չϢչȫ

Ϊʲô˵ͼ AI δ
Դ   ڣ2019-02-10 10:04:21   40090  

Ԫר ߣ٩ ༭ŷ Ԫͼ(Graph NN)ǽһоȵ㣬DeepMindGraph NetworksųѧϰʵƪĻɬѶǼϯAIѧҡҽﴴʼ˵٩ʿ廪ʿڽ...

Ϊʲô˵ͼ AI δ

Ԫר

ߣ٩

༭ŷ

Ԫͼ(Graph NN)ǽһоȵ㣬DeepMind“Graph Networks”ųѧϰʵƪĻɬѶǼϯAIѧҡҽﴴʼ˵٩ʿ廪ʿڽŶӶGNNĻϣDeepMind“ͼ”塣

- 1 -

ع 2018 ѧϰĽչ20186 DeepMind Ŷӷ

“Relational inductive biases, deep learning, and graph networks”

һƪҪģҵ顣

󣬺ܶѧǵ˼·оа廪ѧïŶӡ201812£һƪĿ“Graph neural networks: A review of methods and applications”

20191£ʿڽŶӣҲдһƪƪĸȫ棬Ŀ“A Comprehensive Survey on Graph Neural Networks”

Ϊʲô˵ͼ AI δ

ʿڽŶGNNԴarxiv

DeepMind ŶӵƪģҵôҵĹעԭ

Դ AlphaGo սʤhԺDeepMind ҵ磬ΪѧϰҵŶӣDeepMind Ŷӷģܵͬձע

ԴDeepMind Ŷӷ [1] Ժ󲻾ã Github ϿԴǿϵͳĿƽ Graph Nets [4]

⣺ͿԴҪDzDZҵҪԭҪԭ⣬DeepMind Ŷоǣѧϰͼס

- 2 -

ͼ (Graph) ɵ (Node) ͱ (Edge) ɡ

ͼһҪѧģͣܶ⡣

Ʃǰѳе·ͼͼףÿվһ㣬ڵĵվ֮߾DZߣ㵽յ㣬ǿͨͼ׵ļ㣬㵽յ㣬ʱ̡˴ٵг·ߡ

Ʃ Google Ͱٶȵ棬ÿվÿҳͼеһ㡣ÿҳӣվҳÿӶͼеһߡĸҳõԽ࣬˵ҳԽףǣҲԽǰ

ͼ׵IJȻд

Ʃ뼸ε˾н·ߣÿн·ǰʱеһʱ䡢GPSγȣ顣ΰѼн·ߣһ𣬹еͼ

ѵͼҲһͼףÿ·ڣһ㣬ڵ·ڣһߡ

òƺܼ򵥣ϸں鷳

ٸӣ·кܶʽʮ·ڣdzڣлε——δӶ·Уȷ·ڵλã

Ϊʲô˵ͼ AI δ

ձ챣ɽţȷŵλ

- 3 -

ѧϰͼףܹǶͼ׵Ĵ

ѧϰͼıĴ棬Ѿȡ˾޴ijɹѧϰijɹʹ֮Ӧͼ״

ͼɺƽֱؾɡһǶȣÿΪͼеһ㣬ÿصܱߵ 8 ֮䶼бߣÿ߶ȳͨӽǣͼͼǹͼ׵һ

ͼѧϰֶΣԸͷ棬ӦڹͼףƩ convolutionresidualdropoutpoolingattentionencoder-decoder ȵȡѧϰͼ״뷨ʵܼ򵥡

Ȼ뷨ܼ򵥣뵽ϸڣսÿսζŸǿļŸDZӦó

ѧϰͼ״оҵûͳһijν

ǿͼ׵ѧԵŶӣоΪ Geometric Deep LearningïŶӺʿŶӣǿͼ״еҪԣǿ˼ԴǰΪ Graph Neural NetworksDeepMind Ŷȴ԰ضֶΣʹøƣGraph Networks

ôҪַȥչȴҪѸѧɵĿ궨λͼڼǿ֮ͬ໥⣬ڴٽ֮ͬδ

- 4 -

ʿŶӰѧϰͼ״չ 5 ӷ򣬷dzö

Ϊʲô˵ͼ AI δ

ʿŶӰѧϰͼ״ 5 ӷԴ A Comprehensive Survey on Graph Neural Networks

Graph Convolution Networks

Graph Attention Networks

Graph Embedding

Graph Generative Networks

Graph Spatial-temporal Networks

˵ Graph Convolution Networks (GCNs)

