🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Only 2 buses added to ⁦BMTC⁩ fleet in the past 9 months. In the same period, 475000 new private vehicles have been added to the city roads. And you complain about congestion? ⁦CM of Karnataka⁩ ⁦@BBMP_MAYOR⁩ ⁦Tushar Giri Nath IAS⁩ thehindu.com/news/cities/ba…

Neural nets can generalize well on test data, but often lack robustness to distributional shifts & adversarial attacks. Our #NeurIPS2020 paper on simplicity bias sheds light on this phenomenon. Poster: session #4, town A2, spot C0, 12pm ET today! Paper: bit.ly/39RXDel

While standard SGD cannot learn from correlated data, SGD with reverse experience replay can! To appear in NeurIPS 2021. Joint work with Suhas Kowshik, Dheeraj Nagaraj and Prateek Jain arxiv.org/abs/2103.05896 arxiv.org/abs/2105.11558

Applications now open for pre-doctoral positions at Google AI India! #predoc careers.google.com/jobs/results/1… Come spend 2 years with our ML and Optimization team working on cutting edge problems in DL understanding, efficient ML, RL, privacy! Manish Gupta , Divy Thakkar , Praneeth Netrapalli

Delighted to announce #MLinIndia social at #NeurIPS2021. Do attend if you are interested in exploring long term career opportunities in ML in India. More details: sites.google.com/view/ml-in-ind… RSVP: docs.google.com/forms/d/e/1FAI… iKDD Sandeep Juneja Amit Sharma Shourya Roy

New paper on the arXiv with Ankit Garg, Praneeth Netrapalli, and Suhail Sherif, which will be presented as a spotlight talk at NeurIPS 2021. Summary: We show nearly optimal quantum and randomized lower bounds for smooth convex optimization. arxiv.org/abs/2112.01118 scirate.com/arxiv/2112.011…

Do input gradients highlight discriminative and task-relevant features? Our #NeurIPS2021 paper takes a three-pronged approach to evaluate the fidelity of input gradient attributions. Poster: session 3, spot C0 Paper: bit.ly/3EzdvyH with Prateek Jain and @pnetrapalli

Happening in less than 24 hours. If you are registered for NeurIPS 2021, you can join the social on gather town using NeurIPS login credentials: eventhosts.gather.town/app/8Bob4qi7Gk…

Enjoyed teaching a short course on optimization in Bangalore School on Statistical Physics ICTS last week. Lectures available now on youtube. youtu.be/uL1Yk2NkEpw

On Sept 5th, Karnataka State Board for Wildlife will decide whether Hesaraghatta will be notified as a conservation reserve. Let’s rally to preserve Hesaraghatta. We cannot let Hesaraghatta be turned into a concrete jungle. Please support: bit.ly/SaveHesaraghat… CM of Karnataka

Workshop for 8 to 12 class mathematics school teachers in Karnataka ICTS-TIFR is conducting a Mathematics Teachers Training camp from 4 to 7 Nov' 22 The camp aims to discuss challenging mathematical problems at the pre-olympiad and olympiad level Last date to register: 30 Sept 22

Very nice work on mitigating simplicity bias by my colleagues @tiwarishabh16 and Pradeep Shenoy !

It was a great learning experience interacting with the naval officers and other speakers, and getting to know the various problems in the naval domain where ML/AI can be useful. Thanks INS Valsura for organizing the workshop and for the kind hospitality!

All in due time... 20 years after "Online Newton Step", comes "Bandit Newton Step" :-) work with my fantastic student Jennifer Sun and my colleagues in Google India, Arun Suggala and Praneeth Netrapalli arxiv.org/abs/2402.08929

Very happy to share this work that we have been working on over the last year.

DEJAN I'm enamored of combining many approaches. Here are some that are interesting and public: Various dense retrieval methods TreeFormer (arxiv.org/abs/2208.09015) High-Recall Approximate Top-K Estimation (arxiv.org/abs/2402.09360) Various forms of KV cache quantization and