AI in Math Seminar: Reading the Genome as Language/Measuring Intelligence

Two 20 minute talks in Toll Physics 2211

Speaker 1: Chirag Adwani (UMD) - Genomic sequences can be viewed not just as strings of nucleotides, but as objects with structure that can be learned from data. In this talk, we will explore how ideas from language modeling are being applied to genomes, with splice-site prediction as a concrete example. 

Speaker 2: Todd Rowland (UMD) The second talk is about measuring intelligence with ideas from AIT. Measuring intelligence for general AI is not simple. One approach is based on Algorithmic Information Theory, a set of ideas from the early days of AI, introduced by Ray Solomonoff (1960s) and developed for modern AI by Marcus Hutter (Google/Deep Mind). We will try to make it simple.