RoboPapers cover image

Ep#4: Vision Language Models are In-Context Value Learners

RoboPapers

00:00

Exploring Negative Correlation in Trajectories

This chapter delves into the perplexing negative correlation found in a dataset about trajectories. The speakers ponder potential causes, such as corrupted data and random exploration, while stressing the importance of further analysis to clarify these findings.

Play episode from 45:04
Transcript

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app