Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
PHH_Graph_Representation_Learning_11092020
Search
phanhoang17
September 11, 2020
Research
0
31
PHH_Graph_Representation_Learning_11092020
phanhoang17
September 11, 2020
Tweet
Share
More Decks by phanhoang17
See All by phanhoang17
Building a simple RS for Viblo website
huyhoang17
0
50
Introduction to TensorFlow
huyhoang17
0
120
Deep Learning - Conceptual Understanding and Applications
huyhoang17
0
62
Pandas for Data Analysis
huyhoang17
0
180
Other Decks in Research
See All in Research
業界横断 副業・兼業者の実態調査
fkske
0
220
能動適応的実験計画
masakat0
2
780
問いを起点に、社会と共鳴する知を育む場へ
matsumoto_r
PRO
0
580
Google Agent Development Kit (ADK) 入門 🚀
mickey_kubo
2
1.6k
Adaptive fusion of multi-modal remote sensing data for optimal sub-field crop yield prediction
satai
3
250
数理最適化と機械学習の融合
mickey_kubo
16
9.2k
一人称視点映像解析の最先端(MIRU2025 チュートリアル)
takumayagi
6
3.3k
Mechanistic Interpretability:解釈可能性研究の新たな潮流
koshiro_aoki
1
400
EarthSynth: Generating Informative Earth Observation with Diffusion Models
satai
3
180
大規模な2値整数計画問題に対する 効率的な重み付き局所探索法
mickey_kubo
1
330
Agentic AIとMCPを利用したサービス作成入門
mickey_kubo
0
430
Adaptive Experimental Design for Efficient Average Treatment Effect Estimation and Treatment Choice
masakat0
0
100
Featured
See All Featured
Rails Girls Zürich Keynote
gr2m
95
14k
Documentation Writing (for coders)
carmenintech
73
5k
The Straight Up "How To Draw Better" Workshop
denniskardys
236
140k
CSS Pre-Processors: Stylus, Less & Sass
bermonpainter
358
30k
Learning to Love Humans: Emotional Interface Design
aarron
273
40k
The Psychology of Web Performance [Beyond Tellerrand 2023]
tammyeverts
49
3k
BBQ
matthewcrist
89
9.8k
It's Worth the Effort
3n
187
28k
Exploring the Power of Turbo Streams & Action Cable | RailsConf2023
kevinliebholz
34
6k
Fantastic passwords and where to find them - at NoRuKo
philnash
51
3.4k
Site-Speed That Sticks
csswizardry
10
780
How to train your dragon (web standard)
notwaldorf
96
6.2k
Transcript
Graph Representation Learning An introduction Phan Hoang - AI Research
Team
Do we need deep graph neural network?
Graph Representation Learning?!
Graph Representation Learning?! - Viblo recommender engine - Graph-based recommender
system for paper citation network - https://api.semanticscholar.org/v1/paper/arXiv:2006.07739 - https://jsoneditoronline.org/#left=cloud.f3946e860ca54e98934899f6f30eb475 - https://www.connectedpapers.com/
Graph Representation Learning?!
Graph Representation Learning?!
Prerequisite knowledge - https://viblo.asia/p/6J3ZgP0qlmB#_1-so-ly-thuyet-do-thi-co-ban-1
Tasks - https://viblo.asia/p/6J3ZgP0qlmB#_1-so-bai-toan-dien-hinh-2 - Node Classification - Link Prediction -
Graph Clustering & community detection - ...
Node Embedding - Random Walk / DeepWalk - Word2Vec
Node Embedding - Random Walk
Node Embedding - Skip-gram model
DeepWalk - karate dataset
Node Embedding - Cons - Unseen node? - Node feature?
- Tranductive learning vs inductive learning
Graph Neural Network - an introduction
Example of GCN model
Example of GCN model
Graph Convolution Network - Kipf (2016)
GraphSage - L.Hamilton (2017)
GraphSage - L.Hamilton (2017)
GraphSage - inductive learning
The number of layers
GraphSage - L.Hamilton (2017) - Loss function, un-supervised learning
How to apply?
Real use-case applications - PinSage / UberEat
Real use-case applications - PinSage / UberEat - Based on
GraphSage paper (Haminton / 2017) - Heterogeneous graph / bipartite graph - Node Embedding: image + text feature, user info - Edge: Pin-Board, User-Dish, User-Restaurant relation
Real use-case applications - Decagon
Real use-case applications - Decagon
Real use-case applications - Goal-directed generation
How to model the GNN for a specific dataset?
Task: Text Classification
Task: Relation Extraction / NLP
Task: Key-Information Extraction
Task: Recommender System
Other tasks - 3D Object Detection - Action Recognition /
Pose Estimation - GAN / VAE - GAT / Graph Transformer - Feature Matching - Key Information Extraction - Scene Graph Generation - Recommender System - ...
Papers - Cluster-GCN / GraphSAINT - graph/sub-graph/node sampling - GAT
(graph attention network) - GIN (graph isomorphism network) - Deep-GCN / Deeper-GCN - going deeper with GCN - ...
Cons - non-injective aggregate function
Cons - adversarial attack
Thank you! https://viblo.asia/p/deep-learning-graph-neural-network-a-literature-review-and-applications-6J3ZgP0qlmB