학술행사
- 발표자 최재웅 교수(성균관대학교)
- 개최일시 2025-03-20 10:30-12:30
- 장소 국가수리과학연구소 산업수학혁신센터(판교)
유튜브 스트리밍 예정입니다.
- 일시: 2025년 3월 20일(목), 10:30-12:30
- 장소: 판교 테크노밸리 산업수학혁신센터 세미나실
- 경기 성남시 수정구 대왕판교로 815, 기업지원허브 231호 국가수리과학연구소 / 무료주차는 2시간 지원됩니다.
- 발표자: 최재웅 교수(성균관대학교)
- 주제: Advancing Neural Optimal Transport: Improving Stability and Overcoming Fake SolutionsNeural Optimal Transport (OT) has emerged as a powerful tool for generative modeling and image-to-image translation, but existing methods suffer from instability and fake solutions. In this talk, I will introduce two recent advancements that address these issues. First, I will present Displacement Interpolation Optimal Transport Model (DIOTM), which stabilizes training by leveraging displacement interpolation. Second, I will discuss Optimal Transport Plan learning (OTP), a novel approach that overcomes fake solutions by introducing smoothing on the source measure. These methods improve the reliability of neural OT, expanding its effectiveness in machine learning applications.
유튜브 스트리밍 예정입니다.
- 일시: 2025년 3월 20일(목), 10:30-12:30
- 장소: 판교 테크노밸리 산업수학혁신센터 세미나실
- 경기 성남시 수정구 대왕판교로 815, 기업지원허브 231호 국가수리과학연구소 / 무료주차는 2시간 지원됩니다.
- 발표자: 최재웅 교수(성균관대학교)
- 주제: Advancing Neural Optimal Transport: Improving Stability and Overcoming Fake SolutionsNeural Optimal Transport (OT) has emerged as a powerful tool for generative modeling and image-to-image translation, but existing methods suffer from instability and fake solutions. In this talk, I will introduce two recent advancements that address these issues. First, I will present Displacement Interpolation Optimal Transport Model (DIOTM), which stabilizes training by leveraging displacement interpolation. Second, I will discuss Optimal Transport Plan learning (OTP), a novel approach that overcomes fake solutions by introducing smoothing on the source measure. These methods improve the reliability of neural OT, expanding its effectiveness in machine learning applications.
