4min chapter

"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis cover image

E6: The Computer Vision Revolution with Junnan Li and Dongxu Li of BLIP and BLIP2

"The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis

CHAPTER

Optimizing for Image and Text Domains?

The contrast we're learning, like it's the same as what Blip used, to learn better representations that the aligned is image at text. The idea is that you want to learn some kind of limitations for your data such that if your data is similar, like in semantic minuos, they should have similar limitations. And this very simple mechanism, you just use the dot product, which measures the cosine similarity of these normalized embeddings. But then if you want to do more final green interaction between the image and text, you'll need more than just the dot product. You'll need some parameterized mechanism to interact. So that's where we have this cross attention, where

00:00

Get the Snipd
podcast app

Unlock the knowledge in podcasts with the podcast player of the future.
App store bannerPlay store banner

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Save any
moment

Hear something you like? Tap your headphones to save it with AI-generated key takeaways

Share
& Export

Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode