Special Topic

Description

Algorithms in Data Science

Scalability, the subject of study in this course, has recently become one of the central topics in Data Science (DS), Machine Learning (ML), and Artificial Intelligence (AI). For example, artificial neural network based natural language models have shown their true potential only when their size has reached billions, or even hundreds of billions, of parameters; to achieve human level accuracy at the protein structure prediction problem a deep learning algorithm requires dozens of powerful (A100) GPUs to be trained on.

In this course we will study what it means for an algorithm to scale and how to design and implement scalable algorithms. Starting from basic methods of algorithm design and analysis, we will consider what breakthroughs in algorithms made modern AI, ML, and DS possible, and what algorithmic challenges remain to be solved to enable future generations of scalable data-intensive computational methods.

Learning outcomes of this course include basic skills in algorithm design and analysis in the context of modern deep learning (DL). For example, students will be able to design, implement, and analyze a transformer-based DL model. All algorithms will be illustrated with their applications in biological data science, which will equip the learner with transferable skills.