Software Engineer in Applied Science @ Google.
Joined June 2019
- Tweets41
- Following330
- Followers61
- Likes464
Pinned Tweet
In this new paper from my team @GoogleAI we use ML + a differentiable CFD simulator + TPUs to achieve ~86x speedup over direct numerical simulation.
Coauthored with @shoyer, @dkochkov1, Ayya Alieva, Qing Wang and Michael Brenner.
arxiv.org/abs/2102.01010
1/2
Excited to share "Machine learning accelerated computational fluid dynamics"
arxiv.org/pdf/2102.01010.pdf
We use ML inside a CFD simulator to advance the accuracy/speed Pareto frontier
with/
Jamie A. Smith
Ayya Alieva
Qing Wang
Michael P. Brenner
@shoyer
Jamie Smith retweeted
Google researchers used data from 40 million smartphones to map and continuously monitor the Earth's ionosphere.
This could help improve GPS location accuracy and even improve forecasting for geomagnetic storms. More on how in the link.
newscientist.com/article/245…
Jamie Smith retweeted
Calling all climate enthusiasts! Drop your answers in the comments and let’s see your forecasting skills 🌩️ Discover more about the latest #AI research: goo.gle/3STaz9r
ALT A quiz question asking: NeuralGCM leverages the power of neural networks. What aspect of atmospheric science is NeuralGCM most likely designed to model? A. Jet stream behavior B. Cyclone trajectories C. Weather prediction D. All of the above
Jamie Smith retweeted
NeuralGCM is a method that combines traditional physics-based modeling with ML to accurately and efficiently simulate Earth's atmosphere. Learn how NeuralGCM marks a significant step towards developing more powerful and accessible climate models. → goo.gle/4ff57ah
Jamie Smith retweeted
Exciting news! Meet NeuralGCM, a cutting-edge atmospheric model from Google Research. 🌡🌎
This AI-driven general circulation model (GCM) outperforms traditional atmospheric models in precision, providing faster, more accurate climate predictions. Learn more in @Nature → nature.com/articles/s41586-0… Watch the video here → youtu.be/KUAWw32FjIo
Or read the blog post below ⬇️
NeuralGCM is a method that combines traditional physics-based modeling with ML to accurately and efficiently simulate Earth's atmosphere. Learn how NeuralGCM marks a significant step towards developing more powerful and accessible climate models. → goo.gle/4ff57ah
Jamie Smith retweeted
In @Nature today: NeuralGCM, a breakthrough in climate modeling. It combines physics-based modeling with AI, and is up to 100K times more efficient than other models for simulating the atmosphere, providing scientists with new tools for predicting climate change. nature.com/articles/s41586-0…
Jamie Smith retweeted
Gemini Pro 1.5 a week after Gemini Ultra and 70 days after Gemini Pro 1.0. Who says Google doesn't ship anymore? goo.gle/GeminiV1-5
And with 10M context length, we've never been more back 🕺
Jamie Smith retweeted
New open source release from my team at Google: Dinosaur, a differentiable dynamical core for global atmospheric modeling, written in JAX: github.com/google-research/d…
Dinosaur is a core component of NeuralGCM and we hope it is useful for the weather/climate research community.
Jamie Smith retweeted
We just finished a joint code release for CamP (camp-nerf.github.io/) and Zip-NeRF (jonbarron.info/zipnerf/). As far as I know, this code is SOTA in terms of image quality (but not speed) among all the radiance field techniques out there. Have fun! github.com/jonbarron/camp_zi…
Jamie Smith retweeted
It’s been a privilege to be part of the Gemini pretraining team and overall program, I’m so excited that the world can finally see what we’ve been up to for most of the past year:
tl;dr we’re so back storage.googleapis.com/deepm…
Jamie Smith retweeted
Satya Nadella walking into the Microsoft Monday Morning Meeting after acquiring OpenAI for $0 and 0 regulatory scrutiny
Jamie Smith retweeted
This year, Google's Research Scholar program for early-career professors is specifically solicitating proposals on large-language and multi-modal machine learning models for science:
research.google/outreach/res…
Applications will open next week and are due by the end of November.
Jamie Smith retweeted
Excited to announce our Deep Learning Tuning Playbook, a writeup of tips & tricks we employ when designing DL experiments. We use these techniques to deploy numerous large-scale model improvements and hope formalizing them helps the community do the same! github.com/google-research/t…
Jamie Smith retweeted
Come by our workshop on deep learning training algorithms tomorrow, "Has It Trained Yet?" hity-workshop.github.io/Neur…
We'll be in Theater B swapping tips and tricks about neural net training!
Jamie Smith retweeted
🚨 We've extended the deadline for the "Has It Trained Yet" Workshop @NeurIPSConf 2022!
Submit your papers by September 30, 2022. More info at hity-workshop.github.io/Neur…
This Post is from an account that no longer exists. Learn more
Jamie Smith retweeted
I'm happy to share a new project on using machine learning for computational fluid dynamics, led by @gideoknite:
"Learning to correct spectral methods for simulating turbulent flows"
arxiv.org/abs/2207.00556
Jamie Smith retweeted
In case you were wondering what paths 800 unsteered bicycles take when you push them:
Jamie Smith retweeted
It's only one minute long, but this CCTV footage has a better narrative and more compelling character arcs than most of the arthouse films I've seen:
Readers added context they thought people might want to know
As the original tweeter notes later in the thread, this video is not “CCTV footage” but performance art — “a Dhar Mann-style inspirational video by Cris Elmasry”
twitter.com/ponchorebound/…