“We did clustering of building blocks on similarity and avoided entire clusters from being overlapped between test and train… performance drops when we do that”
♫
Said on Latent Space: The AI Engineer Podcast
“You should do something like XGBoost… we see XGBoost drop steadily as you get further away with building-block cluster approaches”
Latent Space: The AI Engineer Podcast · Lab-in-the-Loop at Compute Speed - Ian Quigley, Leash Bio
Play from 45:14
45:14
Audio streams from the publisher’s own feed. Verbatim does not host or copy it.
What was described
used XGBoost as a baseline model and measured its performance drop as test data moved farther from training building-block clusters
Said nearby on the same subject
“Stress testing the models ... trying to expose weaknesses ... creates a baseline for the performance of your models”
DS
Dan Stocker
Cyber Risk Management Podcast ·
“We trained the final model on the SpaceX cluster ... a few weeks of pre-training, then synthetic data NRL and then shipping it”
ML
Misha Laskin Beam
No Priors: Artificial Intelligence | Technology | Startups ·
“We’re mainly in Azure Shop using Azure OpenAI… testing in GCP with Vertex AI, also using their Anthropic models because we can use GCP as the broker, and we know that the data will stay with us”
JW
Jason Ward
Agentic Conversations (formally mlops.community) ·
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.