Advanced Neural Networks for Multimodal Data Fusion in Interdisciplinary Research
K. Devika Rani Dhivya, S. Niresh Kumar, D. Rosy Salomi Victoria, S. Irin Sherly, G. Durgadevi · Advances in IT standards and standardization research (AISSR) book series/Advances in IT standards and standardization research series · 2024
The chapter will delve into deep neural network architectures developed for interdisciplinary research, specifically aimed at combining multimodal data. Modern research, spanning multiple domains like health and environmental science, necessitates the integration of diverse data sources like images, text, and sensor data. This discussion will explore advanced techniques like convolutional neural networks, recurrent neural networks, and transformer models that effectively combine these modalities. This paper emphasizes the development of robust multimodal systems, primarily focusing on data alignment and feature extraction prior to model training. The text discusses the challenges of heterogeneous data integration, its solutions, and the role of transfer learning in improving model performance. The chapter explores the potential of this approach in medical diagnostics, climate modeling, and social sciences, offering insights into future research directions and applications.