Ϊʲô˵ͼ AI δ

GCN ܣԴ A Comprehensive Survey on Graph Neural Networks

GCN CNN ӦڹͼסCNN ҪΪĸ

֮ںϡͼ򣬵֮ںҪͨ (convolution) ʵ֡ڹͼ֮ĹϵñԣڹͼںϣбȾǿİ취Messsage passing [5] һָǿİ취

ֲCNN ʹ convolution İ취ԭʼؾУ߲ĵ㣬ǹĵ㣬ںԡںڵİ취ҲӦڹͼС

CNN ʹ pooling ֶΣԭʼУԵڱԵУʵʵУ߲ʵ塣CNN ͨ convolution pooling ʹãṹӣܸǿ硣ڹͼףҲڻ Messsage passing Poolingͼס

㡣CNN ͨʹ softmax ֶΣͼз࣬ʶͼ׵ںڹͼ˵ĽԶͼףȵȽҲԤͼijضĵֵҲԤijߵֵ

Ϊʲô˵ͼ AI δ

GCN Graph Attention Networks Դ A Comprehensive Survey on Graph Neural Networks

Graph Attention Networks Ҫ⣬ GCN ƣڵںϡķ

Graph Convolution Networks ʹþʽʵֵںϺͷֲConvolution ʽںڵĵ㣬 attention ۽ʽȴڵĵ㣬ÿںͼ㣬ǷڣǷںںϣȡڵ֮Ĺǿ

Attention ǿ󣬵ǶҪߣΪҪͼ֮Ĺǿ Graph Attention Networks оص㣬νͼɱͨм㣬߼Чʡ

- 5 -

Graph Embedding Ҫ⣬Ǹͼÿÿߣһֵͼ񲻴⣬Ϊֵǣıִʻ乹ɣҪִʻ㣬תֵʹѧϰ㷨

ıеÿֻʻ㣬ͼеһ㣬ͬʱѴ֮﷨ϵͼеһߣôͶ䣬͵ͬıͼеһн·

ܹÿֺʹʻ㣬һеֵôͶӦн··

жʵ Graph Embedding İ취ЧȽϺõİ취 Autoencoder GCN İ취ͼ׵ĵͱתֵ̳Ϊ (encoding)Ȼ֮ͨľ룬ֵϣתΪͼף̳Ϊ (decoding)ͨϵصΣýõͼףԽԽԭʼͼף̳Ϊѵ

Graph Embedding ͼеÿÿߣеֵͼ׵Ľṹ⡣

ͼн·δЩн·УʶЩЩ֮ߣѶȸǣûн·ѵͼ׵ľֲԼ֮Ӧͼ׵ԣΰѾֲƴӳͼȫòЩ Graph Generative Networks Ҫ⡣

Graph Generative Networks ȽDZʵַʹ Generative Adversarial Networks (GAN)

GAN (generator) ͱ (discriminator) ֹɣ1. ѵУƩ纣н·²ݱͼӦóʲô2. ɳͼףαһн·3. Ӵα·ʵ·Уѡ·ñʶļ·αġ

ɵɵֲ˭α·˭ʵ·˵ɳͼףܽӽʵͼס

Ϊʲô˵ͼ AI δ

GCN 4 ͼ磬Դ A Comprehensive Survey on Graph Neural Networks

- 6 -

Ծ̬ͼ׵⣬ͼʱǶ̬ģƩͼбֵĵ·Ǿ̬ģ·Ƕ̬ġ

Ԥⴺڼ䣬찲ŸĽͨӵ⣬Ҫǿռ spatial أƩ찲ܱߵĵ·ṹҲҪʱ temporal أƩ괺ڼõͨӵ Graph Spatial-temporal Networks Ҫ֮һ

Graph Spatial-temporal Networks ܽ⣬ƩһƵÿһ֡ͼУʶλãѵڣƵijЩ֡УпǿģƩ类Աڵˡ

ʱͨ˼· RNN LSTM GRU ȵȡ

DeepMind Ŷ RNN ϣ˱ͽ (encoder-decoder) ơ

- 7 -

DeepMind Ŷӵƪ[1]ԼĹ“part position paper, part review, and part unification”᰸ںϡ⻰ô⣿

Ϊʲô˵ͼ AI δ

DeepMindϹȸԡMITȻ27λ߷ذģ“ͼ”Graph network˵ѧϰϣѧϰ޷йϵ⡣

ǰ˵ʿŶӰѧϰͼ״չ 5 ӷ1) Graph Convolution Networks2) Graph Attention Networks3) Graph Embedding4) Graph Generative Networks5) Graph Spatial-temporal Networks

