Maaz
@mmaaz_98
Followers
429
Following
785
Media
58
Statuses
584
MD & PhD Operations Research (c) @ University of Toronto. Really into polynomials these days.
Toronto
Joined February 2022
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
مدريد
• 1413401 Tweets
دورتموند
• 709481 Tweets
Real Madrid
• 349972 Tweets
Champions
• 242954 Tweets
Vini
• 193157 Tweets
Dortmund
• 174838 Tweets
Eminem
• 130722 Tweets
River
• 123710 Tweets
Celtics
• 106132 Tweets
The Atlantic
• 72995 Tweets
Knicks
• 56842 Tweets
E. Coli
• 56011 Tweets
Lakers
• 49772 Tweets
Tatum
• 44492 Tweets
Mark Meadows
• 33100 Tweets
Mineiro
• 31806 Tweets
John Kelly
• 24178 Tweets
Cano
• 23112 Tweets
Nacho
• 19920 Tweets
#WWENXT
• 18786 Tweets
Chandler
• 17824 Tweets
Gallardo
• 16686 Tweets
Mikal Bridges
• 13368 Tweets
الريال
• 13200 Tweets
#じゃがりこの日
• 11646 Tweets
おはスタ
• 11563 Tweets
Fonseca
• 11531 Tweets
梶谷引退
• 11062 Tweets
#ManiaDeVocê
• 10631 Tweets
Brunson
• 10329 Tweets
Had a (pretty niche) research idea and I see someone else just published it a couple months ago, and remembered this rick rubin quote
1
30
216
@d_feldman
I was talking to a girl once who said she "works in tech"
Later found out she was a marketing intern at chicfila lol
2
0
57
@miniapeur
It’s funny that working over C makes things easier in a lot of fields, eg algebraic geometry
1
0
40
Too many Pubmed papers to read? Try out my new app, Xtract⬇️
Paste your Pubmed search url, and get questions answered directly from those papers, with sources :)
Makes literature reviews way faster 🚀
2
5
32
@brodriguesco
I’ll always maintain R is better than Python for data wrangling and plotting but this is a very unpopular opinion in my department (engineering)
3
0
24
I made a Chrome extension that hides pictures of food so that you don't get tempted while fasting :)
It uses a quantized CLIP model that runs fully locally in the browser, thanks to the amazing Transformers.js library from
@xenovacom
@huggingface
3
1
18
Overjoyed to receive a Vanier Scholarship
@NSERC_CRSNG
for my proposal "Minimal supergraphs for saturating stable matchings", studying combinatorial properties of matching algorithms. I could not have done it without the support of my friends and my advisor Timothy Chan
@uoftmie
0
4
19
Had a great time presenting our work "Siamese Graph Neural Networks for Drug Discovery" and meeting people working on the cutting edge of biotech at the Molecular Machine Learning 2023 conference
@Mila_Quebec
!
1
1
16
Anyone else work hard on a theorem or a paper and then when you finally finish you think it’s not that good because in retrospect it’s “trivial”?
1
0
14
The paper I worked on last summer at
@AmazonScience
was accepted
@IJCAIconf
, Workshop on Generalizing from Limited Resources in the Open World :)
Summary below...
2
2
14
@gabrielpeyre
In case anyone else had the same q as me, Richard von Mises is actually the brother of economist Ludwig von Mises
0
0
13
Very excited to be joining
@AmazonScience
as a research intern, where I'll be working on natural language understanding!
1
1
13
@rcdaut
@paulg
@cremieuxrecueil
In Boolean algebra, the “addition” operation is OR, and “multiplication” is AND.
X and (Y or Z) = (X and Y) or (X and Z)
X or (Y and Z) = (X or Y) and (X or Z)
1
0
11
Just looked at some proofs I wrote a few months ago and it took me a while to understand what I wrote, yet it seemed so obvious at the time lol
1
0
11
I had an incredible time last week at MSRI
@mathmoves
at the "Mathematics and Computer Science of Market and Mechanism Design" workshop!
Also, Berkeley is an extremely beautiful campus and town. I want to move there so badly...
0
1
10
super proud of my younger brother Moeez for placing 1st at the McGill Codejam hackathon - his team made a generative AI powered app which lets you try on clothes virtually. check it out below :)
(he continues to be the smarter brother)
0
0
10
@internetvin
I’ve been wanting to start one in downtown Toronto for years. The closest thing is church street espresso which is open until 10 — but it used to be open until 11 a couple years ago!
2
1
9
@JadeMasterMath
Similar story. The highest undergrad math I took was 2nd year calculus. I self studied a lot of math out of self interest. My department def took a chance on me in admitting me to an OR PhD.
