The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) PaccMann^RL: Designing Anticancer Drugs with Reinforcement Learning w/ Jannis Born - #341
Jan 23, 2020
In this insightful discussion, Jannis Born, a PhD student at ETH Zurich and IBM Research Zurich, dives into his groundbreaking work with 'PaccMann^RL.' He explains how his background in computational neuroscience informs anticancer drug discovery and the role of reinforcement learning in tailoring treatments. Jannis also explores the complexities of RNA sequencing, gene expression, and innovative drug prediction methods using deep learning. Listeners gain a glimpse into the future of personalized medicine and the integration of AI in revolutionizing cancer treatment.
AI Snips
Chapters
Transcript
Episode notes
Jannis's Background
- Jannis Born's background is in cognitive science and computational neuroscience, focusing on brain research.
- He transitioned to applying machine learning in computational systems biology, specifically cancer research.
Brain Science to Cancer Research
- Cognitive science and computational neuroscience offer top-down and bottom-up approaches to understanding the brain, respectively.
- Jannis's desire to apply machine learning to human biology led him to cancer drug modeling.
Pac-Mann and Cancer Diversity
- Pac-Mann, a multimodal attention-based neural network, predicts anticancer compound sensitivity.
- It treats drug sensitivity as a property of the drug-tumor cell pair due to cancer's diversity.