Gautam Kamath(@thegautamkamath) 's Twitter Profileg
Gautam Kamath

@thegautamkamath

Assistant Prof of CS @UWaterloo, Faculty @VectorInst, Canada @CIFAR_News AI Chair. Co-EiC @TmlrOrg. I lead @TheSalonML. Privacy, robustness, machine learning.

ID:97707247

linkhttp://www.gautamkamath.com calendar_today18-12-2009 17:03:35

10,0K Tweets

44,4K Followers

508 Following

Follow People
Vector Institute(@VectorInst) 's Twitter Profile Photo

Meet the 2024-2025 Vector Scholarship in AI recipients who will be studying at University of Waterloo! We are thrilled to have 12 exceptional individuals from the University of Waterloo’s Computer Science, Data Science, Systems Design Engineering, and Electrical and Computer Engineering

Meet the 2024-2025 Vector Scholarship in AI recipients who will be studying at @UWaterloo! We are thrilled to have 12 exceptional individuals from the University of Waterloo’s Computer Science, Data Science, Systems Design Engineering, and Electrical and Computer Engineering
account_circle
Andrew Gordon Wilson(@andrewgwils) 's Twitter Profile Photo

The accepted ICML workshops have been posted. A great opportunity to submit your work! icml.cc/virtual/2024/e…

account_circle
Florian Tramèr(@florian_tramer) 's Twitter Profile Photo

Ted wrote a very nice blog post about our recent paper on empirical ML privacy evaluations (with Michael Aerni & Jie Zhang).

Damien's blog has many cool articles about DP and how to use it in practice. A highly recommended read!

desfontain.es/blog/ml-privac…

account_circle
Gautam Kamath(@thegautamkamath) 's Twitter Profile Photo

I am glad the ML community, across NeurICMLR and beyond, is starting to take this issue more seriously.

If you participate in colluding behaviour, I very strongly recommend you stop *right now*. It's not worth ruining your entire career.

account_circle
Ravid Shwartz Ziv(@ziv_ravid) 's Twitter Profile Photo

I actually think that this is how this kind of paper should look. They are very honest about their algorithm and its pros and cons. And yes, this is not cool to call the work of other people garbage

account_circle
Ken Chiu(@kjw_chiu) 's Twitter Profile Photo

Gautam Kamath This is especially relevant given the history of NN. It took many decades before there was enough compute to show their strengths. An early SGD paper was in 1967! The CDC 6600, considered fast at the time, could do only 3 MFLOPS!
people.idsia.ch/~juergen/amari…

account_circle
Thomas Steinke(@shortstein) 's Twitter Profile Photo

The square root can be uniformly approximated to error ε by a rational function of degree O(log²(1/ε)).

This surprising, since a polynomial approximation (e.g. Taylor series) requires degree Ω(1/ε), so rational functions are exponentially better than polynomials.

The square root can be uniformly approximated to error ε by a rational function of degree O(log²(1/ε)). This surprising, since a polynomial approximation (e.g. Taylor series) requires degree Ω(1/ε), so rational functions are exponentially better than polynomials.
account_circle
Thomas Steinke(@shortstein) 's Twitter Profile Photo

Pre-recorded talks are ubiquitous now, but they were rare pre-COVID.
STOC 2016 made it optional to add a video to the proceedings. This was was so unusual back then that only one paper submitted a video!
And, well, you should watch it (Ctrl+F for 'Video'):
conference-publishing.com/list.php?Event…

account_circle
Arya Mazumdar(@MountainOfMoon) 's Twitter Profile Photo

We continue to accept applications for senior (associate/full) professor positions in Halıcıoğlu Data Science Institute in the areas of 1) Artificial Intelligence Foundations;2) Data Systems: includes Data Security; 3) Intelligent Systems and Applications. apol-recruit.ucsd.edu/JPF03850

account_circle