How Graph Neural Networks Are Transforming Industries

🔑 Get your AssemblyAI API key here: www.assemblyai.com/?...
Graph Neural Networks (GNN) have been rapidly advancing and have recently become the dark horse behind many exciting achievements, from research breakthroughs to actual solutions with large-scale deployment. But why do Graph Neural Networks (GNN) matter in 2024?
This video will recap some impactful applications of GNNs in science and industry, and the most recent research highlights, focusing on applied use cases. With the latest applications of GNNs spanning from Relational Databases, Materials Science, Dynamical System Modeling, to the most advanced Recommendation Systems out there, we take an overview of diverse application areas unlocked by Graph Machine Learning (Graph ML).
For all references made in the video, see below.
▬▬▬▬▬▬▬▬▬▬▬▬ CONNECT ▬▬▬▬▬▬▬▬▬▬▬▬
🖥️ Website: www.assemblyai.com/?...
🐦 Twitter: / assemblyai
🦾 Discord: / discord
▶️ Subscribe: kzbin.info?...
🔥 We're hiring! Check our open roles: www.assemblyai.com/careers
▬▬▬▬ REFERENCES AND SOURCES ▬▬▬▬
Uber Eats (Uber) - www.uber.com/en-GR/blog/uber-...
Pinsage (Pinterest) - arxiv.org/abs/1806.01973
Google Maps (DeepMind) - deepmind.google/discover/blog...
GraphCast (DeepMind) - deepmind.google/discover/blog...
Relational Databases (Stanford/KumoAI/MPII) - relbench.stanford.edu/paper.pdf
GNoME (DeepMind) - deepmind.google/discover/blog...
Antibiotic Discovery / Halicin (MIT) - www.sciencedirect.com/science...
Explainable GNNs (MIT/Harvard) - www.nature.com/articles/s4158...
RFdiffusion (Baker Lab) - www.bakerlab.org/2023/07/11/d...
Chroma (Generate:Biomedicines) - github.com/generatebio/chroma
Paper on Non-Euclidean Structure Data - www.mdpi.com/2073-8994/15/4/837
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
0:00 Intro
0:44 What are Graph Neural Networks?
1:45 Recommendation Systems
3:19 Traffic Prediction
3:53 Weather Prediction
4:44 Data Mining
6:35 Materials Science
8:02 Drug Discovery
10:05 Protein Design
11:41 Final Words
#machinelearning #deeplearning #ai

[@samurai-code]

Amazing video, one of the best intro to GNNs on YT. Thank you!

[@AssemblyAI]

Glad you liked it!

[@adrienforbu5165]

Nice video ! One could also talk about the incorporation of the powerful transformers architecture into GNN

[@NielsSwimberghe]

Fantastic video. So much information!

[@AssemblyAI]

Glad it was helpful!

[@MarkSimithraaratchy]

Excellent video. I hadn't come across the Relational Deep Learning' paper where data-mining use case of taking relational tables as input. Genius idea.

[@AssemblyAI]

Thank you! Indeed, that research came out just a couple of months ago and may unlock so many exciting applications because of its great generality and applicability!

[@graziamarchese692]

Excellent! Very clear, very much well conceived and structured; understendable also by non-experts.

[@Stewz66]

Thank you. I have a node/edge graph problem and i need to get smart on GNN in a hurry.🙏🏽

[@AssemblyAI]

You're welcome and good luck!

[@yosconisi]

Thanks, a good introduction! A friendly comment, there is some dirt on your camera. Looking forward to studying GNNs

[@AssemblyAI]

Thanks for watching!

[@user-du8hf3he7r]

I found the dirt on the camera in the center of the screen very distracting. Perhaps you can use AI to remove it 😀

[@samurai-code]

don't be dramatic, it's a snot not a dirt

[@thesimplicitylifestyle]

😎🤖

[@art-creator]

Graphs can be easily described with sparse matrices, so GNN are not something very special. More tricks are about dataset preparations for their tuning.

[@alexanderlevakin9001]

If you know good video on explaining what GraphNN are, please share. This video requires to know the basics of GNN.