Academics

  /  

Courses

  /  

Descriptions

COMP_ENG 395, 495: AI for Science and Business

Description

For thousands of years of human history, scientific discoveries were performed using observations, experiments, and mathematics. Advent of computers over last several decades enabled simulations based on mathematical models of nature, universe, and any physical phenomena for scientific discoveries. Massive amount of data produced by large-scale computer simulations and instruments resulted in the next phase of scientific discovery, also known as the “fourth paradigm”. Over the last decade+, due to advances in supercomputing power and in Deep Learning algorithms that learn from massive datasets, Artificial Intelligence (AI) has enabled new frontiers and paradigms that have the potential to accelerate scientific discoveries at a pace never imagined.

The main objectives of this class are to discuss, explore and understand how modern AI including Deep Learning, Transformers, Large Language Models (and Foundational Models), and data mining techniques can exploit massive amounts of data to accelerate scientific discoveries. The class will discuss how AI has the potential to accelerate discoveries in various scientific applications including but not limited to materials discovery, climate and weather science, life-science and health-care applications, and cosmology and others In addition, the class will AI’s role in transforming many business applications and models including finance and investment, marketing and advertising, computer vision based applications, and business of AI itself including business models in different verticals. The class will include guest lectures by leaders and experts in AI and various science domains. The class will involve reading papers, presentations, and team project. The evaluation will entail summarization of insights from papers on a particular domain of science or foundation technologies, once per two weeks on average. All students will be responsible for participation in discussion, and 1-2 groups will present on that topic. Concurrently, each team will work on a quarter long project and flexibility will be provided to choose the domain for the project and the type of project based on the team’s interest, including an option to come up with a relevant topic/science/business domain. Given that the field of AI and its impact on scientific discoveries and business is evolving at unprecedented speed, new topics may be introduced based on development in the field in real-time.

The class is expected to be taken mainly by graduate students and seniors. Although the class doesn’t have any specific prerequisite, familiarity with basics of data science, ML and data mining is recommended.

Class: CompE 395/495 CS 396/496.

There are two sections of the class on Wednesdays (3:00pm-5:40pm) and Thursdays (2:00pm-4:40pm). A student can take either one of the classes.