DeepMind Ŷ 5 ӷ 4 򣬷ֱ Graph Attention NetworksGraph EmbeddingGraph Generative Networks Graph Spatial-temporal Networksǰĸijɹ“ں”ͳһĿܣΪ Graph Networks

ǵУĸӷ;ɹ“”Dzû Graph Convolution Networks ijɹȻǴĸӷɹУѡΪDZķγԼ“᰸”ǿԴĴ [4]

DeepMind201810¿ԴGraph Nets libraryTensorFlowй򵥶ǿĹϵ硣Դgithub.com/deepmind/graph_nets

ȻУǵ᰸ĸӷ⣬Dz鿴ǿԴĴ룬ʵǺӷGraph Attention Networks Graph Spatial-temporal Networks

DeepMind ˼·ģȣ [5] message passing ںϵĻƣ [6] ͼȫֵľ۽ϣͨõ graph block ģ飻Σ LSTM Ҫڽ encoder-decoder ܣʱлƣ󣬰 graph block ģڽ encoder-decoder ܣγ Graph Spatial-temporal Networks ͨϵͳ

- 8 -

Ϊʲô DeepMind ijɹҪ¹ļ¡

һѧϰ̵Ľ

ԭϽѧϰƩ CNN ijɹڶͼIJϳҲǣԭʼؾУ߶Ρβ߶Уʵʵ壬ʵ塣

ǣ̽ CNN ÿһмʵϺȷһЩڵ㣬Ҳ֪һЩڵ㣬ʵ塣֮ܶCNN ṹǸԣ޷ȷؽṹصĹ̵ϸڡ

޷͹̵ϸڣҲ̸ΪԤ CNN ⣬ֻѵѵĽǷܴﵽڴЧ޷ϡº«ưȱݣȴȱݡ

˵ȷظ CNN ̵ϸڣͿԵصεĸڵIJΪ׼Ԥ

Сѧϰ

ѧϰѵݣѵݵĹģܴͨ򣬶󼸰򡣴ռôѵݣҪ֯ȥѵݽбעǾ޴ս

ѧϰĹϸڣи˽⣬ǾͿԸƾøٵѵݣѵɵѧϰģ͡

Ḷ́Ḷ́ڵĵ㣬һ©ز׵ؽо

ǶԵ֮Ĺϵиȷ˽⣬ͲҪڵĵ㣬һ©ز׵ؽоֻҪйĵ㣬о

ݵ֮Ĺϵ磬ǹͼסͼ׵Ľṹͨ CNN Ӽ򵥣ԣҪѵҲ١

Ǩѧϰ

õ CNNԴӴͼƬУʶijʵ壬Ʃè

ǣʶè CNN ʶèʶ𹷣Ҫʶ𹷵ѵݡǨѧϰĹ̡

ܲܲṩʶ𹷵ѵݣֻùķʽߵè빷Ȼõʶ𹷣Ŀꡣ

ѧϰи׼˽⣬֪ܰʶ͹ڽѧϰ

ӹ巶Χ˵ѧϰ֪ʶͼףǻѧϰӪѧɵѧɡΪֹѧɸ󣬸ʤںѧɣȡ̣ѧܾõ⡣ѧϰ쵽ͼ״ѧɵںϣϣ

ġռʱںϣں

Ƶ˵ѧϰ߾硣

ƵںͼĿռָͼʵʶʵӦ⡣

̬֡ͼһγƵʵʱСͬһʵ壬ڲͬ֡λã̺ʵ˶˶ıɺ

δһƵܽı⡣߷θһı⣬ҵеƵƵľҲѶȳ

ο

Relational inductive biases, deep learning, and graph networkshttps://arxiv.org/abs/1806.01261

Graph neural networks: A review ofmethods and applicationshttps://arxiv.org/abs/1812.08434

A Comprehensive Survey on Graph Neural Networkshttps://arxiv.org/abs/1901.00596

Graph netshttps://github.com/deepmind/graph_nets

Neural message passing for quantum chemistryhttps://arxiv.org/abs/1704.01212

Non-local neural networkshttps://arxiv.org/abs/1711.07971


վ

ȴʣ AI ˹

չ

ĿHotCates

Copyright © 2010-2024 AiLab Team. ˹ʵ Ȩ    | ϵ | | ˾̬ | | ˽ | | չ