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

LLM-Rec: Personalized Recommendation via Prompting Large Language Models, Hanjia Lyu+, N/A, arXiv'23 #911

Open
AkihikoWatanabe opened this issue Aug 2, 2023 · 1 comment

Comments

@AkihikoWatanabe
Copy link
Owner

AkihikoWatanabe commented Aug 2, 2023

URL

Affiliations

  • Hanjia Lyu, N/A
  • Song Jiang, N/A
  • Hanqing Zeng, N/A
  • Yinglong Xia, N/A
  • Jiebo Luo, N/A

Abstract

  • We investigate various prompting strategies for enhancing personalizedcontent recommendation performance with large language models (LLMs) throughinput augmentation. Our proposed approach, termed LLM-Rec, encompasses fourdistinct prompting strategies: (1) basic prompting, (2) recommendation-drivenprompting, (3) engagement-guided prompting, and (4) recommendation-driven +engagement-guided prompting. Our empirical experiments show that combining theoriginal content description with the augmented input text generated by LLMusing these prompting strategies leads to improved recommendation performance.This finding highlights the importance of incorporating diverse prompts andinput augmentation techniques to enhance the recommendation capabilities withlarge language models for personalized content recommendation.

Translation (by gpt-3.5-turbo)

  • 私たちは、大規模言語モデル(LLMs)を用いたパーソナライズされたコンテンツ推薦の性能向上のためのさまざまなプロンプティング戦略を調査しました。提案されたアプローチであるLLM-Recは、4つの異なるプロンプティング戦略を含んでいます:(1)基本的なプロンプティング、(2)推薦に基づくプロンプティング、(3)エンゲージメントに基づくプロンプティング、および(4)推薦に基づく+エンゲージメントに基づくプロンプティング。実証実験の結果、これらのプロンプティング戦略によって生成されたLLMによる拡張入力テキストと元のコンテンツの説明を組み合わせることで、推薦の性能が向上することが示されました。この結果は、多様なプロンプトと入力拡張技術を大規模言語モデルに取り入れることが、パーソナライズされたコンテンツ推薦の能力を向上させる上で重要であることを強調しています。

Summary (by gpt-3.5-turbo)

  • LLMsを用いたパーソナライズされたコンテンツ推薦のためのプロンプティング戦略を調査し、LLM-Recというアプローチを提案した。実験の結果、プロンプティング戦略によって生成されたLLMによる拡張入力テキストと元のコンテンツの説明を組み合わせることで、推薦の性能が向上することが示された。これは、多様なプロンプトと入力拡張技術がパーソナライズされたコンテンツ推薦の能力を向上させる上で重要であることを示している。
@AkihikoWatanabe AkihikoWatanabe changed the title LLM-Rec: Personalized Recommendation via Prompting Large Language Models, Hanjia Lyu+, N/A, arXiv'23 Aug 2, 2023
@AkihikoWatanabe
Copy link
Owner Author

AkihikoWatanabe commented Aug 2, 2023

LLMのpromptingの方法を変更しcontent descriptionだけでなく、様々なコンテキストの追加(e.g. このdescriptionを推薦するならどういう人におすすめ?、アイテム間の共通項を見つける)、内容の拡張等を行いコンテントを拡張して活用するという話っぽい。WIP

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant