PhD Student Talk, David Čechák, CEITEC Masaryk University

On Friday, May 6, at 4:30pm, our MSc programme will have the pleasure of physically hosting Mr. David Čechák, PhD Student at CEITEC Masaryk University, who will give a talk entitled: “A report on fine-tuning transformers for genomic tasks”. This event will take place in hybrid format, mainly physically in Lecture Theater Δ22 at the School of Sciences, but also via Zoom for our friends outside Thessaloniki at https://authgr.zoom.us/j/96979945626?pwd=ZTYvSERaaGM0aWJETCtFWkhlK00zUT09

Abstract: Transformers have achieved SOTA performance in numerous natural language processing tasks. We review the current state of Transformers usage in genomics, introduce a collection of benchmark datasets for the classification of genomic sequences, and compare the performance of several model architectures on those benchmarks. In particular, we explore the effect of pre-training on a large DNA corpus vs training from scratch. The results presented here can be used to identify functional elements in human and other genomes.

Bio: David Čechák  is a PhD student focused on deep learning in genomics, with a diverse background in artificial intelligence, image processing, and front-end web development. His latest work was on transfer learning in genomics, the composition of genomic benchmark datasets, and teaching deep learning to biologists. He graduated in artificial intelligence from Masaryk University in the Czech Republic and is currently pursuing a PhD in bioinformatics at the Central European Institute of Technology, Brno, Czech republic.