Inspired by the learning mechanism of the human brain

jeudi, 07. janvier 2021 um 18:16 Uhr

Chanjong Im is a researcher at the Department for Information Science and Natural Language Processing at the University of Hildesheim. Currently he is a Ph.D. candidate and associate researcher taking part in the project »Distant Viewing«. After his bachelor‘s degree in International Business and East Asian culture in South Korea he completed his master‘s degree in Information Science and Telecommunication Engineering in South Korea and Germany at the program »Global Studies on Management and Information Science« (GLOMIS) in Hildesheim. His research interests are deep learning, computer vision, information retrieval and image retrieval.

"It is very important to work in international research projects. The data and information differs from country to country which means different problems are being seen. Understanding and analyzing the phenomenon and seeking new ideas to solve problems with a different perspective are extremely valuable."

Chanjong Im, PH.D. candidate

Mr. Im, what is deep learning?

Chanjong Im: I just talked to my colleague Sophie März, a doctoral candidate in our department, about this topic. Deep learning is a specific form of information science and a subfield of machine learning. Inspired by the learning mechanism of the human brain, diverse information is learned by the neural networks from a huge amount of data. This enables a complex analysis and building of patterns which are continously improved by new results and solutions.

What is your main research topic?

The title of my Ph.D. proposal is »Image search analysis tool with deep neural networks generated metadata from illustrations depicted on 19th century children and youth literature«. I am building an image information system that utilizes computer vision techniques on the illustrations depicted in the 19th century children and youth literature. The system is aimed to provide support and enrich the studies in the humanities. The scholars can view, explore, and analyze the historical illustrations using the system and get the information generated by artificial intelligence in addition to metadata created by librarians. Much complex information can be retrieved by utilizing various Computer Vision deep models such as printing techniques that were used for imprinting illustrations in the 19th century, finding similar illustrations to detect reuse cases, locating objects depicted in the old days. Ultimately, I will try to find out through several expert interviews whether the system does benefit and enhance their research.

Why is this research important?

The system provides information that is very time consuming, expensive, and requires domain expertise to obtain if they were to be processed and annotated by the experts. Of course, human experts will be able to get complex information in a very detailed and accurate way. However, it is often better to utilize the technologies that are recently advancing rapidly and let the experts worry about further/advanced topics with the information given by the machines.

How does Covid-19 affect your work as a researcher?

It is not convenient nor pleasant. I am not able to meet friends nor colleagues as frequently as before which makes me a bit depressed. It was really difficult for me to accept the fact that I won’t be able to see some close Indian researchers who canceled their trips to visit Hildesheim University. Also, I wasn’t able to attend conferences nor meetings as they got postponed due to the situation. However, many of the things are now happening online which is working out quite well. I could have been more steps further in the research if it weren’t the pandemic. Nevertheless, I am using the time to improve the system and writing the thesis.

Together with Professor Thomas Mandl you are cooperating with Pai Chai University, Daejeon, and Chungbuk National University, Cheongju, in South Korea. Why did you choose the University of Hildesheim to work here?

It is very important to work in international research projects. The data and information differs from country to country which means different problems are being seen. Understanding and analyzing the phenomenon and seeking new ideas to solve problems with a different perspective are extremely valuable. It is important to know the people when it comes to deciding the workplace. I was very lucky as I spent one year earlier here in Hildesheim during my master’s degree as part of a joint degree program »GLOMIS« and got to know some people who already were in the department. I have an excellent time during my stay here. Many people I met are very supportive and comforting. I especially enjoy working together with Professor Thomas Mandl and Professor Christa Womser-Hacker which was one of the biggest reasons that made me start working here as a researcher and pursue the Ph.D since 2017. They have been providing me a ton of support not only on research but also for small problems from life in Germany. The people I met here in Hildesheim all made me feel very comfortable and I thought it would be okay for me to give a shot on my new Ph.D. journey.

Hildesheim is not Berlin or Hamburg. Hildesheim is much smaller...

...Hildesheim is a perfect city for studying. We have here a rather quiet and relaxed atmosphere with great nature around. At the same time, it is not very small that I feel isolated or inconvenient.

Interview by Isa Lange.

Dieses Interview ist erschienen im aktuellen
Universitätsjournal DIE RELATION, Ausgabe 7, Wintersemester 2020/21.


Chanjong Im is a Ph.D candidate at the Department for Information Science and Natural Language Processing at the University of Hildesheim. Foto: Isa Lange/Uni Hildesheim