The Thesis Review cover image

[44] Hady Elsahar - NLG from Structured Knowledge Bases (& Controlling LMs)

The Thesis Review

CHAPTER

The Importance of Distributional Constraints in Regression Learning

We have gained some attention by this work I would say internally definitely externally also. This is always the trade-off of doing something that could be easily understood by the reviewers or revisiting some theory and trying to like put it until the end so unlike prompting for example if you read if you read ten other papers about prompting you need to read less to review a paper for prompting but if you get something that's completely not completely but not following the same line of work for example it's hard to sell because people assume so much. We lay the landscape between pure reward maximization distributional matching or something in between which is like the KL control where you like to maximize a reward but you still take

00:00
Transcript
Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner