Graph Attention Networks (GAT) | GNN Paper Explained
In this video, I do a deep dive into the graph attention network paper!
GATs have a lot in common with transformers a reason more to keep an eye out for them!
You'll learn about:
✔️ Basic graph theory
✔️ All the nitty-gritty details behind GAT
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
✅ GAT website: https://petar-v.com/GAT/
✅ GAT paper: https://arxiv.org/abs/1710.10903
✅ M. Bronstein's blog: https://towardsdatascience.com/do-we-need-deep-graph-neural-networks-be62d3ec5c59
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
⌚️ Timetable:
00:00 - A note on geometric deep learning
00:50 - Graph theory basics
08:35 - Intro to GATs (related work)
10:15 - A detailed explanation of the method
16:20 - A multi-head version of the GAT
19:05 - Visualizations, spatial pooling, GNN depth
21:20 - A recap of GAT properties
23:35 - Receptive field of spatial GNNs
24:35 - Datasets, transductive vs inductive learning
30:35 - Results on transductive/inductive benchmarks
35:30 - Representations visualization (t-SNE)
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💰 BECOME A PATREON OF THE AI EPIPHANY ❤️
If these videos, GitHub projects, and blogs help you,
consider helping me out by supporting me on Patreon!
The AI Epiphany ► https://www.patreon.com/theaiepiphany
One-time donation:
https://www.paypal.com/paypalme/theaiepiphany
Much love! ❤️
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💡 The AI Epiphany is a channel dedicated to simplifying the field of AI using creative visualizations and in general, a stronger focus on geometrical and visual intuition, rather than the algebraic and numerical "intuition".
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
👋 CONNECT WITH ME ON SOCIAL
(I'm mostly active on LinkedIn and Twitter)
LinkedIn ► https://www.linkedin.com/in/aleksagordic/
Twitter ► https://twitter.com/gordic_aleksa
Medium ► https://gordicaleksa.medium.com/
Instagram ► https://www.instagram.com/aiepiphany/
Facebook ► https://www.facebook.com/aiepiphany/
FOLLOW ME ON GITHUB FOR COOL PROJECTS:
GitHub ► https://github.com/gordicaleksa
Connect with Petar Veličković:
LinkedIn ► https://www.linkedin.com/in/petarvelickovic/
Twitter ► https://twitter.com/PetarV_93
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
#graphs #attention #deeplearning
Видео Graph Attention Networks (GAT) | GNN Paper Explained канала The AI Epiphany
GATs have a lot in common with transformers a reason more to keep an eye out for them!
You'll learn about:
✔️ Basic graph theory
✔️ All the nitty-gritty details behind GAT
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
✅ GAT website: https://petar-v.com/GAT/
✅ GAT paper: https://arxiv.org/abs/1710.10903
✅ M. Bronstein's blog: https://towardsdatascience.com/do-we-need-deep-graph-neural-networks-be62d3ec5c59
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
⌚️ Timetable:
00:00 - A note on geometric deep learning
00:50 - Graph theory basics
08:35 - Intro to GATs (related work)
10:15 - A detailed explanation of the method
16:20 - A multi-head version of the GAT
19:05 - Visualizations, spatial pooling, GNN depth
21:20 - A recap of GAT properties
23:35 - Receptive field of spatial GNNs
24:35 - Datasets, transductive vs inductive learning
30:35 - Results on transductive/inductive benchmarks
35:30 - Representations visualization (t-SNE)
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💰 BECOME A PATREON OF THE AI EPIPHANY ❤️
If these videos, GitHub projects, and blogs help you,
consider helping me out by supporting me on Patreon!
The AI Epiphany ► https://www.patreon.com/theaiepiphany
One-time donation:
https://www.paypal.com/paypalme/theaiepiphany
Much love! ❤️
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
💡 The AI Epiphany is a channel dedicated to simplifying the field of AI using creative visualizations and in general, a stronger focus on geometrical and visual intuition, rather than the algebraic and numerical "intuition".
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
👋 CONNECT WITH ME ON SOCIAL
(I'm mostly active on LinkedIn and Twitter)
LinkedIn ► https://www.linkedin.com/in/aleksagordic/
Twitter ► https://twitter.com/gordic_aleksa
Medium ► https://gordicaleksa.medium.com/
Instagram ► https://www.instagram.com/aiepiphany/
Facebook ► https://www.facebook.com/aiepiphany/
FOLLOW ME ON GITHUB FOR COOL PROJECTS:
GitHub ► https://github.com/gordicaleksa
Connect with Petar Veličković:
LinkedIn ► https://www.linkedin.com/in/petarvelickovic/
Twitter ► https://twitter.com/PetarV_93
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
#graphs #attention #deeplearning
Видео Graph Attention Networks (GAT) | GNN Paper Explained канала The AI Epiphany
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
Graph SAGE - Inductive Representation Learning on Large Graphs | GNN Paper ExplainedGraph Attention Network Project WalkthroughCS480/680 Lecture 19: Attention and Transformer NetworksAlphaGo - Mastering the game of Go with deep neural networks and tree search | RL Paper ExplainedGraph Convolutional Networks (GCN) | GNN Paper ExplainedSpiking Neural Networks for More Efficient AI AlgorithmsC5W3L07 Attention Model IntuitionGraph Node Embedding Algorithms (Stanford - Fall 2019)Overfitting in a Neural Network explainedAI Language Models & Transformers - ComputerphileDeep learning on graphs: successes, challenges, and next steps | Graph Neural NetworksGoogle DeepMind's AlphaFold 2 explained! (Protein folding, AlphaFold 1, a glimpse into AlphaFold 2)EfficientNetV2 - Smaller Models and Faster Training | Paper explainedUnderstanding Graph Attention NetworksBatch Normalization (“batch norm”) explainedPytorch Geometric tutorial: Graph attention networks (GAT) implementationNeural Network Learns to Play SnakeHow to get started with Graph ML? (Blog walkthrough)Temporal Graph Networks (TGN) | GNN Paper ExplainedVision Transformer (ViT) - An image is worth 16x16 words | Paper Explained