CALL FOR PAPERS
Special Section: Recommender Systems with Big Data
Journal of Computer Science and Technology (JCST)
Aims and Scope
Recommender systems have become an important means to improve user experience, engagement, and revenue and conversion rates, which are critical to the current and future success of most businesses and companies. Although recommender systems research has made significant advances over the past decades, traditional recommendation techniques are not powerful enough to address new challenges arising from the “4Vs” (volume, variety, velocity, veracity) of the big data era. Firstly, large volumes of user behaviour data and media data are generated at unprecedented and ever-increasing scales. Existing recommendation techniques are designed for the conventional scale datasets, struggling to meet the requirements of scalability and storage. Secondly, media or product content data involve a great variety of data formats in different modalities: texts, images, videos and arbitrary combinations of them. Thirdly, user behaviour and product data are generated in real time and continually arrive in the form of streams. Data streams are temporally ordered, continuous and come in high velocity. Fourthly, large amounts of biases, noises and abnormality exist in user generated behaviour and content data (e.g., user comments). What dimensions of data quality are particularly important for recommenders and what methods can address them? The data veracity is another big challenge for recommender systems. This special section aims to address these new challenges to enable both accurate and scalable recommendations, in this era of big data.
This special section of JCST journal papers will focus on technologies and solutions related, but not limited to:
* large-scale parallelization and distributed processing techniques to speed up the offline training of complex recommender models;
* scalable hash and indexing techniques to speed up the online recommendation and reduce the storage cost;
* algorithms and models, especially deep learning techniques that exploit heterogeneous and multi-modal content information to make better recommendations, and that address data sparsity and cold start issues;
* context-aware recommendation systems incorporating various contextual information such as location-based recommendation and social recommendation;
* incremental recommendation solutions and online learning models to deal with continuous updates, especially real-time streaming data for recommendations;
* active learning techniques to acquire high-quality and informative user feedback data;
* data cleaning techniques to improve the quality of user generated behaviour and content data;
* recommender system survey that focuses on at least one of the new challenges.
Besides original research papers, we also strongly encourage high quality survey papers, systems papers and applications papers.
Xiaofang Zhou, The University of Queensland, Australia
Hongzhi Yin, The University of Queensland, Australia
December 27, 2017 January 15, 2018
First review completed: February 20, 2018
Revision due: March 20, 2017
Final decision: April 20, 2017
Final manuscript due: April 27, 2018
Expected publication: July 5, 2018
All submissions must be done electronically through JCST's e-submission system at https://mc03.manuscriptcentral.com/jcst, with a manuscript type: "Special Section on Recommender Systems with Big Data".