UVR Lab. was formed in Feb. 2001 at GIST to study and develop “Virtual Reality in Smart computing environments” that process multimodal input, perceive user’s intention and emotion, and respond to user’s request through Augmented Reality. Since 2012, UVR Lab moved to KAIST GSCT and restarted with a theme of “FUN in Ubiquitous VR.”
작성일 : 18-08-04
[ARRC] 제29회 KAIST 증강현실연구센터(ARRC) 콜로키움
 글쓴이 : UVR
조회 : 56  

KAIST KI-ITC 증강현실연구센터(ARRC)에서 제29회 증강현실연구센터 콜로키움을 아래와 같이 개최하였습니다.      


                                 -       -    




■ 주제특징 슬라이더 준지도 학습을 통한 유저-인터렉티브 데이터 정렬

         Criteria Sliders: Interactive Semi-supervised Learning of Continuous Dataset Criteria


■ 연사: Kwang In Kim, Senior Lecturer, Department of Computer Science, University of Bath.


■ 일시: 07월 05일 (오후 4시 (3시 30분부터 다과회)


■ 장소: KAIST KI빌딩 3층 교수회의실(D304)


■ 주관: KAIST KI-ITC 증강현실연구센터(ARRC)


■ 후원KAIST CT대학원한국HCI학회 DCH(디지털문화유산연구회대한전자공학회 AH(증강휴먼연구회


■ 요약 :

Large databases are often organized by hand-labeled metadata, or criteria, which are expensive to collect. We can use unsupervised learning to model database variation, but these models are often high dimensional, complex to parameterize, or require expert knowledge. We learn low-dimensional continuous criteria via interactive ranking, so that the novice user need only describe the relative ordering of examples. This is formed as semi-supervised label propagation in which we maximize the information gained from a limited number of examples. Further, we actively suggest data points to the user to rank in a more informative way than existing work. In this talk, we will discuss our new semi-supervised and active regression learning strategies designed to address challenges in instantiating such systems.



**This lecture will be held in Korean. I apologize for announcing this information in Korean.