Skip to content
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

STUDY: Socially Aware Temporally Casual Decoder Recommender Systems, Eltayeb Ahmed+, N/A, arXiv'23 #728

Open
AkihikoWatanabe opened this issue Jun 16, 2023 · 0 comments
Labels

Comments

@AkihikoWatanabe
Copy link
Owner

AkihikoWatanabe commented Jun 16, 2023

URL

Affiliations

  • Eltayeb Ahmed, N/A
  • Diana Mincu, N/A
  • Lauren Harrell, N/A
  • Katherine Heller, N/A
  • Subhrajit Roy, N/A

Abstract

  • With the overwhelming amount of data available both on and offline today,recommender systems have become much needed to help users find items tailoredto their interests. When social network information exists there are methodsthat utilize this information to make better recommendations, however themethods are often clunky with complex architectures and training procedures.Furthermore many of the existing methods utilize graph neural networks whichare notoriously difficult to train. To address this, we propose Socially-awareTemporally caUsal Decoder recommender sYstems (STUDY). STUDY does jointinference over groups of users who are adjacent in the social network graphusing a single forward pass of a modified transformer decoder network. We testour method in a school-based educational content setting, using classroomstructure to define social networks. Our method outperforms both social andsequential methods while maintaining the design simplicity of a singlehomogeneous network that models all interactions in the data. We also carry outablation studies to understand the drivers of our performance gains and findthat our model depends on leveraging a social network structure thateffectively models the similarities in user behavior.

Translation (by gpt-3.5-turbo)

  • 現在、オンラインおよびオフラインで利用可能な膨大なデータ量に直面して、ユーザーが興味に合わせたアイテムを見つけるために必要なレコメンドシステムが必要とされています。ソーシャルネットワーク情報が存在する場合、この情報を利用してより良い推薦を行う方法がありますが、これらの方法は複雑なアーキテクチャやトレーニング手順を持つことが多く、グラフニューラルネットワークを利用するものも多く、トレーニングが困難であることが知られています。この問題に対処するために、本研究ではSocially-aware Temporally caUsal Decoder recommender sYstems (STUDY)を提案します。STUDYは、修正されたトランスフォーマーデコーダーネットワークの単一のフォワードパスを使用して、ソーシャルネットワークグラフ上で隣接するユーザーグループ全体に対して共同推論を行います。本研究では、学校教育コンテンツの設定で、教室の構造を使用してソーシャルネットワークを定義し、提案手法をテストしました。結果、提案手法は、ソーシャルおよびシーケンシャルな方法を上回り、すべての相互作用をモデル化する単一の均質ネットワークの設計の簡素さを維持します。また、アブレーション研究を実施して、パフォーマンス向上の要因を理解し、ユーザーの行動の類似性を効果的にモデル化するソーシャルネットワーク構造を活用することがモデルの成功に重要であることがわかりました。

Summary (by gpt-3.5-turbo)

  • 本研究では、膨大なデータ量に直面する中で、ソーシャルネットワーク情報を利用したレコメンドシステムの提案を行いました。提案手法であるSTUDYは、修正されたトランスフォーマーデコーダーネットワークを使用して、ソーシャルネットワークグラフ上で隣接するユーザーグループ全体に対して共同推論を行います。学校教育コンテンツの設定で、教室の構造を使用してソーシャルネットワークを定義し、提案手法をテストした結果、ソーシャルおよびシーケンシャルな方法を上回り、単一の均質ネットワークの設計の簡素さを維持しました。また、アブレーション研究を実施して、ユーザーの行動の類似性を効果的にモデル化するソーシャルネットワーク構造を活用することがモデルの成功に重要であることがわかりました。
@AkihikoWatanabe AkihikoWatanabe changed the title STUDY: Socially Aware Temporally Casual Decoder Recommender Systems, Eltayeb Ahmed+, N/A, arXiv'23 Jun 16, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant