PaccMannRL: Designing Anticancer Drugs with Reinforcement Learning w/ Jannis Born

800 800 The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)

Today we’re joined by Jannis Born, Ph.D. student at ETH & IBM Research Zurich.

We caught up with Jannis a few weeks back at NeurIPS, where he was presenting his poster on “PaccMannRL: Designing anticancer drugs from transcriptomic data via reinforcement learning,” a framework built to accelerate new anticancer drug discovery. In our conversation, Jannis details how his background in cognitive science and computational neuroscience applies to his current ML research. We then transition into a conversation about PaccMann, or Prediction of Anticancer Compound Sensitivity with Multimodal Attention-based Neural Networks, discussing how reinforcement learning fits into the goal of cancer drug discovery, how deep learning has changed this research, interesting observations made during the training of their DRL learner, and of course, Jannis offers us a step-by-step walkthrough of how the framework works to predict the sensitivity of cancer drugs on a cell and then discover new anticancer drugs.

Community Update

We’re excited to announce a couple of new study group offerings in conjunction with our ever-expanding array of TWIML Community programs. One is focused on causality and the other on enterprise AI workflows. Click here to learn more about these programs!

About Jannis

Mentioned in the Interview

  • PaccMannRL Project Page: Deep learning for cancer precision medicine — Designing anticancer drugs from omics profiles
  • Paper: PaccMannRL: Designing anticancer drugs from transcriptomic data via reinforcement learning
  • Paper: PaccMann: Prediction of anticancer compound sensitivity with multi-modal attention-based neural networks
  • McCulloch-Pitts Neuron
  • Rosenblatt’s perceptron, the first modern neural network
  • Simplified molecular-input line-entry system
  • Paper: Deep Sets
  • Check it out

    “More On That Later” by Lee Rosevere licensed under CC By 4.0

    Leave a Reply

    Your email address will not be published.