MDSA, 2023 1st seminar

The first seminar of the Data Science Management Association was held at Forest Hall on May 12, 2023 / Photo = Data Science Management Association

The Data Science Management Association successfully held the ‘Data Science Management Association 2023 1st Seminar’ on the 12th at Yeoksam Forest Hall under the theme of ‘Corporate Management Activities of AI Algorithms’.

The seminar was conducted in the following order: topic presentation, Q&A, and general discussion. Starting with the topic presentation by President Ho-yong Choi, topic presentations were made in that order by Academician Jeong-hoon Song, Hye-young Park, Bo-hyun Yoo, Min-cheol Kim, Jeong-woo Park, and finally Gyeong-hwan Lee, CEO of Pabii.

First, President Hoyong Choi gave a presentation on ‘Deep Learning as Solution Methods in Finance’ and introduced how machine learning and deep learning techniques can be used to find solutions to partial differential equations related to cash asset dividends of big tech companies. .

Academic member Song Jeong-hoon pointed out the problems with existing electricity/gas usage forecasts under the theme of ‘Monthly electricity/gas usage forecast for each building’ and further predicted monthly energy usage through statistical techniques that calculate the off-diagonal component of the second moment matrix. A model that predicts more accurately was introduced.

In order to find bubbles in the real estate sales market or auction market under the topic ‘Is bubble in housing auction market really bubble?’, academy member Park Hye-young defined the ‘difference between first and second place in the auction market’ as ‘bubble index’ and used it as a statistical index. The verification process through testing was explained.

Academician Bo-Hyun Yoo introduced a paper on the topic of ‘Discount/surcharge and momentum in the real estate auction market’ in which the factors that make up the winning bid rate in the real estate auction market were extracted using Fourier transform and the results were statistically verified.

Under the theme of ‘Interpretable Topic Analysis,’ academic member Mincheol Kim discussed a true ‘big data’ service that can be of practical help in matching between overseas buyers and domestic companies.

Under the theme of ‘Advertising time series modeling under measurement error,’ Jeongwoo Park, a member of the Academy of Sciences, introduced an advertising performance prediction model that statistically verifies and corrects the impact of measurement error included in user data of digital advertising.

Lastly, Professor Keith Lee discussed the interpretation and application cases of the recently controversial mathematical model related to ChatGPT, as well as expected usage methods, under the topic of ‘Use and Limitations of ChatGPT’.

In the general discussion that followed, SIAI (Swiss Institute Artificial Intelligence) students and MDSA academic members had a heated discussion about the direction of innovation and development in the Korean data science industry.