Personalized Image Retrieval with Sparse Graph Representation Learning

Xiaowei Jia; Handong Zhao; Zhe Lin; Ajinkya Kale; Vipin Kumar

doi:10.1145/3394486.3403324

Personalized Image Retrieval with Sparse Graph Representation Learning

Xiaowei Jia, Handong Zhao, Zhe Lin, Ajinkya Kale, Vipin Kumar

Computer Science and Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

11 Scopus citations

Abstract

Personalization is essential for enhancing the customer experience in retrieval tasks. In this paper, we develop a novel method CA-GCN for personalized image retrieval in the Adobe Stock image system. The proposed method CA-GCN leverages user behavior data in a Graph Convolutional Neural Network (GCN) model to learn user and image embeddings simultaneously. Standard GCN performs poorly on sparse user-image interaction graphs due to the limited knowledge gain from less representative neighbors. To address this challenge, we propose to augment the sparse user-image interaction data by considering the similarities among images. Specifically, we detect clusters of similar images and introduce a set of hidden super-nodes in the graph to represent clusters. We show that such an augmented graph structure can significantly improve the retrieval performance on real-world data collected from Adobe Stock service. In particular, when testing the proposed method on real users' stock image retrieval sessions, we get promoted average click position from 70 to 51.

Original language	English (US)
Title of host publication	KDD 2020 - Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
Publisher	Association for Computing Machinery
Pages	2735-2743
Number of pages	9
ISBN (Electronic)	9781450379984
DOIs	https://doi.org/10.1145/3394486.3403324
State	Published - Aug 23 2020
Event	26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2020 - Virtual, Online, United States Duration: Aug 23 2020 → Aug 27 2020

Publication series

Name	Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

Conference

Conference	26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2020
Country/Territory	United States
City	Virtual, Online
Period	8/23/20 → 8/27/20

Bibliographical note

Publisher Copyright:
© 2020 ACM.

Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.

Keywords

graph convolutional network
image retrieval
sparse graph

Publisher link

10.1145/3394486.3403324

Find It

Find It at U of M Libraries

Cite this

Jia, X., Zhao, H., Lin, Z., Kale, A., & Kumar, V. (2020). Personalized Image Retrieval with Sparse Graph Representation Learning. In KDD 2020 - Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 2735-2743). (Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining). Association for Computing Machinery. https://doi.org/10.1145/3394486.3403324

Personalized Image Retrieval with Sparse Graph Representation Learning. / Jia, Xiaowei; Zhao, Handong; Lin, Zhe et al.

KDD 2020 - Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Association for Computing Machinery, 2020. p. 2735-2743 (Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Jia, X, Zhao, H, Lin, Z, Kale, A & Kumar, V 2020, Personalized Image Retrieval with Sparse Graph Representation Learning. in KDD 2020 - Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Association for Computing Machinery, pp. 2735-2743, 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2020, Virtual, Online, United States, 8/23/20. https://doi.org/10.1145/3394486.3403324

Jia X, Zhao H, Lin Z, Kale A, Kumar V. Personalized Image Retrieval with Sparse Graph Representation Learning. In KDD 2020 - Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Association for Computing Machinery. 2020. p. 2735-2743. (Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining). doi: 10.1145/3394486.3403324

Jia, Xiaowei ; Zhao, Handong ; Lin, Zhe et al. / Personalized Image Retrieval with Sparse Graph Representation Learning. KDD 2020 - Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Association for Computing Machinery, 2020. pp. 2735-2743 (Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining).

@inproceedings{50facd08df5140078b8bbfe9face7025,

title = "Personalized Image Retrieval with Sparse Graph Representation Learning",

abstract = "Personalization is essential for enhancing the customer experience in retrieval tasks. In this paper, we develop a novel method CA-GCN for personalized image retrieval in the Adobe Stock image system. The proposed method CA-GCN leverages user behavior data in a Graph Convolutional Neural Network (GCN) model to learn user and image embeddings simultaneously. Standard GCN performs poorly on sparse user-image interaction graphs due to the limited knowledge gain from less representative neighbors. To address this challenge, we propose to augment the sparse user-image interaction data by considering the similarities among images. Specifically, we detect clusters of similar images and introduce a set of hidden super-nodes in the graph to represent clusters. We show that such an augmented graph structure can significantly improve the retrieval performance on real-world data collected from Adobe Stock service. In particular, when testing the proposed method on real users' stock image retrieval sessions, we get promoted average click position from 70 to 51.",

keywords = "graph convolutional network, image retrieval, sparse graph",

author = "Xiaowei Jia and Handong Zhao and Zhe Lin and Ajinkya Kale and Vipin Kumar",

note = "Publisher Copyright: {\textcopyright} 2020 ACM. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.; 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2020 ; Conference date: 23-08-2020 Through 27-08-2020",

year = "2020",

month = aug,

day = "23",

doi = "10.1145/3394486.3403324",

language = "English (US)",

series = "Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining",

publisher = "Association for Computing Machinery",

pages = "2735--2743",

booktitle = "KDD 2020 - Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining",

}

TY - GEN

T1 - Personalized Image Retrieval with Sparse Graph Representation Learning

AU - Jia, Xiaowei

AU - Zhao, Handong

AU - Lin, Zhe

AU - Kale, Ajinkya

AU - Kumar, Vipin

PY - 2020/8/23

Y1 - 2020/8/23

N2 - Personalization is essential for enhancing the customer experience in retrieval tasks. In this paper, we develop a novel method CA-GCN for personalized image retrieval in the Adobe Stock image system. The proposed method CA-GCN leverages user behavior data in a Graph Convolutional Neural Network (GCN) model to learn user and image embeddings simultaneously. Standard GCN performs poorly on sparse user-image interaction graphs due to the limited knowledge gain from less representative neighbors. To address this challenge, we propose to augment the sparse user-image interaction data by considering the similarities among images. Specifically, we detect clusters of similar images and introduce a set of hidden super-nodes in the graph to represent clusters. We show that such an augmented graph structure can significantly improve the retrieval performance on real-world data collected from Adobe Stock service. In particular, when testing the proposed method on real users' stock image retrieval sessions, we get promoted average click position from 70 to 51.

AB - Personalization is essential for enhancing the customer experience in retrieval tasks. In this paper, we develop a novel method CA-GCN for personalized image retrieval in the Adobe Stock image system. The proposed method CA-GCN leverages user behavior data in a Graph Convolutional Neural Network (GCN) model to learn user and image embeddings simultaneously. Standard GCN performs poorly on sparse user-image interaction graphs due to the limited knowledge gain from less representative neighbors. To address this challenge, we propose to augment the sparse user-image interaction data by considering the similarities among images. Specifically, we detect clusters of similar images and introduce a set of hidden super-nodes in the graph to represent clusters. We show that such an augmented graph structure can significantly improve the retrieval performance on real-world data collected from Adobe Stock service. In particular, when testing the proposed method on real users' stock image retrieval sessions, we get promoted average click position from 70 to 51.

KW - graph convolutional network

KW - image retrieval

KW - sparse graph

UR - http://www.scopus.com/inward/record.url?scp=85090418831&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85090418831&partnerID=8YFLogxK

U2 - 10.1145/3394486.3403324

DO - 10.1145/3394486.3403324

M3 - Conference contribution

AN - SCOPUS:85090418831

T3 - Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

SP - 2735

EP - 2743

BT - KDD 2020 - Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

PB - Association for Computing Machinery

T2 - 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2020

Y2 - 23 August 2020 through 27 August 2020

ER -

Personalized Image Retrieval with Sparse Graph Representation Learning

Abstract

Publication series

Conference

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this