NeurIPS Approximate Inference in BDL Competition (@bdl_competition) profile picture

Official account for NeurIPS 2021 competition on approximate inference in Bayesian deep learning.

Joined July 2021
Don't miss the Bayesian deep learning workshop tomorrow! Our winner @tmoellenhoff will be presenting their solution 16:40-17:00pm GMT and @andrewgwils will also give a brief overview of the competition. Event link: neurips.cc/virtual/2021/work…
I will be giving another talk about our winning solution at the Bayesian deep learning workshop on Dec 14 in the slot from 16:40-17:00 GMT. Please feel free to check it out and thanks to the organizers for setting up this fantastic competition!
2
4
Check out the presentations by our winners @tmoellenhoff, @niket096, A. Thin and @ArnaudDelaunoy 👏👏👏 Many interesting insights and new ideas pushing our understanding of Bayesian deep learning forward! youtu.be/LQInlN5EU7E
NeurIPS Approximate Inference in BDL Competition retweeted
I heard a rumour there is this amazing Approximate Inference in Bayesian Deep Learning competition at #NeurIPS2021 tomorrow, starting at 1 pm ET. From what I understand, the winners will be revealing their solutions, and the link to join is neurips.cc/virtual/2021/comp…. 🤫
42
1
186
NeurIPS Approximate Inference in BDL Competition retweeted
The method itself is based on the Bayesian learning rule, a variant originally proposed by our PhD student Wu Lin from @CAIDA_UBC. More details will be released, I believe, in the next few days.
Our new paper on "The Bayesian Learning Rule" is now on arXiv, where we provide a common learning-principle behind a variety of learning algorithms (optimization, deep learning, and graphical models). arxiv.org/abs/2107.04562 Guess what, the principle is Bayesian. A very long🧵
3
24
NeurIPS Approximate Inference in BDL Competition retweeted
Congratulations to all the winners! So happy to see our group’s work @RIKEN_AIP_EN win this competition! Thanks to @tmoellenhoff @ShhhPeaceful @PeterNickl_ Y Shen. Thanks to all who organized and participated. Hope this will help the future work in uncertainty in deep learning!
Excited to announce the winners of our competition! 🥇Team @riken_en (@tmoellenhoff, Y. Shen, @ShhhPeaceful, @PeterNickl_, @EmtiyazKhan) wins both competition tracks! 🥈🥈@niket096 and A. Thin are second in the extended and tie with @ArnaudDelaunoy for second in the light track.
3
7
2
61
NeurIPS Approximate Inference in BDL Competition retweeted
Congratulations to all winners, but especially to my PhD student @ArnaudDelaunoy for his 2nd position! 💪🍾
Excited to announce the winners of our competition! 🥇Team @riken_en (@tmoellenhoff, Y. Shen, @ShhhPeaceful, @PeterNickl_, @EmtiyazKhan) wins both competition tracks! 🥈🥈@niket096 and A. Thin are second in the extended and tie with @ArnaudDelaunoy for second in the light track.
2
5
40
NeurIPS Approximate Inference in BDL Competition retweeted
Congratulations to Nikita @niket096, Ph.D. student at our group at Skoltech on winning 2nd place in the NeurIPS competition on Bayesian DL (together with Achille Thin). Kudos to all the participants and winners. Very interested in looking at solutions.
Excited to announce the winners of our competition! 🥇Team @riken_en (@tmoellenhoff, Y. Shen, @ShhhPeaceful, @PeterNickl_, @EmtiyazKhan) wins both competition tracks! 🥈🥈@niket096 and A. Thin are second in the extended and tie with @ArnaudDelaunoy for second in the light track.
2
19
Excited to announce the winners of our competition! 🥇Team @riken_en (@tmoellenhoff, Y. Shen, @ShhhPeaceful, @PeterNickl_, @EmtiyazKhan) wins both competition tracks! 🥈🥈@niket096 and A. Thin are second in the extended and tie with @ArnaudDelaunoy for second in the light track.
4
14
6
99
Congratulations to our winners and a huge thank you to all participants for working on the competition🙂 Full results are available here: docs.google.com/spreadsheets…. Stay tuned for details on the winning solutions!
11
The submissions to the competition are now officially closed! We will now start evaluating the submissions on the private data to determine the winners. If you participated, please send us ([email protected]) the following information by Thursday Dec 2 ↓
1
1
5
1. The submission id you want to be evaluated on the private data; by default, we will use your best submission according to the public leaderboard. 2. The list of your team members 3. The code for your submission 4. A paragraph describing your solution
1
1
We will announce the results by Friday Dec 3! We invite all participants to prepare a short presentation on their method and join us on Dec 9 (6-8pm GMT) for discussion, we will post further details soon, stay tuned!
1
Hi everyone! We have a few updates about the competition: - We decided to use CIFAR-10 for our light track instead of the diabetic retinopathy dataset, so now it is easier to get involved! - We are extending the deadline to November 26, there is still two weeks left!
1
3
5
Note that to contest for the prizes you need to use the correct architecture for each dataset. All details at izmailovpavel.github.io/neur…
2
NeurIPS Approximate Inference in BDL Competition retweeted
Submissions are closing on November 15th for the Approximate Inference in Bayesian Deep Learning competition at #NeurIPS2021. This is a last chance to participate! See the overview video for more information: youtube.com/watch?v=2gYhaiXy…
15
41
We are officially entering the final stage of the competition! Now is a great time to get involved :) Competition website: izmailovpavel.github.io/neur… Submission system: competitions.codalab.org/com…
1
5
2
21
There are cash prizes (up to $2000) and the best submissions will be presented as a featured talk at at the #NeurIPS 2021 Bayesian deep learning workshop! We provide detailed starting guides, and there is a light track making it easy to participate!
1
1
Our #NeurIPS2021 competition "Approximate Inference in Bayesian Deep Learning" has started! The goal is to provide high quality approximate inference for Bayesian neural networks, using high-fidelity HMC from hundreds of TPUs as a reference. izmailovpavel.github.io/neur…
1
30
4
112
See our getting started guide where we walk you through the submission process: izmailovpavel.github.io/neur… We made the competition language agnostic, so you should be able to use your favorite deep learning framework! There is plenty of opportunity to get involved.
1
2
We are looking forward to your submissions! If you run into any issues or have questions, please send us a DM or email us at [email protected]. Stay tuned to this account for further announcements about the competition :)
2