AI with AI: Artificial Intelligence with Andy Ilachinski · CNA

Journey to the Cause of Reason

July 9, 2021·37 min·3 clips
Researchers used AI to analyze 45,000 samples and discover a universal host response to viruses like COVID-19.
This episode of AI with AI, hosted by Andy Ilachinski, explores recent developments in artificial intelligence research and its applications. The host examines several specific studies and projects, focusing on AI's role in analyzing complex biological and operational data. Topics include COVID-19 research, autonomous aerial combat, and machine learning for surveillance. A study from UC San Diego School of Medicine, published in eBioMedicine, used AI to analyze gene expression data from 45,000 samples across viral infections like Ebola, Zika, influenza, and COVID-19. The research aimed to identify a common host response pattern to viral pandemics by examining terabytes of data on which genes activate during infection. Another segment revisits DARPA's work on AI for aerial dogfights, following previous discussions on autonomous combat simulations. The episode also covers machine learning applications for analyzing footage from CCTV surveillance cameras. A portion is dedicated to the concept of causality in machine learning models. The host notes the UC San Diego paper's use of AI is described as "fairly rudimentary" despite the complex, high-dimensional data involved. The research process involved sifting through massive gene expression datasets to find invariant patterns across different viruses. The discussion on AI in dogfights connects to ongoing military research into autonomous aerial combat systems. Using machine learning with CCTV footage highlights AI's growing role in mass surveillance and data analysis. Exploring causality touches on a fundamental challenge in making AI systems' decisions interpretable and grounded in reason. The tone is educational and conversational, with the host guiding listeners through technical research in an accessible manner. The style is analytical, breaking down scientific papers and project updates into their core components. Listeners interested in the intersection of AI with medical research and defense technology would find this episode engaging. Those seeking highly speculative futurism or deep technical tutorials might find the survey-style coverage too broad.
Listen to the show on