Salva Rühling Cachay
@salvaRC7
PhD student @ucsd_cse & intern @nvidia. Into AI for Science, especially climate&weather. Prev: @allen_ai @PARCinc @Mila_quebec @CMU @TUDarmstadt
Sharing a blog post about our #Neurips2023 paper "DYffusion: A Dynamics-informed Diffusion Model for Spatiotemporal Forecasting" salvarc.github.io/blog/2023/dyff… It's faster than standard diffusion models, has low memory needs, and generates stable & probabilistic rollout forecasts. 1/7
We are delighted to invite submissions to the #NeurIPS2025 workshop "Tackling Climate Change with Machine Learning". Important dates: ▶️Mentorship program: Jul 28 ▶️Papers, proposal, and tutorial submissions: Aug 20 ▶️Workshop: Dec 6 or 7 Learn more: climatechange.ai/events/neurips…
We're organizing the AI for Music workshop at @NeurIPSConf in San Diego! We'll be accepting both papers + demos w/an initial deadline of August 22, well timed for early visibility on your ICASSP/ICLR drafts 👀 Check out the website for more: aiformusicworkshop.github.io
🔥Happy to announce that the AI for Music Workshop is coming to #NeurIPS2025! We have an amazing lineup of speakers! We call for papers & demos (due on August 22)! See you in San Diego!🏖️ @chrisdonahuey @Ilaria__Manco @zawazaw @huangcza @McAuleyLabUCSD @zacknovack @NeurIPSConf
📢📢 Elucidated Rolling Diffusion Models (ERDM) How can we stably roll out diffusion models for sequence generation in data-scarce dynamical systems? We elucidate the design of rolling diffusion, inspired by prob. flow ODEs and nonisotropic noise. 📄 arxiv.org/pdf/2506.20024
Come to our poster today 11 - 2 pm East Exhibition Hall A-C #3905! @salvaRC7 and I will be there!
How can diffusion models accelerate climate emulation? Check out our #NeurIPS2024 Spotlight paper on Spherical Dyffusion! Paper: openreview.net/pdf?id=Ib2iHIJ… Code: github.com/Rose-STL-Lab/s… (1/3)
Come talk to us at Poster session 3: Thursday 11am-2pm at East Hall A-C #3905! #NeurIPS Btw, our code is now available at github.com/Rose-STL-Lab/s… I look forward to discussions @NeurIPS and will be there for the whole week, so please reach out if you'd like to chat! :)
Computer scientists at #UCSanDiego and @Allen_AI have developed a new climate model—combining generative #AI and physics data—that is capable of predicting #climate patterns 25 times faster than the state of the art. ➡️ bit.ly/3ZByZbc
Computer scientists at #UCSanDiego and @Allen_AI have developed a new climate model—combining generative #AI and physics data—that is capable of predicting #climate patterns 25 times faster than the state of the art. ➡️ bit.ly/3ZByZbc
Check out our new work on modeling heavy-tailed distributions with diffusion models. Key highlight: good empirical performance with minimal implementation overhead 🔥. Shoutout to my amazing colleagues during my internship @nvidia x.com/MardaniMorteza…
🌪️ Can Gaussian-based diffusion models handle heavy-tailed data like extreme scientific events? The answer is NO. We’ve redesigned diffusion models with multivariate "Student-t" noise to tackle heavy tails! 📈 📝 Read more: arxiv.org/abs/2410.14171
🎉Our ICML ML4ESM Best Paper is now accepted as a #NeurIPS2024 spotlight!🏆➡️🔦 📄Stay tuned for our camera-ready version with exciting updates, including 100-year inference runs 🌏🌎🌍. Huge thanks to our reviewers for their invaluable feedback & unanimously strong support. 🙏
🌍We've developed the first conditional generative model for accurate & efficient ensemble climate simulations! 📄Read the full paper: arxiv.org/abs/2406.14798 🎉Excited to share that a preliminary version was accepted as an oral presentation @ml4esmworkshop at #ICML2024! 🧵
We are looking for a postdoc to work on scientific foundation models! If you are excited about multi-modal #LLM #AI4Science, #Physics-Guided #DeepLearning pls reach out. Appreciate any repost! roseyu.com/postdoc.html
Deeply honored to receive the best paper award @ml4esmworkshop! Immense gratitude to the organizers and reviewers for this recognition. And of course once again thank you to my amazing co-authors! :) 📄Read the full paper: arxiv.org/abs/2406.14798 #ML4ESM #GenAI #Climate
We are delighted to announce the #ICML2024 ML4ESM Best Paper Award winner: [2406.14798] Probabilistic Emulation of a Global Climate Model with Spherical DYffusion (arxiv.org) Congrats to @salvaRC7 and co-authors @ai2_climate @yuqirose for their amazing work!
How to discover multiple causal graphs from heterogeneous time series? Check out our #icml2024 paper on ``Discovering Mixtures of Structural Causal Models from Time Series Data''! Paper: arxiv.org/abs/2310.06312 Code: github.com/Rose-STL-Lab/M… (1/3)
After almost a year, our review paper on #Physics-Guided #DeepLearning finally appears at @PNASNews pnas.org/doi/10.1073/pn…! It is part of the special issue on #Physics Meet #MachineLearning pnas.org/topic/559 (1/3)
What drives innovation in machine learning? In a new paper, we argue that application-driven work is systemically under-valued in the machine learning community, but that it's essential for both innovation and impact. arxiv.org/abs/2403.17381 1/3
We're excited to announce the @ClimateChangeAI Summer School 2024!☀️ Are you an #AI expert who wants to tackle #ClimateChange? Are you a #climate expert trying to use #ML in your work? Are you curious about the topic? Apply & register now! Learn more👉climatechange.ai/events/summer_…