Optimal Network Architecture and Inference Dependencies for Efficient Training of Deep Neural Networks in Bioinformatics
Jaganathan Logeshwaran, Durgesh Kumar Srivastava, Manoj Kumar Pal, S. Dhanasekaran, Anuradha Sagar Nigade, Keshav Kaushik · 2024
Optimal Network Architecture and Inference Dependencies for Deep Learning in Bioinformatics are all about how to efficiently design and train deep neural networks for bioinformatics use-cases referring to training or inference dependencies. It attempts to optimize the network architecture and reduce the entanglement between different parts of the network, which can train deep neural networks in bioinformatics more eventfully and faster. The performance of the deep learning models and their accuracy can be enhanced overall. It is beneficial for the study, analysis, and interpretation of biological data by cautiously opting for the network architecture and minimizing the dependencies. Moreover, such optimization can prevent overfitting as well as save the computational resources and time consumed during training, making it more feasible to upscale it for deep learning on big data in bioinformatics.