Interp-ML

Advances in the field of Biotechnology are increasingly reliant on the widespread use of large-scale biological data generated by various technologies, such as second, third, and fourth-generation sequencers, as well as gene expression analysis platforms, proteomics, and molecular simulation. The abundance of data creates opportunities for both researchers to deepen scientific knowledge and for Biotechnology companies. However, given the immense amount of information from these technologies and the high dimensionality of the data, conducting cutting-edge research in Biotechnology becomes largely unfeasible without the use of Data Science-based analysis, both in the generation and processing, analysis, and construction of hypotheses and models based on this data. This research project involves four major areas: Computer Science, Molecular Biology, Bioinformatics, and Computational Mathematics. The general objective of this project is the development of feature selection approaches in Machine Learning based on interpretability strategies such as Attention Mechanism and Layer-wise Relevance Propagation and multi-objective and self-adaptive optimization techniques.