AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
3min chapter
Common Sense Reasoning in NLP with Vered Shwartz - #461
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
How Many Questions Would the Model Generate for a Given Challenge Paragraph?
Most of the knowledge was grammatical and understandable, which is expected because we know that neural language models are very good at generating fluent text. But something that was interesting is that when we asked human evalators to judge this generated knowledge, they didn't find we looked at the knowledge that was in practice helpful for the model. So i think there is also some limitation here about when you ask people to consider whether some kind of knowledge is helpful for making a decision. We're still thinking of ways to guide the generation to towards a helpful knowledge or maybe ways to judge whether knowledge is its going to be helpful or not.
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