학술행사
- 발표자 최재영(한국외국어대학교)
- 개최일시 2025-02-04 14:00-16:00
- 장소 국가수리과학연구소 산업수학혁신센터(판교)
유튜브 스트리밍 예정입니다.
-
일시: 2025년 2월 4일(화), 14:00-16:00
-
장소: 판교 테크노밸리 산업수학혁신센터 세미나실
-
경기 성남시 수정구 대왕판교로 815, 기업지원허브 231호 국가수리과학연구소 / 무료주차는 2시간 지원됩니다.
-
발표자: 최재영(한국외국어대학교)
-
주제: Ensemble Deep Network Learning and Its Applications
Ensemble learning is one of the most popular techniques in the area of machine learning (ML) and pattern recognition. In traditional ensemble learning, multiple ML models are constructed and outcomes each coming from individual ML models are optimally combined for the purpose of improving recognition, classification, and prediction performances. In this talk, I explain “ensemble deep network (EDN) learning” that apply ensemble learning techniques to recently developed deep network models, considering the following two important issues: (a) how to construct an optimal set of deep network ensemble with minimum correlation and (b) how to optimally combine deep network ensemble with the goal of obtaining maximum performance improvement. Also, this talk presents how the proposed EDN learning has been used to improve performance in practical applications such as image recognition, fine dust predictions, and time-series data predictions.
유튜브 스트리밍 예정입니다.
-
일시: 2025년 2월 4일(화), 14:00-16:00
-
장소: 판교 테크노밸리 산업수학혁신센터 세미나실
-
경기 성남시 수정구 대왕판교로 815, 기업지원허브 231호 국가수리과학연구소 / 무료주차는 2시간 지원됩니다.
-
발표자: 최재영(한국외국어대학교)
-
주제: Ensemble Deep Network Learning and Its Applications
Ensemble learning is one of the most popular techniques in the area of machine learning (ML) and pattern recognition. In traditional ensemble learning, multiple ML models are constructed and outcomes each coming from individual ML models are optimally combined for the purpose of improving recognition, classification, and prediction performances. In this talk, I explain “ensemble deep network (EDN) learning” that apply ensemble learning techniques to recently developed deep network models, considering the following two important issues: (a) how to construct an optimal set of deep network ensemble with minimum correlation and (b) how to optimally combine deep network ensemble with the goal of obtaining maximum performance improvement. Also, this talk presents how the proposed EDN learning has been used to improve performance in practical applications such as image recognition, fine dust predictions, and time-series data predictions.
