AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
3min chapter
Learning Long-Time Dependencies with RNNs w/ Konstantin Rusch - #484
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
Recurrent Models - How to Learn Long Time Dependencies?
The big problem for recurrent models, arin ends, is that it's quite hard to learn very long time dependencies. And so this is called the wenshing and exploding credit problem because exactly of that. So all we do we suggest a new aren n architecture, recurrent nor network architecture. It may sound well o and stabilize discredit in some way, but it's actually not that easy.
00:00
Transcript
Episode notes
Get the Snipd
podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered
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