James Zou

speaker

371 appearances 1 recordings 1 series first heard Dec 2024 last heard Dec 2024

James Zou’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

No recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.

Appearances

newest first · ▶ plays the moment

"The Cognitive Revolution" · Breakthroughs in AI for Biology: AI Lab Groups & Protein Model Interpretability with Prof James Zou · 18 Dec 2024 podcast

And how it works is that you can actually give it like a new protein sequence where you can modify some protein sequence by introducing some mutations.

And the ESM model then will produce some sort of basically like a log-likely score, right?

Similar to perplexity in large language models that will tell you how likely is this modified sequence.

And the general intuition is that you know if the this modified sequence are stable, right, if they're actually functional proteins, then the log likelihood sc would be relatively good, right, relatively high score.

But there's actually a lot of fine art in how exactly to compute those log likelihood scores and exactly which ones to use.

So the agent here actually developed like a new formula for computing a specific way of this likelihood ratio score, which is actually different from previous publications where people had used DSM.

So that's actually quite an interesting, let's say, local modification or innovation by the agents.

Yeah, so I think

A lot of that really comes down to because the AI agents decided that they want to make a sequence of mutations, right?

So they'll start with some

wild type nan nanobody, which are basically existing nanobodies people have tried before, but those nanobodies do not really work for the recent variants of COVID.

So the agents decided to take those as a starting point and then make some sequential modifications by introducing mutations on top of that nanobody.

I think a lot of the f uh the fine art i if you will in computing the likelihood score depends on how do you measure the changes in the likelihood.

Do you say compare to the original wild type, uh does the likelihood increase after every mutation you make?

But then what happens if you make the second mutation?

Does it then you do you wanna compare that with the first mutation or do you compare that with the wild type?

So that's where there's even debates in the literature about where and how should want to approach that with different approaches and the agents actually developed like a a specific formula for computing those likelihood ratios essentially that uh in you know after the fact in my view is actually pretty reasonable for doing.

Yeah.

So it's actually a pretty sophisticated workflow that AI agents developed.

And you know, I I wish I could claim credit for that, but it's really developed by the agents, right?

Showing 81–100 of 371 · page 5 of 19