Enhancing Data Rather Than Hallucinating

1min Snip

00:00

Play full episode

Summary

Transcript

Episode notes

Concerns exist regarding the use of synthetic data for training AI models, as generating data without real-world basis risks reinforcing existing biases. A more effective approach involves taking actual measurements and augmenting them to explore unmeasured or previously unavailable data modalities. This method enhances the quality of training data, especially in fields like biology, by using AI to create additional valuable data from already collected measurements, thereby improving model performance and insights without relying solely on artificial data generation.

How can scientists leverage AI and machine learning to more effectively research, develop, and deliver new drugs? This week, Reid and Aria talk drug development with renowned computer scientist and executive Dr. Daphne Koller, whose company, insitro, uses machine learning to improve the quality and speed of drug discovery. She addresses several ways AI and ML are already being used to redefine diseases and create better therapeutic interventions. Plus, she shares her experience transitioning from academia to industry.

Read the transcript of this episode here: https://www.possible.fm/podcasts/dkoller

For more info on the podcast and transcripts of all the episodes, visit https://www.possible.fm/podcast/

Topics:

00:01: Cold open

00:43: Hellos and intros

3:20: Early involvement in AI

5:00: Overview of drug discovery and its evolution over the last five years

7:50: AI and machine learning impact on biology and drug development

9:09: Pi aside defining AlphaFold

11:37: Areas of acceleration and opportunity in AI and drug discovery

17:21: Synthetic data

21:21: AI implications on therapeutic hypotheses

23:54: Personalized vs. precision medicine

27:04: Closing the data feedback loop

30:31: Exciting announcements

34:24: Privacy and data

36:41: GPT-4 story

38:48: Stuart Russell’s comments on the applications of Koller’s thesis

43:36: AI as a moving target

47:24: How can academia and industry positively impact society and humanity

51:58: insitro’s culture

54:18: - Rapid-fire questions

Select mentions:

Not the End of the World: How We Can Be the First Generation to build a Sustainable Planet by Hannah Ritchie

Stuart Russell

Possible is an award-winning podcast that sketches out the brightest version of the future—and what it will take to get there. Most of all, it asks: what if, in the future, everything breaks humanity's way? Tune in for grounded and speculative takes on how technology—and, in particular, AI—is inspiring change and transforming the future. Hosted by Reid Hoffman and Aria Finger, each episode features an interview with an ambitious builder or deep thinker on a topic, from art to geopolitics and from healthcare to education. These conversations also showcase another kind of guest: AI. Whether it's Inflection’s Pi, OpenAI’s ChatGPT or other AI tools, each episode will use AI to enhance and advance our discussion about what humanity could possibly get right if we leverage technology—and our collective effort—effectively.