1. NCCU Academic Hub
  2. 學術產出
  3. College of Informatics
  4. Proceedings
  5. Item

學術產出-Proceedings

Article View/Open

Publication Export

Google ScholarTM

政大圖書館

Citation Infomation

題名 ICE: Item Concept Embedding via Textual Information
作者 蔡銘峰
Wang, Chuan-Ju
Wang, Ting-Hsiang
Yang, Hsiu-Wei
Chang, Bo-Sin
Tsai, Ming-Feng
貢獻者 資訊科學系
關鍵詞 concept embedding; conceptual retrieval; information network; textual information
日期 2017-08
上傳時間 29-Aug-2017 13:24:20 (UTC+8)
摘要 This paper proposes an item concept embedding (ICE) framework to model item concepts via textual information. Specifically, in the proposed framework there are two stages: graph construction and embedding learning. In the first stage, we propose a generalized network construction method to build a network involving heterogeneous nodes and a mixture of both homogeneous and heterogeneous relations. The second stage leverages the concept of neighborhood proximity to learn the embeddings of both items and words. With the proposed carefully designed ICE networks, the resulting embedding facilitates both homogeneous and heterogeneous retrieval, including item-to-item and word-to-item retrieval. Moreover, as a distributed embedding approach, the proposed ICE approach not only generates related retrieval results but also delivers more diverse results than traditional keyword-matching-based approaches. As our experiments on two real-world datasets show, ICE encodes useful textual information and thus outperforms traditional methods in various item classification and retrieval tasks.
關聯 Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval, Pages 85-94
資料類型 conference
DOI http://dx.doi.org/10.1145/3077136.3080807
dc.contributor 資訊科學系zh_TW
dc.creator (作者) 蔡銘峰zh_TW
dc.creator (作者) Wang, Chuan-Juen_US
dc.creator (作者) Wang, Ting-Hsiangen_US
dc.creator (作者) Yang, Hsiu-Weien_US
dc.creator (作者) Chang, Bo-Sinen_US
dc.creator (作者) Tsai, Ming-Fengen_US
dc.date (日期) 2017-08en_US
dc.date.accessioned 29-Aug-2017 13:24:20 (UTC+8)-
dc.date.available 29-Aug-2017 13:24:20 (UTC+8)-
dc.date.issued (上傳時間) 29-Aug-2017 13:24:20 (UTC+8)-
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/112296-
dc.description.abstract (摘要) This paper proposes an item concept embedding (ICE) framework to model item concepts via textual information. Specifically, in the proposed framework there are two stages: graph construction and embedding learning. In the first stage, we propose a generalized network construction method to build a network involving heterogeneous nodes and a mixture of both homogeneous and heterogeneous relations. The second stage leverages the concept of neighborhood proximity to learn the embeddings of both items and words. With the proposed carefully designed ICE networks, the resulting embedding facilitates both homogeneous and heterogeneous retrieval, including item-to-item and word-to-item retrieval. Moreover, as a distributed embedding approach, the proposed ICE approach not only generates related retrieval results but also delivers more diverse results than traditional keyword-matching-based approaches. As our experiments on two real-world datasets show, ICE encodes useful textual information and thus outperforms traditional methods in various item classification and retrieval tasks.en_US
dc.format.extent 1574416 bytes-
dc.format.mimetype application/pdf-
dc.relation (關聯) Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval, Pages 85-94en_US
dc.subject (關鍵詞) concept embedding; conceptual retrieval; information network; textual informationen_US
dc.title (題名) ICE: Item Concept Embedding via Textual Informationen_US
dc.type (資料類型) conference
dc.identifier.doi (DOI) 10.1145/3077136.3080807
dc.doi.uri (DOI) http://dx.doi.org/10.1145/3077136.3080807