Ilan Shomorony

Assistant prof of ECE at UIUC. Interested in information theory, bioinformatics, and data science

[Ilan Shomorony @ilanshom]

@krsahlin @shenwei356 Thanks, Kristoffer! We are strobemers fans around here too :)

[Mo Tiwari @mo_tiwari]

Thrilled that a paper from my PhD, "Faster Maximum Inner Product Search in High Dimensions" has been accepted to ICML 2024!

In the paper, we accelerate state-of-the-art for the Maximum Inner Product Search (MIPS) problem. MIPS is a core subroutine in systems like recommendation engines, multi-class prediction, matching pursuit, retrieval, and more. We improve the state-of-the-art by using techniques from multi-armed bandits.

Paper: https://t.co/Agp9HgBoxP

Let me know if you'll be in Vienna for ICML!

Congratulations to my coauthors Ryan Kang, Donghyun (Luke) Lee, Jeyong Lee, @chrispiech, @SebastianThrun, @ilanshom, and @martinjzhang!

[Ilan Shomorony @ilanshom]

@lewislehe Awesome! Congratulations!

[Ilan Shomorony @ilanshom]

@martinjzhang @dntse @james_y_zou @bpasaniuc Congratulations, Martin! Very well deserved!

[Mo Tiwari @mo_tiwari]

Want something better than k-means? I'm happy to announce our SOTA k-medoids algorithm from NeurIPS 2020, BanditPAM, is now publicly available in Python, R, and C++! `pip install banditpam` or `install.packages("banditpam")` and you're good to go! 1/n

[Ilan Shomorony @ilanshom]

Join us for the 6th Health Data Analytics Summit!
"Generative AI: A Promising Catalyst for Healthcare?"
Wednesday, April 5, 8:45am to 3pm (CT)
More information and registration here: https://t.co/qNtQYSVI8c

[Ilan Shomorony @ilanshom]

@lewislehe have you thought about teaching a tikz class?

[Ilan Shomorony @ilanshom]

Proud of this work with @mo_tiwari, @martinjzhang and others!

[Stanford AI Lab @StanfordAILab]

Want something better than k-means? Great work from @mo_tiwari & co on BanditPAM, a SOTA k-medoids algorithm. `pip3 install banditpam` and you're good to go! Check out the blog post: https://t.co/4DMw6BoVTd

[Illinois ECE @ECEILLINOIS]

Illinois ECE Asst. Prof. Ilan Shomorony (@ilanshom) earns prestigious @NSF CAREER Award to better understand genomic data problems

Learn more 🔗 https://t.co/4m9QDL4sPM

[Ilan Shomorony @ilanshom]

@Dr_A_Wooldridge @uofigrainger @IllinoisISE Now you are too busy hanging out with the chancellor all the time! 😂

[Ilan Shomorony @ilanshom]

@Dr_A_Wooldridge @uofigrainger @IllinoisISE Congratulations, @Dr_A_Wooldridge!

[Ilan Shomorony @ilanshom]

New paper with @GoviKamath on how to look at sketching and min-hashes from a rate-distortion perspective!

[Govinda Kamath @GoviKamath]

A recent paper with @ilanshom on doing a rate-distortion analysis of min-hashing. We were a bit surprised it had not been done till now.

https://t.co/QFpRaz8Wfc

[Govinda Kamath @GoviKamath]

A recent paper with @ilanshom on doing a rate-distortion analysis of min-hashing. We were a bit surprised it had not been done till now.

https://t.co/QFpRaz8Wfc

[Martin Jinye Zhang @martinjzhang]

Our recent work, Bandit-PAM, was accepted to NeurIPS! It utilizes adaptive sampling to make PAM, one of the most popular k-medoids clustering algorithm, orders of magnitude faster. Great collaboration with @mo_tiwari James Mayclin @SebastianThrun @chrispiech @ilanshom

[Ilan Shomorony @ilanshom]

@hjpimentel @UCLA @CompMedUCLA @dgsomucla Congrats, Harold!

[Govinda Kamath @GoviKamath]

Delighted to say that my work with @TBaharav and @ilanshom on sequence alignment through a spectral algorithm has been accepted into @recomb2020

https://t.co/vJseXnfAgo

[Ilan Shomorony @ilanshom]

Proud to have been part of this work!

[J. Craig Venter Institute @JCVenterInst]

Our recent work on identification of novel biomarker signatures of health and disease states is now published. Great work by the team!

https://t.co/ECzc8uOWOF

cc @JCVenter

[Jason Chin @infoecho]

Experiment: represent 8 different MHC from GRCh38 in a single "sketch graph".

[Ilan Shomorony @ilanshom]

@TBaharav @KamathGovinda @dntse Also, a big thanks to @NCTC_3000. Their easily usable datasets are invaluable for validating algorithms in this space.

[Ilan Shomorony @ilanshom]

Excited about this work with @TBaharav, @KamathGovinda and @dntse. We show that a rank-1 approximation of the min-hash collision matrix provides good estimates for pairwise sequence alignments. We call this Spectral Jaccard Similarity! https://t.co/TGItjDBXb8