In RecSys2020 Tutorial Kaggle GMs share quite a few of their secrets.
One of them is the Combining Categories technique which can give your tree model a significant boost.
But how do you implement this out without writing a lot of boilerplate, error-prone code? 🤔
Plus you can use how you combine features to control overfitting.
On top of that, the authors of the tutorial also show you a simple method for how to evaluate features!
Wow, what a read!
The RecSys2020 tutorial bt @NVIDIAAI, where KGs spill their secrets 😄
One year after our "carousel personalization" work at #RecSys2020, I am extremely honored to be once again shortlisted among the best papers at #RecSys2021. Thank you @ACMRecSys, and congratulations to
@olivierjeunen who eventually won the "student" award for his excellent work!
In our experience with #RecSys2020, authors were asked to submit a short 2min teaser for advertisement a prerecorded 10min talk to serve as a backup. Then we had each author present live twice with an 11h interval to suit attendees across the globe.
If you are interested in our work on probing BERT for conversational recommendation from #RecSys2020---or other interesting papers from dutch/belgian-based researchers in top-tier IR conferences---register to @DIR_2020 at dir2020.be/ for🆓 !
@_Guz_ with "What does BERT know about books, movies and music? Probing BERT for Conversational Recommendation", published at #RecSys2020; analysing the knowledge stored in BERT’s parameters and studies how BERT performs in conversational recommendation tasks. 4/10
@_Guz_ with "What does BERT know about books, movies and music? Probing BERT for Conversational Recommendation", published at #RecSys2020; analysing the knowledge stored in BERT’s parameters and studies how BERT performs in conversational recommendation tasks. 4/10