Black hole binary populations & constraints on fundamental physics | Flatiron Research Fellow, Flatiron Institute | Prev. PhD Physics, UT Austin
NYC/Austin/Karachi
Joined June 2013
- Tweets4.8K
- Following1.2K
- Followers406
- Likes28.7K
Asad Hussain retweeted
There is an (unpatchable?) supply-chain zero-day in PyTorch autograd that allows a regularization term to use the model's weights as RAM and SGD as its CPU clock to run any computation during training.
Asad Hussain retweeted
A new paper on the potential existence of 0d holography
arxiv.org/abs/2608.14781
Asad Hussain retweeted
After 9 years of work, we're proud that our new vision of Urdu typing is available on the most important platform – Android. This Urdu keyboard, @matnsaz (matn - text, saz - instrument) learns from how Arabic script is taught to children.
Asad Hussain retweeted
Flow-LLM Blogpost :D flow-based-llms.github.io
In the last few weeks, a bunch of work on flows for language came out 🌊
That is exciting, because it makes truly parallel text generation feel real: generation where models can keep refining the whole response during inference, instead of committing token by token.
I wrote an intuitive and animated introduction to the area — why autoregression has a structural ceiling, why discrete diffusion only partly escapes it, and why flows may be the first genuinely parallel alternative.
Here's an overview of the key parts of the blog - and let's chat at #ICLR2026 :)
Asad Hussain retweeted
Sabina Sagynbayeva, in work with @farrwill @FlatironInst, develops a hierarchical Bayesian model that can exploit multiple exoplanet transit light curves to better map stellar surfaces. (link in replies)
If you google “how to be a good theoretical physicist” you get his website where he basically says learn math and do math! And then gives book/lecture notes recommendations from starting from pre-calculus all the way to supergravity. 1/2
Over more than 6 decades, Gerard 't Hooft's discoveries have been critical in the development of modern quantum field theory and the Standard Model of particle physics – not to mention his other contributions, from black holes to holography. This year, a Special Breakthrough Prize in Fundamental Physics recognizes his pioneering career. breakthroughprize.org/News/9…
@UniUtrecht
1
1
438
Asad Hussain retweeted
🎉 We have reached 200 candidates in O4! 🎉
The fourth observing run (O4) of our detector network has had the best performance so far, with more candidates than ever before! We are currently busy analysing all these wonderful data and look forward to sharing results
#O4IsHere
Asad Hussain retweeted
Yearly reminder: What do polar coordinates, polar matrix factorization, & Helmholz decomposition of a vector field have in common?
All are implied by Brenier’s Thm: a cornerstone of Optimal Transport theory. It’s a fundamental decomposition result that’s not widely known
1/n
Asad Hussain retweeted
Dec 27 - Weinberg Soft Theorems from Weinberg adiabatic modes
arxiv.org/abs/1602.05196
Mirbabayi & Simonovic show that the photon and graviton soft theorems for scattering in flat space are a consequence of physical long wavelength configurations tied to large diffs (BSM)
Asad Hussain retweeted
SymbolicRegression.jl → 1.0 🎉
After several years of work, I'm thrilled to announce some major new features! Let me show you what's possible now:
Asad Hussain retweeted
Our new initiative to accelerate AI/ML in Physics & Astro at @UCSDPhySci between @ucsdastro @HDSIUCSD @SDSC_UCSD & @UCSDPhysics got highlighted today!
SMASHing records in the big data era of astronomy and physics! A new initiative led by @UCSDAstro's @FloorAstro brings together researchers from across campus to help spark new collaborations and strengthen their skills and knowledge in machine learning and data science.
Asad Hussain retweeted
On the arXiv today: two papers about the @ESAGaia selection function for astrometric orbits!
First: a full forward-model of Gaia astrometry for binaries, which can tell you whether any hypothetical binary would have gotten a solution in DR3 or not.
arxiv.org/abs/2411.00088
1/n
Asad Hussain retweeted
This year I was going through a lot of transition so I didn't push a lot of papers, but here comes two of them: the first one is we use a diffusion model, i.e. the AI model people use to generate image, to learn the gravitational wave likelihood directly arxiv.org/abs/2410.19956
Asad Hussain retweeted
And here comes the double whammy! This is another paper that just came out, which integrate harmonics, an open source package for computing evidence, with jim, a gravitational wave PE code I am developing. arxiv.org/abs/2410.21076