Jumping from 2nd yr Calc to grad level real analysis was def crazy tho lol
1
0
8
@TonyTheLion2500
Combinatorial Optimization by Papadimitrou. Introduced me to the wonderful world of complexity theory.
0
2
8
Living in Boston makes me grind harder because two slices of pizza, a movie, and a popcorn cost me $50 USD
1
0
7
Back at UofT — campus is much greener (in the summer!) and more beautiful than MIT lol
0
1
7
@AlgebraFact
I find the American mathematical monthly often has interesting papers about such topics
0
0
7
I think I finally understand the multivariate discriminant thanks to GPT o1
This may have just saved me for my upcoming committee meeting
0
0
6
What sort of jobs can you work on scientific computing software? Things like optimization algorithms, solvers, simulators, etc.
I think some applied math or engineering departments value this? But I’m not sure.
1
0
6
It’s cool how if you have a niche math q you can just email one of the experts in that topic and they’ll actually talk to you.
Like I once had a question about the Hong projection operator from CAD, so I emailed Hong himself and he got on a Zoom call with me!
0
0
6
This is such an interesting definition because it really captures the essence of what differentiates the real numbers and the complex numbers
0
0
5
@yacineMTB
yes and the worst is when you ask it a coding question and it spends all this time on boilerplate and then errors out just before it writes the actual code I need. (I have premium)
I've had to go back to actually reading pandas documentation 💀
1
0
5
SymPy can now compute roots of polynomials with *algebraic* coefficients -- previously it only worked for rational coefficients. Code contributed by me :)
1
0
5
Found these excellent lecture notes by Heng Yang at Harvard on semidefinite programming. I especially like the engineering applications!
2
1
5
When I cite a classical result, I'm unsure if I should cite the *original* or a textbook. Eg I just wrote:
"...may be impossible if we sum the rewards of the Markov process for 5 periods or more, due to the Abel-Ruffini Theorem (Ruffini 1799, Abel 1824)."
idk just feels cooler?
2
0
3
I kinda love how they keep calling the second-order cone the “ice cream cone”
0
0
5
Had a stimulating day today
@FieldsInstitute
at the Workshop on Quantum Computing and Operations Research :)
0
2
4
Apparently the UofT Mathematica license expired a month ago so I'm trying to convert people in my department to SymPy
0
0
5
The RAND Corporation in the early 1950s contained “what may have been the most remarkable group of mathematicians working on optimization ever assembled” [6]: Arrow, Bellman, Dantzig, Flood, Ford, Fulkerson, Gale, Johnson, Nash, Orchard-Hays, Robinson, Shapley, Simon, Wagner...
1
0
5
Very interested to see if a similar approach can be used to further optimize fundamental algorithms in combinatorics (for example)
Today in
@Nature
:
#AlphaTensor
, an AI system for discovering novel, efficient, and exact algorithms for matrix multiplication - a building block of modern computations. AlphaTensor finds faster algorithms for many matrix sizes: & 1/
114
2K
8K
0
0
5
Listening to stories from my friends who work in hospitals makes me genuinely shocked that hospitals work at all
2
0
5
Reading a math paper. Can’t understand it after multiple rereads. ChatGPT can’t understand it either.
Might have to pull out old reliable: printing it out and working through every step by hand
1
0
5
@fahdananta
V true
Many of the smartest ppl I know are from scarbs / east York / north York
Some of the best high schools there too
0
0
5
In hindsight, it is pretty crazy that I used a laptop without a working speaker, mic, headphone jack, bluetooth, and sometimes screen, for more than a year
2
0
5
Do kids just not know how to use computers anymore, like wdym you don’t know how to take a screenshot… I thought that was a common thing
3
0
5
@CaltechEconThry
The paper I remember off the top of my head is "The Implementation Duality" (Nöldeke and Samuelson, ECTA 2018) which uses Galois connections to generalize various results in matching
0
0
2
I was telling my friend (big co senior software engineer) about NP-hard problems the other day and how many important computer science problems are NP-complete and hence are widely believed to have no polynomial time algorithm. This was all new to him!
1
0
4
I love it when older papers use the first-person. Feel like you don't see it much anymore.
1
0
4
Recently
@UpshotNYT
had this wonderful illustration (left) of how countries would rank under different weightings of Olympic medals.
I tried to develop a way to characterize this analytically (right), using a mix of some optimization and geometry.
0
0
4
It took me like twice as long to mark as the other TAs bc I actually wrote “code reviews” on the students’ code. I just really want them to be better coders man 🥲
0
0
5
Well one of my students said that the quiz was really similar to the lab and that made me happy bc that was kinda the point when I wrote them 😀
0
0
4
I was struggling with leaning cylindrical algebraic decomposition, and then I finally found all the answers in Christopher W. Brown's PhD thesis (1999, University of Delaware).
