304 karma · joined November 4, 2014
I had a similar experience few years back when participating in a ML competitions [1,2] for detecting and typing phrases in a text. I submitted an approach based on Named Enttiy Recognition using Conditional Random Field (CRF) which has been quite robust and well known in the community and my solution beat most of tuned Deep learning solutions by quite a large margin [1].
I think a lot of folks underestimate the complexity of using some of these models (DL, LLM) and just throw them at the problem or don't compare it well against well established baselines.
[1] https://scholar.google.com/citations?view_op=view_citation&h... [2] https://scholar.google.com/citations?view_op=view_citation&h...
Used in non generative language models like BERT but should help with generative models as well.
Old HN thread: https://news.ycombinator.com/item?id=12932183
Sorry about that. I can't delete this post hence posting this comment.
[1] https://en.wikipedia.org/wiki/Unified_Payments_Interface
I recently visited B&N and the feeling of moving around and sitting with books after spending all week in front of screen was quite refreshing.
The coin toss forces the suboptimal move, thereby giving the other player a chance to play the optimal move next time. The original game is quite easy, and becomes boring because most games tend to be tied. This configurations will allow individuals to keep playing, for a longer duration, knowing that the other player might not be able to play the most optimal move.
All kind of feedback in welcome. Do contribute to the repository if you like.