Ativ Joshi (@ativsc) profile picture

संसारोऽयम अतीव विचित्रः | CS Ph.D. Student @UMassAmherst | Online Optimization, ML Theory | Prev: Research Assistant @TIFRScience, CMI, @ahduniv

Ahmedabad, India
Joined December 2013
Photo by Ativ Joshi (@ativsc)Photo by Ativ Joshi (@ativsc)Photo by Ativ Joshi (@ativsc)Photo by Ativ Joshi (@ativsc)Photo by Ativ Joshi (@ativsc)Photo by Ativ Joshi (@ativsc)Photo by Ativ Joshi (@ativsc)Photo by Ativ Joshi (@ativsc)
Ativ Joshi retweeted
Replying to @lu__jasper
both use the same principle of checking a few bits every once in a while and saying its probably fine
1
4
261
Ativ Joshi retweeted
We should talk about "Existentialist adults" the way we talk about "Disney adults." Everyone goes through an Existentialist phase as a teen, but at some point, you gotta grow out of it.
86
36
15
510
24,037
Ativ Joshi retweeted
Unpopular opinion: “Everything happens for a reason” is the most successful post-hoc rationalization in human history. We don’t find meaning. We manufacture it, then forget we built the factory.
1
1
2
166
"There is but one truly serious philosophical problem, and that is suicide."
Without naming the book, tell me a line that readers of it will instantly recognize.
1
6
375
hot take i know many friends will disagree with me on: this is a deeply selfish act of theft from civilization if you’re a smart capable young person, society poured millions of dollars into raising you you have a moral obligation to pay it back with the surplus of your labor
my smartest friend in college left tech & is now a farmer at a buddhist monastery his days are: harvesting & meditating & walking through the forest & cooking his meals he could excel at anything he wanted. he chose a life that is simple and good, with all his intelligence
1,404
171
755
4,696
2,733,855
With recommender systems becoming extremely powerful, using social media feels like being watched by a human suggesting content based on overall activity. Feels creepy 😐 particularly when the recommendations are based on personal chats. Privacy needs a revision in the LLM era.
1
4
422
We are thrilled to announce that the work of Sreejata Bhattacharya, a graduate student in the School of Technology & Computer Science at #TIFRMumbai, has jointly won the Best Student Paper Award in the upcoming Innovations in Theoretical Computer Science (ITCS 2025) conference.
1
8
1
48
4,598
This is a beautiful illustration of an apparent paradox in deep learning that ~the weights don't move~ I think we resolved this paradox in prior work, so I just want to share our perspective. And before you ask: yes, it's a question of norms 😅 (1/6) nitter.cf/norabelrose/status/184…
If you make a drawing in the weight matrices of your neural network at initialization, it will likely still be visible at the end of training arxiv.org/abs/2012.02550
4
50
4
540
100,264
Postulates are believed to be known to be true while axioms are known to be believed to be true. The distinction is subtle and often blurred in our mathematical treatment. It took me a good while to appreciate this.
1
1
5
359
16 popular math youtube channels (thread 🧵) comment with the ones missing from our list.
69
442
46
2,473
279,657
Happy to share that two of our papers have been accepted to the ICML workshop on "Foundations of RL and Control" (rl-control-theory.github.io/). Details below.
1
2
14
1,465
of course this will be bad, but i feel the need to comment specifically on why, pre-emptively. although it reads like the 'to be fair, you have to have a very high IQ' rick and morty post, its easy to misunderstand king of the hill. its not a sitcom, or a comedy. [...]
Pamela Adlon says “a new generation of young writers” are working on the ‘KING OF THE HILL’ revival. She adds that they are adapting “sweet spots of the classic” into the new era. (Source: @Variety)
291
1,449
208
21,861
10,249,197
For the past few years, I've been putting my (LaTeXed) math notes on my website (sharmaeklavya2.github.io/not…) so that I can access them from any device and share them with others if needed. Turns out they're getting 4K hits per month on Google search! 😱 Totally unexpected.
1
1
1
20
2,120
Ativ Joshi retweeted
You are brave. You are smart. You’ve overcome a lot in your life to get to where you are. But remember. You’re not the only one who has.
7
9
102
13,328
Interesting exchange here with evolutionary biologist and ornithologist (bird expert) Richard Prum: conversationswithtyler.com/e…
73
103
30
699
143,394
Seen at NeurIPS 👀
10
48
12
843
80,752
Do stop by at booth 1001 during the evening poster session today at #NeurIPS23! This is joint work with @sinha_mit, @mhajiesmaili and @CameronMusco.
Replying to @_rajarshiB
We give an online policy for fairly allocating resources among a number of agents in an adversarial setup. Our problem setup is very general and can be applied in a wide variety of settings like online caching, job scheduling etc.
3
18
1,798
We give an online policy for fairly allocating resources among a number of agents in an adversarial setup. Our problem setup is very general and can be applied in a wide variety of settings like online caching, job scheduling etc.
1
1
2
2,043
My coauthor Ativ Joshi and I will be presenting our work on no-regret algorithms for online fair resource allocation today at booth 1001 at #NeurIPS23 from 5-7 pm CST! nips.cc/virtual/2023/poster/…
1
1
9
446
Ativ Joshi retweeted
I once wanted to enroll in a course called Graph Theory thinking it will teach cutting edge dashboard tools like ggplot2 and Tableau or at least advanced Excel. Instead it was telling how many ways one can draw lines between many points and how to cross the bridges in some random town called Konigsberg. No wonder the youth is demotivated to study these days
7
2
2
32
11,268