Pure gold!
Somewhere out there is a phd thesis 250+ pages, read by maybe 2-3 people (who couldn’t comprehend it’s significance), that has all the answers to the questions you have.
14
24
359
0
0
4
@NTFabiano
@ICESOntario
@solmi_marco
@anderskk3
@Chris_Perlman
@PTanuseputro
@MWpugliese
@DanielMyran
@JAMAPsych
thats sad, seems like public discourse doesn't talk enough about the risks of cannabis
0
0
4
@KaiyuYang4
Really cool work! I think I saw Logan present the beginnings of this last year in Xujie's automated reasoning class, glad to see it out.
0
0
3
Living at an MIT dorm is funny, I just found out one of the new directors at Boston Dynamics lives in the same building as me
0
0
4
I remember the first thing that blew my mind when I tried Chatgpt in 2022 was asking it to write a rap about Brenier's theorem from optimal transport in the style of Lil Uzi Vert. And it was actually good.
I tried again yesterday and it was awfully generic.
GPT-2 (2019) vs. GPT-4 (2024)
look how beautifully original gpt-2 is! it's like comparing gpt's poetic child drawings vs. corporate emails from middle age.
89
671
10K
2
0
4
@SantiRCastillo
The reimbursement system is so ridiculous and arduous that it's put me off from going to conferences.
The only reason it doesn't affect me *much* is bc I had savings from doing an industry internship, which is kinda an indictment on academia LMAO
0
0
2
Hearing stories of med students pasting patient info into chatgpt, totally unaware with the privacy issues of sending that to the openai server
Hospitals seem behind on opsec in the face of new technology
0
0
3
One of the best books I've read recently is
Nonnegative Matrices in the Mathematical Sciences (Berman and Plemmons, 1979)
Absolutely full of useful results about types of matrices that often arise in applications
0
1
4
who leaked my schedule
ideal work day
- wake up 2pm
- eat blinner (breakfast, lunch, dinner)
- browse twitter
- go for a walk / hit the gym
- code from 5pm-4am
- watch the entire history of mario speedruns or something of the like on youtube
- have a restful 8 hours of sleep from 6am-2pm
165
1K
16K
0
0
3
I was busy working but then
@AahilDayani
asked me if I wanted to go to the Hans Zimmer concert and an hour later there we were
1
0
4
After TAing undergrad analytics the main thing I’ve realized is reading documentation is a lost art
0
0
4
One of the most interesting facts in complexity to me is that the determinant is O(n^3) but the permanent is Sharp-P-hard. Even tho they’re the same thing except for the signs of the summands.
1
0
4
I used it to recreate the systematic review from Tahir et al (2021) "The Cost-Effectiveness of Mammography-Based Breast Cancer Screening in Canada". Their search string yields 1000+ results, but Xtract finds exactly the <5 papers that they screened, in a few seconds!
1
0
4
@ccanonne_
I mean I like these names better than what the ML community does w their acronyms haha
0
0
4
@NTFabiano
@J_Psychosom_Res
@AntMikocka
@uOPsychiatry
Very prescient. I've always been annoyed by this. In math/cs/engineering type disciplines, my experience is it is much less painful as one simply submits a LaTeX file, which the journal can seamlessly load their own style file on to. Imo more people in med should use LaTeX.
0
0
4
Cutting the last 30 words to meet a word limit is def the toughest part of writing... like there's no more contractions I can make lol
0
0
3
The other side of this
0
0
3
@cxbln
@PierreChambon6
@Dr_ASChaudhari
@curtlanglotz
Really surprised at how "crisp" the images are. I didn't think SD could be used for this as most of the AI art I've seen seems to be in a rather impressionistic style. Great work!
1
0
3
Crawling through mathoverflow for questions about real algebraic geometry (the field of math I know best I think) is like a weekend pastime now
0
0
3
As a Pakistani I never know what to select on these demographic questionnaires
0
0
3
I think the two big steps in math education are:
- going from numbers to symbols (~elementary school)
- getting mostly rid of the numbers and just having symbols and words (~university)
In both of those steps, a lot of people get lost and never progress
1
0
3
Wrote about some interesting connections between polynomials and complexity theory: counting the monomials in a polynomial is #-P-hard (it's related to the permanent), but you can use Schwartz-Zippel to do it probabilistically in polynomial time
1
0
3
Started using Docker to collab on something with my friend... wow it's amazing
0
0
3
@gautamcgoel
In Knuth's book on the stable marriage problem, he says he'd rather the field was called "Algorithmics"
1
0
3