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.

Transcript
Play full episode

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