Tweet Disematkan
Here is my recent talk at IFIC on deep learning applications in cosmology!
youtube.com/watch?v=07k2JH…
YouTube
English
Pablo Villanueva Domingo
45 posts
Pablo Villanueva Domingo
@CosmoPabloVD
PhD in Physics | Cosmologist | Deep Learning Scientist at @CVC_UAB and @carlasimulator
València, Spain Bergabung Kasım 2021
246 Mengikuti110 Pengikut
Pablo Villanueva Domingo me-retweet
CARLA version 0.10.0 has arrived with a huge upgrade to Unreal Engine 5.5. The upgrade allows CARLA to render detailed high-poly environments with incredible lighting through Unreal Engine's Lumen and Nanite technologies. Check out the release notes: carla.org/2024/12/19/rel…
English
@TingAstro @wfastronomer @zexpe I started around 2016 using Fortran and Mathematica, I moved a few years later to full Python.
English
A colleague inspired me to explore how programming languages have evolved in astronomy, especially since we've been curating astronomical literature.
Fascinating.
English
@DarkCosmologist Funny that the "spirit of European unity" gauge field is not coupled to any field... 😂
English
"You're using GPT-4 for research purposes, right?"
English
Pablo Villanueva Domingo me-retweet
CARLA 0.9.15 is now released!!
CARLA now supports SimReady assets through the #NVIDIAOmniverse Unreal Engine plugin.
CARLA 0.9.15 brings two great new maps, a procedural map generation tool, an HGV vehicle and more!
Read all about it here: carla.org/2023/11/10/rel…
English
I have received a "2021 relevant PhD thesis" prize from @CSIC ! Thanks to all that helped me to make it possible!
He recibido un premio del CSIC a tesis doctoral relevante 2021! Gracias a todos los que me ayudaron a hacerlo posible!
English
Español
Pablo Villanueva Domingo me-retweet
El CSIC ha entregado hoy los Premios Margarita Salas a Mejor trayectoria en supervisión de personal investigador y Premio Tesis doctoral más relevante de 2021 🥇
¡Enhorabuena a todos!
Puedes ver el acto aquí ▶️
youtube.com/live/CfiY7haKo…
YouTube
Español
Pablo Villanueva Domingo me-retweet
Pablo Villanueva Domingo me-retweet
Paper day: arxiv.org/abs/2302.14101. We train GNNs to perform field-level likelihood-free inference using galaxy catalogs from @camels_project. Our models have no scale cutoff, achieved a precision of ~12% when inferring Ωm, and it is robust for 5 != subgrid models! Check it out!
English
Pablo Villanueva Domingo me-retweet
🚀Excited+nervous to share our latest work on understanding geometric GNNs for biomolecules, materials, etc.
"On the Expressive Power of Geometric GNNs" with @crisbodnar @SimMat20 @TacoCohen @pl219_Cambridge
PDF: arxiv.org/abs/2301.09308
Code: github.com/chaitjo/geomet…
Findings👇
English
@a51776901 @rishabh16_ @PetarV_93 When dealing with sparse and irregular data, point clouds... Examples are publications in a citation network, atoms in molecules or galaxies in the universe.
English
Happy Lunar NY 🎉
Just read @PetarV_93's "Everything is Connected". I spent most of 2022 studying Transformers for graphs and this paper intuitively connects most of what I've learned!
🔗: arxiv.org/abs/2301.08210
Here's an executive summary if you haven't caught it yet!
1/9
English
Pablo Villanueva Domingo me-retweet
Current and future neutrino limits on the abundance of primordial black holes. (arXiv:2203.14979v2 [hep-ph] UPDATED) relevance:100% arxiv.org/abs/2203.14979 #darkmatter @carambolos @CosmoPabloVD
English
Happy to announce our new paper on robust cosmological inference in halo catalogs with deep learning. Project led by Helen Shao, undergraduate student at @PU_Astro, and @paco_astro.
Francisco Villaescusa-Navarro@paco_astro
Really proud of our latest paper: "Robust field-level inference with dark matter halos". We study the robustness of field-level inference to differences in N-body codes, hydrodynamics, and astrophysics using thousands of N-body and hydrodynamic simulations. We have developed a
English
Pablo Villanueva Domingo me-retweet
Robust field-level inference with dark matter halos. (arXiv:2209.06843v1 [astro-ph.CO]) relevance:100% arxiv.org/abs/2209.06843 #darkmatter @paco_astro @CosmoPabloVD
English
Pablo Villanueva Domingo me-retweet
Boosting the 21 cm forest signals by the clumpy substructures. (arXiv:2209.01305v1 [astro-ph.CO]) relevance:31% arxiv.org/abs/2209.01305 #darkmatter @CosmoPabloVD
English
Graph neural networks become an ideal tool for these tasks, since they are specially suited to deal with sparse and irregular data, like galaxy catalogues, and they are not limited by a minimum scale.
See more info in the paper arxiv.org/abs/2204.13713 ! (4/4)
English
Our models can also be trained to predict the power spectrum of a galaxy catalogue with a 3% accuracy. This illustrates how graph neural networks can learn clustering information. (3/4)
English
Happy to announce a new paper with @paco_astro !
We apply deep learning methods to infer clustering and cosmological parameters on galaxy catalogues from CAMELS simulations @camels_project. (1/4)
arxiv.org/abs/2204.13713
GIF
English