Dr Dmitri Roussinov - Neural architectures for bioinformatics, information systems and data analytics (venia legendi)

Klipi teostus: Ahti Saar 22.03.2019 179 vaatamist Arvutiteadus


Abstract: Recently, artificial neural networks (deep, recurrent, convolutional, etc.) have resulted in major breakthroughs in processing human languages, computer vision, game playing, and many other domains. I will present an overview of artificial neural network architectures and discuss their important properties, limitations, and historical evolution. I will use my own experience from published studies and works by other researchers to illustrate how specific architectures help particular applications such as those in bioinformatics, information systems and data analytics.