Publication Type
Conference Proceeding Article
Version
publishedVersion
Publication Date
1-2021
Abstract
Deep neural networks have gained great success in a broad range of tasks due to its remarkable capability to learn semantically rich features from high-dimensional data. However, they often require large-scale labelled data to successfully learn such features, which significantly hinders their adaption in unsupervised learning tasks, such as anomaly detection and clustering, and limits their applications to critical domains where obtaining massive labelled data is prohibitively expensive. To enable unsupervised learning on those domains, in this work we propose to learn features without using any labelled data by training neural networks to predict data distances in a randomly projected space. Random mapping is a theoretically proven approach to obtain approximately preserved distances. To well predict these distances, the representation learner is optimised to learn genuine class structures that are implicitly embedded in the randomly projected space. Empirical results on 19 real-world datasets show that our learned representations substantially outperform a few state-of-theart methods for both anomaly detection and clustering tasks. Code is available at: https://git.io/ RDP
Discipline
Databases and Information Systems | OS and Networks
Research Areas
Data Science and Engineering
Publication
Proceedings of the 29th International Joint Conference on Artificial Intelligence IJCAI 2020, Virtual Conference, January 7-15
First Page
2950
Last Page
2956
Identifier
10.24963/ijcai.2020/408
Publisher
IJCAI
City or Country
Virtual Conference
Citation
WANG, Hu; PANG, Guansong; SHEN, Chunhua; and MA, Congbo.
Unsupervised representation learning by predicting random distances. (2021). Proceedings of the 29th International Joint Conference on Artificial Intelligence IJCAI 2020, Virtual Conference, January 7-15. 2950-2956.
Available at: https://ink.library.smu.edu.sg/sis_research/7136
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.