IEEE Region 10 Postgraduate Research Paper Contest 2nd Prize Winner

Son T. Huynh, Vietnam National University

Nowadays, there are a huge number of scientific works reviewed and published in different research fields, especially in computer science, biology, physics, and applied sciences. There are an increasing number of new journals being launched and new conferences happening every year, giving researchers multiple choices to submit their research work based on the submission’s quality and novelty. This creates a dilemma for researchers in selecting the most suitable journal or conference venue for paper submission. Well-known publishers (e.g., IEEE, Springer, and Elsevier) created a webpage on their portals to support researchers in quickly checking an appropriate journal based on the title, the abstract, and the list of keywords of their papers.

Recently, recommendation systems have become an indispensable part of many industries, including e-commerce (Amazon, eBay, Walmart, or Lazada), video/audio entertainment (YouTube, Spotify), and social networking sites (Facebook, Twitter, LinkedIn). There are a lot of studies related to recommendation systems in terms of both algorithms and applications. According to our literature survey, there are a few recommendation systems for article submission.

Figure 1. Three students with their supervisor (from the left to the right): Son Huynh, Phong Huynh, Dac Nguyen, and Dr. Binh Nguyen. All students are working at AISIA Research Lab and studying at John Von Neumann Institute and University of Science, Vietnam National University in Ho Chi Minh City, Vietnam.

Our paper aims to investigate an efficient algorithm for paper submission recommendation systems that can help scientists (especially young researchers) easily find an appropriate journal or conference for submitting their research work. Such a system can take the title, abstract, and keywords of the submission as the input data, and then provide top conferences/journals related to the inputs. To construct a suitable algorithm, we apply deep learning methods to improve the recommendations’ effectiveness of the system. By concatenating two different types of features using one-dimension convolutional neural networks (CNN) and pre-trained embedding techniques (Glove or FastText), the proposed model can achieve much higher performance in terms of both computational speed and recall compared to other state-of-the-art methods. We measured the performance of different approaches using the dataset proposed by Wang et al., 2018, including articles and proceedings paper in computer vision. These experimental results are promising, and we are motivated to continue enhancing the recommendation models’ performance for paper submission recommendation systems. In the future, we want to build a real application for this system that can serve different researchers in various research areas to find the most proper journals or conference venues efficiently.


  1. D. Wang, Y. Liang, D. Xu, X. Feng, and R. Guan, “A content-based recommender system for computer science publications,” Knowledge-Based Systems, vol. 157, pp. 1 – 9, 2018.

All the 2020 Region 10 Postgraduate Research Paper Contest winning papers will be invited for presentation at IEEE TENCON 2020 and the papers will be included in the conference proceedings.