Research Engineer Dmitrii Kochkov's work contributed to the development of a hybrid AI and differentiable solver approach for atmospheric modeling. 🌡⛈️ → https://goo.gle/3Yb9OMf & →https://goo.gle/4cUeif0 Learn more about the groundbreaking work from the NeuralGCM team at Google Research, and explore how hybrid models have the potential to improve modeling of weather and climate.
Google Research’s Post
More Relevant Posts
-
We humans resist change but once we accept it why do we normalize it so fast? Like we now see it as totally normal that if you organize rocks in a specific way and pass electricity between them we can make them think and draw images at this quality??? Just some random thoughts while reading papers on early-fusion multimodal models. The image is from Meta's Chameleon paper, which I think is a little sneak-peak of future Llama models.
To view or add a comment, sign in
-
NED'28 || Electronics Engineering | Python, Graphic Design & Frontend Dev | Exploring TypeScript & Next.js
Today I earned my "Classify types of space rocks in random photos by using artificial intelligence" badge!🎉 I’m so proud to be celebrating this achievement and hope this inspires you to start your own Microsoft Learn journey!❤ #PythonAI #ML #DataScience #AIwithPython #MachineLearning #DataAnalysis #PythonDeveloper #AIEngineer #DataScientist #ai #machinelearning #artificialintelligence #deeplearning #neuralnetworks #dataanalysis #datamining #microsoftlearn
To view or add a comment, sign in
-
📍Come at poster board A58, Hall A 🟥, on Thursday morning for further insights on this great work led by Arash Rahi. 📊Topic: modeling #runoff coefficient - a key indicator of #freshwater availability - through #LSTM. #AI #deeplearning #hydrology #EGU2024
A preview of my poster for the upcoming #EGU24. Excited to discuss AI algorithms? Discover our innovative approach using Discrete Wavelet Transform (DWT) combined with Long Short-Term Memory (LSTM) to model the noisy runoff coefficient (Rc)💧 and address the impact of noise on modeling accuracy. Learn how DWT helps overcome this challenge in our Germany case study. Join me at session HS3.4: "Deep Learning in Hydrology" in Hall A 🟥, second floor, poster board A.58, on Thursday, April 18th, 2024, from 10:45 to 12:30. Let's delve into the intersection of AI and hydrology! See you there!
To view or add a comment, sign in
-
I'm really excited to share this news article about BLASTNet, the first large dataset for fundamental fluid dynamics. 🔺This dataset can enable machine learning models to tackle complex problems in fluid flows, such as turbulence, combustion, and multiphase flows. These problems are relevant for many scientific and engineering domains, such as biomedical science, physics, energy, and neuroscience. 🔺I believe this dataset can lead to exponential breakthroughs in these fields and help us better understand the natural phenomena around us. 👨🏾💻Check out the article and the dataset on GitHub. Kudos to the Stanford HAI team for creating this amazing resource! Shout out to Paul Golding for finding this.
By collecting data from the field of computational fluid dynamics into a single dataset, AI researchers at Stanford hope to do for rocket science, oceanography, and climate modeling what web-scale data did for language. https://lnkd.in/gtdqRsea
BLASTNet – The First Large Machine Learning Dataset for Fundamental Fluid Dynamics
hai.stanford.edu
To view or add a comment, sign in
-
The future is NeuroSymbolic. The inefficiency of linear approximation by brute force is just too overwhelming. Trying to multiply that inefficiency by using a system that only uses the most inefficient information encoding, namely "Human Language", only amplifies that problem. Its like inefficiency on top of inefficiency. This is why intelligence or autonomous systems will get more and more out of reach, the more people try to solve it end-to-end gradient-based. The future is intelligently designed cognitive architectures.
As we have spoken about in great detail on MLST over the years, basic NN architectures have severe computational limitations. They are not Turing machines, they can not perform the types of computation a simple calculator can perform. But what if we could have our cake and eat it? The future might just be neurosymbolic! Dr. Petar Veličković is a rising star scientist at DeepMind and is taking his Geometric Deep Learning framework to the next level in collaboration with his collegues. This is just a teaser clip but there will be lots more info coming.
To view or add a comment, sign in
-
The Frontiers journal reacts to the AI rat image scandal In this episode, we discuss.. ✅Published papers disappearing from the internet ✅AI-generated image of rat retracted ✅Climate scientists modelling the Planet of Dune 😱 ✅+ much more 😁 Listen to episode 77: (Link in the comments)
To view or add a comment, sign in
-
We're thrilled to share a groundbreaking piece of research we've come across - an innovative approach to applying deep reinforcement learning in the complex domain of fluid dynamics. This study unveils a system, dubbed "Box o Flows," designed to navigate the intricacies of fluid behaviors in the physical world, a challenge often deemed too complex for accurate simulation. What stands out to us is the system's ability to derive intricate behaviors from straightforward rewards, highlighting the potential of offline learning to maximize data utility. This research resonates deeply with our ethos at Pivot-al, where we're passionate about harnessing the transformative power of data to solve real-world problems. Check out the full study here https://lnkd.in/exRniXSb
Real-World Fluid Directed Rigid Body Control via Deep Reinforcement Learning
arxiv.org
To view or add a comment, sign in
-
As we have spoken about in great detail on MLST over the years, basic NN architectures have severe computational limitations. They are not Turing machines, they can not perform the types of computation a simple calculator can perform. But what if we could have our cake and eat it? The future might just be neurosymbolic! Dr. Petar Veličković is a rising star scientist at DeepMind and is taking his Geometric Deep Learning framework to the next level in collaboration with his collegues. This is just a teaser clip but there will be lots more info coming.
To view or add a comment, sign in
-
Proud to have reached 100 citations in my first 3 years of publications! The first of many milestones. Next, the PhD. P.S.: the 100th citation was from "A Study of Current Socio-Technical Design Practices in the Industry 4.0 Context among Small, Medium, and Large Manufacturers in Minnesota and North Dakota", a pleasant sign of how our research reaches beyond the field of AI.
To view or add a comment, sign in
-
Hi! Here is my first talk at the University of Potsdam in the Seminar of AI in Software Engineering. The presentation is based in the article "Is Deep Learning Good Enough in SDP?". The focus is on two CNN models: SqueezeNet and Bottleneck. The authors conducted a detailed comparative study across seven datasets from the NASA repository, comparing the performance of these models against baseline models. Paper's link: https://lnkd.in/e88-rp_v
To view or add a comment, sign in
196,562 followers
PhD Candidate @ Sorbonne Université
1moIncredible work ! Is the team currently looking for junior researchers ?