Matthew Ho (@MattHoCosmo) profile picture

Postdoctoral Researcher @astroIAP interested in cosmology, astronomy, statistics, and machine learning.

Joined February 2022
Photo by Matthew Ho (@MattHoCosmo)Photo by Matthew Ho (@MattHoCosmo)Photo by Matthew Ho (@MattHoCosmo)Photo by Matthew Ho (@MattHoCosmo)Photo by Matthew Ho (@MattHoCosmo)Photo by Matthew Ho (@MattHoCosmo)Photo by Matthew Ho (@MattHoCosmo)Photo by Matthew Ho (@MattHoCosmo)Photo by Matthew Ho (@MattHoCosmo)
Very proud to announce... LtU-ILI, an all-in-one framework for ML parameter inference in astro/cosmo!💻🔭🌌 github.com/maho3/ltu-ili We unified the neural nets, training methods, and validation metrics used at the forefront of ML+astro in one accessible package... (1/9)
2
5
1
28
3,593
Lastly, a cool demonstration of combined data constraints: Three trials infer galactic dust parameters, from either the [g-r] colors (blue), the r-band (orange), or both (green). Using the combination of data results in much tighter constraints than using each separately. (8/9)
1
1
212
For more info, checkout: - the code (github.com/maho3/ltu-ili) - the paper (arxiv.org/abs/2402.05137) Of course, with any new code, we want your feedback, so don't hesitate to request new features or point out 🐛's on the Github. :) (9/9)
2
196
Watch the important concepts emerge much faster with our information-ordering (IOB) vs standard PCA compression of semantic latent space. Visualisations generated by unCLIP diffusion from compressed latent space vectors. @MattHoCosmo nitter.cf/bwandelt/status/166103…
What if we could use machine learning to turn information into insights? Our new paper, led by the outstanding Matt Ho @MattHoCosmo discovers key degrees of freedom for describing data, ordered by importance. It is now on the arXiv: arXiv.org/abs/2305.11213 1/5
1
1
5
1,678
What if we could use machine learning to turn information into insights? Our new paper, led by the outstanding Matt Ho @MattHoCosmo discovers key degrees of freedom for describing data, ordered by importance. It is now on the arXiv: arXiv.org/abs/2305.11213 1/5
2
2
1
16
3,520
Unravel the mysteries of the universe's birth! Our paper uses score-based generative models to infer the early universe's density field from present-day dark matter density with stunning accuracy. Check it out: arxiv.org/abs/2304.03788 Led by Ronan Legin, @MattHoCosmo
4
13
1
40
5,753
Matthew Ho retweeted
The Horn Antenna in Holmdel, NJ was used by Nobel laureates and Bell Labs scientists A. Penzias & B Wilson to discover the cosmic background radiation. This petition is asking the town to turn the site into a park & not sell it to a developer. actionnetwork.org/petitions/…
9
35
6
156
Matthew Ho retweeted
🫂Welcome to the newcomers of 2022: PhD students, post-docs and engineers @astroIAP🌟! @CNRS_Paris @INSU_CNRS @Sorbonne_Univ_ @ScienceSorbonne
5
2
35
New desktop wallpaper!
🌟 A star is born! Behind the curtain of dust and gas in these “Cosmic Cliffs” are previously hidden baby stars, now uncovered by Webb. We know — this is a show-stopper. Just take a second to admire the Carina Nebula in all its glory: nasa.gov/webbfirstimages/ #UnfoldTheUniverse
3
Matthew Ho retweeted
It's here–the deepest, sharpest infrared view of the universe to date: Webb's First Deep Field. Previewed by @POTUS on July 11, it shows galaxies once invisible to us. The full set of @NASAWebb's first full-color images & data will be revealed July 12: nasa.gov/webbfirstimages
7,480
120,969
34,757
524,822
In 1933, Zwicky’s measurement of the Coma cluster's mass led him to infer the existence of dark matter. @MattHoCosmo, @astro_michelle et al. have applied a new DL approach to revisit Zwicky's measurement and demonstrate the progress of modern astronomy. nature.com/articles/s41550-0…
2
1
14
The PhD candidates are ready! So freakin’ proud of all of them! 🎉 #CMUgrad #tartanproud
1
22
A wonderful moment when we shared the image with the world.
261
2,165
395
34,439
I've been grading my physics final exam, and I'm happy to report that student priorities haven't changed.
17
62
12
805