<MSRA-Yonsei Joint Workship 개최 안내>

개최일시 : 2015년 5월 6일 수요일 14:00 ~ 18:00  

개최장소 :  제 2공학관 B201

세미나 제목 : Solving Math Number Word Problem – What can NLP, Knowledge Engineering and Machine Learning contribute?

발표초록 : With the availability of personal agents such as Cortana, Siri and Google Now, it seems a world of humans and machines communicate and solve problems together in natural language is not far away. The scene of freely chatting with HAL in 2001: A Space Odyssey and Samantha in Her could happen to us seems within reach. The question is are we ready to go there? Do we have sufficient and necessary technologies to make it happen? In this talk, I will use solving math number word problem as an example to show how the emerging NLP, knowledge engineering and machine learning technologies can pay the way to this holy grail and what challenges that we have to address to travel down the path.

강연자 성함&직함 / 소속 : Dr. Chin-Yew Lin / Knowledge Mining Group Microsoft Research Asia, Beijing

개최일시 : 2015년 5월 6일 수요일 14:00 ~ 18:00  

개최장소 :  제 2공학관 B201

세미나 제목 : Massive MIMO Research in MSRA

발표초록 : In this talk, I will first give a brief overview of wireless and networking group and then focus on the massive MIMO research in MSRA. Massive MIMO is an active research area and holds the promise to give another order of magnitude capacity improvement compared to existing wireless technologies. I will introduce the latest progress in this direction and also discuss the remaining challenges.

강연자 성함&직함 / 소속 : Dr. Kun Tan / Wireless and Networking Group Microsoft Research Asia 

개최일시 : 2015년 5월 6일 수요일 14:00 ~ 18:00  

개최장소 :  제 2공학관 B201

세미나 제목 : Intrinsic Image Decomposition and Its Application to Facial Image Enhancement

발표초록 : The decomposition of an image into its intrinsic components, namely an albedo and a shading layer, is a fundamental task in computer vision which can facilitate a variety of applications such as segmentation and shape-from-shading. In this talk, we will present an introduction to this problem and describe contributions we have made toward this goal on different forms of input, including single images, image sequences, and RGB-D video. We will also present our work on intrinsic image decomposition for human faces, and show how such decompositions can be used to enhance the appearance of facial images through the efficient and realistic application of virtual cosmetics.

강연자 성함&직함 / 소속 : Dr. Steve Lin / Internet Graphics Group Microsoft Research Asia 

개최일시 : 2015년 5월 6일 수요일 14:00 ~ 18:00  

개최장소 :  제 2공학관 B201

세미나 제목 : When Multimedia Information Retrieval Meets Big User Data

발표초록 : Multimedia information retrieval has been studied for decades. Previous research has focused on content-based understanding, indexing and ranking from large scale multimedia data. With the popularity of commercial search engines as well as social networks, users are contributing huge amount of click data on the Web. Along with this trend, multimedia information retrieval is witnessing a new paradigm shift—from content-based retrieval to leveraging user data for more effective yet efficient retrieval. In this talk, we will show several research projects which leverage big user data to conduct multimedia retrieval. In particular, we will demonstrate that how big user data can be used for 1) video understanding and embedding, 2) predicting popularity and trending search, and 3) reranking.

강연자 성함&직함 / 소속 : Dr. Tao Mei / Media Computing Group Microsoft Research Asia

개최일시 : 2015년 5월 6일 수요일 14:00 ~ 18:00  

개최장소 :  제 2공학관 B201

세미나 제목 : Systems Research on Big-Data Platforms at Microsoft

발표초록 : Big-Data applications impose significant challenges in computing infrastructures. In this talk, I will introduce some of our recent projects on the systems support for Big-Data computation at Microsoft Research, in collaboration with product teams. Our journey begins with understanding real production systems at Microsoft: focusing on efficient batch processing (Dryad) and effective sharing of cluster resources (Apollo). Based on the experience gained, we identified and enabled new capabilities such as matrix computation (MadLINQ) and real-time streaming (TimeStream, SCOPE Streaming). I will focus on the abstractions for building related Big-Data systems, as well as the design choices we made according to the real, production requirements and workloads at Microsoft. I will also briefly talk about other related work in the Systems Research Group at Microsoft Research Asia.

강연자 성함&직함 / 소속 : Dr. Zhengping Qian / System Research Group Microsoft Research Asia