“We represented molecules as a vector of the probability of creator… 1800 element vector… top-5 accuracy of 60%… it does as well as if you use chemical formulas”
♫
Said on Latent Space: The AI Engineer Podcast
“We got a top five accuracy of 60%… we represent the molecules… as a vector of the probability of creator”
Ian Quigley, LeashBio · Latent Space: The AI Engineer Podcast · Lab-in-the-Loop at Compute Speed - Ian Quigley, Leash Bio
Play from 0:45
0:45
Audio streams from the publisher’s own feed. Verbatim does not host or copy it.
What was described
Created a 1,800-element vector representing the probability that each of 1,815 chemists authored a given molecule, then trained a classifier on that representation to predict the chemist from the molecule alone
top-5 accuracy 60%
Said nearby on the same subject
“Trained a cost estimator model… emits a probability distribution over multiple buckets… I’m used to training these classifiers by giving it different enums”
♫
Said on The a16z Show
“The model is basically predicting at very high frequency… at least 50-dimensional vector… 10-20 times a second… potentially predicting one to two seconds out”
♫
Said on Latent Space: The AI Engineer Podcast
“I’ve developed a tool specifically trained on authors ... scans your website ... looks at meta tags, title tags, creates schema for you and does a social platform scan”
♫
Said on The Self Publishing Show
Moments like this, every Monday
The strongest new moments on your topics, in the speakers' own words. Free for three topics.
No card. No password. One-click unsubscribe.
Track a competitor?
Pro watches up to 25 names or phrases and emails you the morning after they are said.
See Pro, $29 a monthVerbatim quotes a short extract and links to the episode; the episode belongs to its publisher. Remove a show.