Details

Autor(en) / Beteiligte

• Nadia
• Gandotra, Ekta
• Siddiqui, Mohd Faizan

Titel

Chapter 1 - Deep learning ensembles in translational bioinformatics

Ist Teil von

• Deep Learning Applications in Translational Bioinformatics, 2024, p.1-15

Ort / Verlag

Elsevier Inc

Erscheinungsjahr

2024

Link zum Volltext

Beschreibungen/Notizen

• In the era of big data, transformation of biomedical big data into valuable knowledge has been one of the most important challenges in bioinformatics. Since the early 2000s, deep learning has developed rapidly, and it now performs at the forefront of technology in several areas. As a result, deep learning applications in bioinformatics have received attention from both academia and industry. Deep learning ensembles are predicted to play a bigger role in the area and be incorporated into many analysis pipelines. Deep learning models and ensemble approaches are increasingly being used in bioinformatics research due to their exceptional adaptability and flexibility. In bioinformatics applications, these two machine learning techniques have typically been viewed primarily as independent procedures. To achieve synergistic advances in model accuracy, stability, and reproducibility, two machine learning techniques are recently emerging called ensemble deep learning. This has sparked a new wave of study and application. In this chapter, we discuss recent major advancements in ensemble deep learning and examine how their contributions have aided a variety of bioinformatics studies, from systems biology to simple sequence analysis. Although there are many different and varied ways that ensemble deep learning is used in bioinformatics, we highlight and explore the typical problems, opportunities, and challenges in the context of bioinformatics research. This chapter intends to help researchers working in the field of machine learning and bioinformatics to foster future research and development in ensemble deep learning and inspire new bioinformatics applications that cannot be attained by using traditional machine learning models.