ITUScalableSimSearch (@ERC_SSS) header
ITUScalableSimSearch (@ERC_SSS) profile picture

ITUScalableSimSearch

@ERC_SSS

Scalable Similarity Search project at the IT University of Copenhagen. Led by professor Rasmus Pagh, financed by the European Research Council (ERC).

Joined May 2014
  • Tweets139
  • Following5
  • Followers64
8Media
Photo by ITUScalableSimSearch (@ERC_SSS)Photo by ITUScalableSimSearch (@ERC_SSS)Photo by ITUScalableSimSearch (@ERC_SSS)Photo by ITUScalableSimSearch (@ERC_SSS)Photo by ITUScalableSimSearch (@ERC_SSS)Photo by ITUScalableSimSearch (@ERC_SSS)Photo by ITUScalableSimSearch (@ERC_SSS)
This Twitter feed will feature news from the Scalable Similarity Search project, sss.projects.itu.dk which started May 1, 2014.
1
kNN is now available for Elasticsearch on AWS (opendistro.github.io/for-ela…). Nice to see that the choice of algorithm was informed by ANN-Benchmarks (ann-benchmarks.com/), co-developed by the @ERC_SSS project!
1
3
ITUScalableSimSearch retweeted
Rasmus Pagh (@RasmusPagh1) profile picture
Rasmus Pagh@RasmusPagh15 Aug 2019
Enjoyed speaking at WADS 2019 on Set Similarity! Slides available at itu.dk/people/pagh/papers/WA…
1
14
Serious attempts at improving nearest neighbor algorithms in practice should demonstrate their worth on ann-benchmarks (co-developed by @ERC_SSS team members Martin Aumüller and Alec Faithful)!
Erik Bernhardsson (@bernhardsson) profile pictureErik Bernhardsson
@bernhardsson7 Jul 2019
ann-benchmarks, my project to benchmark approximate nearest neighbor algorithms, just crossed 1k stars on GitHub 🎉 github.com/erikbern/ann-benc…
GitHub - erikbern/ann-benchmarks: Benchmarks of approximate nearest neighbor libraries in Python

GitHub - erikbern/ann-benchmarks: Benchmarks of approximate nearest neighbor libraries in Python

Benchmarks of approximate nearest neighbor libraries in Python - erikbern/ann-benchmarks

github.com
1
1
We are proud to release the PUFFINN library for approximate nearest neighbors, github.com/puffinn/puffinn and the corresponding paper at arxiv.org/pdf/1906.12211 !
GitHub - puffinn/puffinn: Parameterless and Universal FInding of Nearest Neighbors

GitHub - puffinn/puffinn: Parameterless and Universal FInding of Nearest Neighbors

Parameterless and Universal FInding of Nearest Neighbors - puffinn/puffinn

github.com
1
1
12
Huge thanks to Martin Aumüller, Tobias Christiani, and Michael Vesterli for making this possible.
1
PUFFINN was developed to eliminate the parameter tuning that is required for most nearest neighbor search algorithms with theoretical guarantees. Let us know what you think of it!
ITUScalableSimSearch retweeted
Francesco Silvestri (@SilvestriFr) profile picture
Francesco Silvestri@SilvestriFr6 Jun 2019
New paper! “Fair Near Neighbor Search: Independent Range Sampling in High Dimensions“ by M. Aumüller @RasmusPagh1 and myself: arxiv.org/abs/1906.01859. Given a query q, we show how to evenly sample near neighbors of q in a high dimensional space. (1/2)
2
1
7
Speaking at TOCA-SV today! sites.google.com/corp/view/t…
ITUScalableSimSearch retweeted
Microsoft Research (@MSFTResearch) profile picture
Microsoft Research
@MSFTResearch15 May 2019
Microsoft’s Space Partition Tree And Graph (SPTAG), now available on @GitHub, takes advantage of vectors to search through billions of pieces of info in milliseconds: aka.ms/AA52hxp

1
21
2
66
Today marks the end of the generous funding of @ERC_SSS by @ERC_Research. Special thanks to the talented people who contributed, directly or as collaborators. But we are not done! New results are in the pipeline, and I am excited about the progress that has been made. Stay tuned!
3
ITUScalableSimSearch retweeted
Rasmus Pagh (@RasmusPagh1) profile picture
Rasmus Pagh@RasmusPagh114 Dec 2018
Last day of the Foundations of Data Science at @SimonsInstitute - thanks to the organizers and the institute for a great program!
Image from Rasmus Pagh (@RasmusPagh1)
1
14
I thought about whether monochromatic = bichromatic closest pair for an annoyingly long time (probably a few days a year, since 2011) so I am glad to see progress on it! eccc.weizmann.ac.il/report/2…
2
8
21
Thank you @johanvts and Alexander Faithfull for your contributions to the @ERC_SSS project, and best wishes for the next stage of your careers!
3
ITUScalableSimSearch retweeted
Gautam Kamath (@thegautamkamath) profile picture
Gautam Kamath@thegautamkamath27 Nov 2018
The start of the final @SimonsInstitute workshop in the data science program: sublinear algorithms and nearest neighbor search. Piotr Indyk kicks it off! Schedule and stream here (simons.berkeley.edu/workshop…). Stay tuned for talks by @ceacy_ @RasmusPagh1 @ilyaraz2 and more!
Image from Gautam Kamath (@thegautamkamath)
4
16
Join us for next week's workshop on Sublinear Algorithms and Nearest-Neighbor Search at @SimonsInstitute! Many exciting talks by leading experts, live-streamed at simons.berkeley.edu/workshop…
1
6
The talk on "How to Prepare Your Vectors" will be given again at Nvidia on November 15, 11 AM.
ITUScalableSimSearch (@ERC_SSS) profile pictureITUScalableSimSearch@ERC_SSS11 Oct 2018
Favorite recipes for preparing high-dimensional vectors from @RasmusPagh1 presented at @Google, Mountain View, October 11 at 1 PM.
Image from ITUScalableSimSearch (@ERC_SSS)
Favorite recipes for preparing high-dimensional vectors from @RasmusPagh1 presented at @Google, Mountain View, October 11 at 1 PM.
Image from ITUScalableSimSearch (@ERC_SSS)
1
1
8
Great introduction to approximate nearest neighbor algorithms by @ilyaraz2 today at Simons Institute for Theory of Computation Part 1: youtu.be/IgyifV7dsxY Part 2: youtu.be/99STqyJOXmA
2
9
Martin Aumüller, who has been a post-doc in @ERC_SSS since 2016, has started as assistant professor at @ITUkbh. Congratulations! Martin will continue to collaborate with us on similarity search.
7
We thank Samuel McCauley for two great years at a post-doc in @ERC_SSS, and wish him good luck with his next post-doc in Boston!
2
Load more