2min chapter

The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) cover image

Learning Long-Time Dependencies with RNNs w/ Konstantin Rusch - #484

The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)

CHAPTER

The Limits of the Gradients

The problem was first proposed in the original alistean paper, actually. Its just synthetic sequential classific nowis actually recression. The idea here is that you have sequences of arbitrary length and their two dimensional. And so it is just adding them together. So we can use sequence of a hundred, we can use a thousand, and we can make it harder, harder, harder. But for corn we actually had that even in the case for five thousand, you get more or less direct convergence,. which is really nice. I'm so very curious how this, what kind of result you saw, and how it compares from a performance perspective.

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