[행사/세미나] 인공지능대학원 전문가 초청 세미나 개최 안내( Dr.Heuna Kim, Senior machine learning researcher, 7/9 (금) 16시)
- 인공지능학과
- 조회수25705
- 2021-06-29
■ Title : Challenges of integrating classical numerical simulations and DL-based image processing (with examples in meteorology and eye tracking)
■ Speake : Dr.Heuna Kim, Senior machine learning researcher at Pupil Labs, Berlin Germany
■ Time : 2021 July 9th 16:00 ~ 18:00
■ Location : Webex room
▶ 미팅 번호 : 184 804 6823
▶ 비밀번호 : W9xGh4tpU2V
▶ 링크 :
https://skku-ict.webex.com/skku-ict/j.php?MTID=mbec120f1bc3e8225f00a9c037e81be0d
■ 1:1 meeting with Dr. Kim:
After the talk, we provide one-on-one meetings with Dr. Kim (10 mins each). Please write your name and email address (only school email) if you want to meet her personally until July 5th.
https://forms.gle/yQDEAdsGR9vbjiNF7
■ Abstract :
Image processing has gone through a critical paradigm shift since AlexNet came out in 2012 and the whole neural network approach was revived from the AI winter. Some popular image processing techniques 10 years ago have become hard to find now. Nevertheless, many fundamental concepts in the traditional methods have survived and are inspiring the new technological advancement. With no doubt, some of the fundamental methodologies that have seemed to disappear will be again popular at some point in future, as happened to neural networks.
This raises a question: can we know in advance what will continue to become the fundamental components in image processing? Can we even define what are the fundamentals in some scientific research?
In this talk, I would like to bring reflections on our perspectives of fundamental and traditional methods by explaining contemporary challenges of applying deep-learning-generation's computer vision on meteorology and eye tracking. Meteorology and eye tracking both have been historically the pinnacles of numerical simulations involving interdisciplinary research collaborations due to their complicated nature and essential use cases in our life. Furthermore, they are still going through this above-mentioned paradigm shift; in other words, deep learning has not yet beaten - maybe is even far away from beating - the traditional methodologies.
■ Bio :
Heuna Kim is a senior machine learning researcher at Pupil Labs with the largest market share of head-mounted mobile eye trackers.
She specializes in satellite image and video analysis; she has previously worked on solar and wind power forecasting at Haezoom, and temporal localization and classification of video data at Twenty Billion Neurons.
Heuna received a doctoral degree in Theoretical Computer Science from Freie Universität Berlin in Germany, a masters degree in Computer Science from KAIST and a bachelors degree in Mathematics from POSTECH.
She is also currently active in helping refugees get IT education and Internet at their homes and shelters and in supporting a renewable energy discussion group in Berlin.