Graduate to present at top data-mining conference

19 Aug 2015

A Xi’an Jiaotong-Liverpool University computer science graduate has had his final year project work accepted for presentation at a top international conference, with reviewers praising the quality of his work.

Chunchuan Lv, who graduated from BSc Information and Computing Science in July with the highest final average grade of his year group, co-authored a paper that has been accepted for regular presentation at the Institute of Electrical and Electronics Engineers (IEEE) International Conference on Data Mining (ICDM).

His paper is one of only 8.4 percent of papers that have been accepted for presentation at the conference.

ICDM attracts researchers and application developers from a wide range of data mining-related areas such as statistics, machine learning, pattern recognition, databases and data warehousing, data visualisation, knowledge-based systems, and high-performance computing. It is the world’s premiere research conference on data mining.

Data mining is the process of analysing data from different perspectives and summarising it into useful information. Regular papers present work that is mature and provides significant contributions to the field.

Conference reviewers praised the innovative approach and technical quality of Chunchuan’s work, saying it focussed on an "important topic" for the data mining and machine learning communities.

The project proposes a theoretical framework to allow accurate and robust classification of adversarial examples, an important area of research in image classification. Adversarial examples are images that have been changed slightly. These changes are almost imperceptible to the human eye, but are significant enough to cause computers to see the original image and changed one very differently. This creates issues when it comes to computers classifying and grouping these images.

Dr Kaizhu Huang, one of Chunchuan’s two supervisors for his final year project, said: “Chunchuan's research is both novel and significant. He has proposed a unified model that not only incorporates previous famous approaches as special cases, but more importantly provides a new insight on learning theory, especially deep learning theory. Such research output will be of big impact on the field.”

Dr Hai-Ning Liang, second supervisor, said: “Chunchuan’s final year project work is one of the best examples of the kind of high-quality research work that XJTLU students can achieve if they are given the opportunities, platform, and guidance to explore their own ideas and interests. I am very proud of having played a small part in helping to unlock the potential inherent in students like Chunchuan.”

Professor Yong Yue, Head of the Department of Computer Science and Software Engineering, said: “Our department endeavours to offer our students a research-led teaching environment and our efforts are yielding fruit.

“We are dedicated to enhancing student learning by providing them with research and entrepreneurship experiences. We are delighted to pass the knowledge and skills onto our students so that they can start to tread the path to become leaders in their fields.”

Chunchuan will continue his studies at the University of Edinburg in Scotland. It is the sixth-oldest university in the English-speaking world and has some of the finest researchers in the area that Chunchuan is interested in pursuing.

19 Aug 2015