-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
DKN: Deep Knowledge-Aware Network for News Recommendation, Wang+, WWW'18 #363
Comments
OverviewContents-basedな手法でCTRを予測しNews推薦。newsのタイトルに含まれるentityをknowledge graphと紐づけて、情報をよりリッチにして活用する。 CNNのinputを[[word_ embedding], [entity embedding], [contextual entity embedding]](画像のRGB)のように、multi-channelで構成し3次元のフィルタでconvolutionすることで、word, entity, contextual entityを表現する空間は別に保ちながら(同じ空間で表現するのは適切ではない)、wordとentityのalignmentがとれた状態でのrepresentationを獲得する。 ExperimentsBingNewsのサーバログデータを利用して評価。 word/entity embeddingの次元は100, フィルタのサイズは1,2,3,4とした。loss functionはlog lossを利用し、Adamで学習した。 DeepFM超えを達成。 |
#365 によって経験的にRNN, Recursive Neural Network等と比較して、sentenceのrepresentationを獲得する際にCNNが優れていることが示されているため、CNNでrepresentationを獲得することにした模様(footprint 7より) |
Factorization Machinesベースドな手法(LibFM, DeepFM)を利用する際は、TF-IDF featureと、averaged entity embeddingによって構成し、それをuser newsとcandidate news同士でconcatしてFeatureとして入力した模様 |
content情報を一切利用せず、ユーザのimplicit feedbackデータ(news click)のみを利用するDMF(Deep Matrix Factorization)の性能がかなり悪いのもおもしろい。やはりuser-item-implicit feedbackデータのみだけでなく、コンテンツの情報を利用した方が強い。 |
(おそらく)著者によるtensor-flowでの実装: https://github.com/hwwang55/DKN |
https://arxiv.org/pdf/1801.08284.pdf
The text was updated successfully, but these errors were encountered: