalability, and generalization to diverse datasets. The evolving landscape of SSL is highlighted, including recent breakthroughs and emerging trends that shape the future of utilizing unlabeled data for model training.
Pages: 01-04 | 147 Views 65 Downloads
How to cite this article:
Dr. Sunil Kumar Mishra, AM Tripathi and Mukesh Chauhan. The role of semi-supervised learning in harnessing unlabeled data for model training. The Pharma Innovation Journal. 2019; 8(4S): 01-04. DOI: 10.22271/tpi.2019.v8.i4Sa.25255
Journal code(s)
- E-ISSN Number: 2277-7695
- P-ISSN Number: 2349-8242
- CODEN Code: PIHNBQ
- ZDB-Number: 2663038-2
- IC Journal ID: 7725
Main Menu
Special Issue
Copyright Form
Identifier
The Pharma Innovation Journal
