I will summarize our work on in-database solvers for relational machine learning and artificial intelligence. We implement a declarative query language that offers support for model (or instance) finding. The capability is used to express predictive and prescriptive analytics. The presentation gives an overview of the platform and the language. In particular, it focuses on the use of algebraic (e.g. semi-rings) and combinatorial structure (e.g. query width) for asymptotic improvements of gradient descent based solvers used in most in most machine learning methods.
Mr. Molham Aref is the Chief Executive Officer of Relational AI. Mr. Aref has more than 25 years of experience in developing and implementing enterprise-grade analytic, predictive, optimization and simulation solutions for the demand chain, supply chain, and revenue management across various industries. Relational AI combines the latest advances in Artificial Intelligence with a understanding of business processes to develop solutions that shape better decisions, improve agility, and reduce risk. Prior to Relational AI, he was co-founder and CEO of LogicBlox where he led the company from inception through a successful sale to Infor. Previously, he was CEO of Optimi (acquired by Ericsson), a leader in wireless network simulation and optimization, and co-founder of Brickstream (renamed Nomi and then acquired by FLIR), a leading provider of computer-vision-based behavior intelligence solutions.
Faculty Hosts: Ben Moseley (Tepper) / Andy Pavlo (CSD)