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

Predicting Search Intent Based on Pre-Search Context(SIGIR2015) #3

Open
kirinsannnnnnnnnn opened this issue Aug 14, 2019 · 0 comments

Comments

@kirinsannnnnnnnnn
Copy link
Collaborator

matome information

  • matome author: Yuta Tsuchizawa
  • read date: 2019/08/15

Predicting Search Intent Based on Pre-Search Context(SIGIR2015)

paper information

  • authors
    • Weize Kong: Center for Intelligent Information Retrieval, University of Massachusetts Amherst
    • Rui Li: Yahoo! Labs
    • Luo Jie: Yahoo! Labs
    • Aston Zhang: Department of Computer Science, University of Illinois at Urbana-Champaign
    • Yi Chang: Yahoo! Labs
    • James Allan: Center for Intelligent Information Retrieval, University of Massachusetts Amherst
  • URL
    [URL]

1. What did authors try to accomplish?

  • 直感的には検索を開始するコンテキストは異なるはずだが、隠れていて見えない
    • 例: 「ビットコイン」と検索 ← コンテキスト: ニュースで読んだとか、メールで見たとか
  • 事前検索コンテキストの分析を行うことで、1) 実世界での隠れたトリガー関係を検証し、2) 事前コンテキストとクエリの関係性を調べる

2. What were the key elements of the approach?

  • 教師なし学習で、隠されたトリガーとクエリの関係を予測できる混合生成モデルを作った
  • このモデルを使用して、クエリ予測とクエリ自動補完を改善した

3. What can you use yourself?

  • シチュエーションが特殊で微妙か

4. What other references do you want to follow?

  • なし

その他資料

github

  • なし

official blog

  • なし

author's resource

third person's explanation

  • なし

メモ

  • あるweb pageを見た後に、その内容に起因するクエリとそうでないクエリがある
    • ファイナンスのビットコインの記事を見た直後に
      • triggered: 「bitcoin」など
      • non-triggered: 「facebook」など無関係のクエリ
    • harold ramis(映画監督)が死んだというニュース記事を読んだ直後に、
      • triggered: 「harold ramis」(監督作についてなど記事のトピックそのものでは無いかもしれない)
      • non-triggered: craiglistなど無関係
      • leading: 「harold ramis dead」などトピックそのものについて(cos simが違う(table 4))
  • 結論
      1. 事前検索コンテキストは、検索意図の予測に役立つ
      1. 検索が特定のコンテキストからのみトリガーされるような非現実的なケースでは、そのコンテキストのみに商店を当てた教師ありモデルは、提案された教師なし混合モデルよりも効果的
      1. トリガーが隠されていると、提案モデルは準備にコストがかかるラベルなしでも正確に検索意図を予測する

感想

  • first authorのインターンの成果らしい。すごい

分野全体での立ち位置

  • クエリを入力する前の段階からクエリを入力するところに着目
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