
Kowshik Thopalli
@kowshik0808
ML Scientist at Lawrence Livermore National Labs, CA. Ph.D from ASU.
ID: 85289778
26-10-2009 11:08:24
389 Tweet
269 Followers
1,1K Following

If you’re at #ICML2023 and are interested in uncertainty quantification for GNNs, come talk to Jay Thiagarajan at our poster tomorrow! 😁



Sadanand Dhume Every media outlet frames the situation from POV of asking what are the implications if allegation is true. Given there is hitherto no evidence to substantiate, what happens if *false*? Why are we lending credibility to claims with no evidence? Reminds me of election denial.


Delighted to share that our recent work on reliably detecting failures in vision models under a variety of shifts has been accepted by ECCV. This was a terrific collaboration with Rakshith Subramanyam Vivek Sivaraman, led by our amazing mentor Jay Thiagarajan Lawrence Livermore National Laboratory #ECCV2024

Birth anniversary of Padma Vibhushan and Sangeet Natak Akademi awardee, singer, composer and the multifaceted musical genius Dr Mangalampalli Balamurali Krishna Garu. ☺️🙏🏼☺️🙏🏼☺️


✨⭐️✨If you are at #ICML2024, check out PAGER, a new approach for proactive failure detection in deep networks (both regression and classification)! Paper: tinyurl.com/pager-paper Session: tinyurl.com/pager-ICML Vivek Sivaraman Puja Trivedi Rushil LLNL Computing


One of my favorite papers this year, written by my awesome colleagues and friends at Lawrence Livermore National Laboratory. PAGER has far-reaching implications with its proactive ability to flag samples on which a deep learning model could fail. It goes beyond thresholding epistemic uncertainties.



