On August 28, 2017

Speaker

Justine Cassell, Carnegie Mellon University


Title

Modeling Dyadic Phenomena in Intelligent Conversational Agents


Abstract

In this talk I propose a particular computational sociocultural approach to the study of the so-called "social emotions" - intrinsically dyadic states such as rapport, friendship, intimacy, interpersonal closeness. I rely on this approach to describe the surface level observable verbal and nonverbal behaviors that function to evoke, deepen, demonstrate, and destroy these dyadic social emotions. I highlight the need for differentiating the observable behaviors from inferable underlying states by demonstrating how putatively negative visible behaviors may play a positive role in underlying states. Finally, I describe some important roles that these often discounted aspects of human behavior play in learning, commercial transactions, and other facets of day-to-day life. Each step of this talk is illustrated by experiments that involve human-human and human-computer interaction. I include novel approaches to modeling and generating behaviors for human-computer interaction on the basis of the human-human corpora. And finally, lessons are drawn both for the study of human behavior, and the improved design of technologies capable of engaging in conversational interaction with people over the long-term.


Bio

Justine Cassell is Associate Dean of Technology Strategy and Impact in the School of Computer Science at Carnegie Mellon University, and Director Emerita of the Human Computer Interaction Institute. She co-directs the Yahoo-CMU InMind partnership on the future of personal assistants. Previously Cassell was faculty at Northwestern University where she founded the Technology and Social Behavior Doctoral Program and Research Center. Before that she was a tenured professor at the MIT Media Lab. Cassell received the MIT Edgerton Prize, and Anita Borg Institute Women of Vision award, in 2011 was named to the World Economic Forum Global Agenda Council on AI and Robotics, in 2012 named a AAAS fellow, in 2016 made a Fellow of the Royal Academy of Scotland, and in 2017 made a Fellow of the ACM. In 2017-2018 Cassell will hold the Chaire Blaise Pascal in Paris. Cassell has spoken at the World Economic Forum in Davos for the past 5 years on topics concerning the impact of AI and Robotics on society.

On September 04, 2017

Speaker

김광현(Kwang-Hyun Kim), NAVER Corp.


Title

Naver Search X AI


Abstract

Naver has been the most popular search engine for over a decade in Korea. As a search engine, Naver aims to match a user’s search intent with the information in the web pages and databases. To achieve this, Naver has utilized traditional information retrieval models and machine learned models to understand users, queries and web documents. The recent breakthrough of deep learning made big progress in image classification and text analysis. With this powerful method, Naver could provide new search services that meets user needs.

In this talk, I will cover research topics and search services especially recent AI based projects that Naver Search is working on.


Bio

Kwang Hyun Kim is a Director of Naver Search at Naver Corp. Joined Naver at 2000, He ran Naver’s core search department where he and his team were responsible for the Naver search modeling and R&D. Kwang Hyun received Ph.D in Information Retrieval at SoongSil University, His research interests are in the area of Search Modeling, NLP, Machine Learning and AI. In 2013, He was named a ‘Technical Fellow’ which is given to the Naver’s top technical specialists. Kwang Hyun is currently running Naver Search’s many ongoing projects, including AI based projects like Clova, Smart-lens and Naver-i.

On September 11, 2017

Speaker

강미선(Miseon Kang), Ewha Womans University


Title

당신의 집은 안녕하십니까? (Say Hello to Your House!)


Abstract

청년주택, 노인주택, 하우스푸어, 주거빈곤자 등 우리 삶의 가장 기본재인 주택의 문제를 거시적으로, 그리고 미시적으로 들여다 보고 대안을 함께 모색해 본다.


Bio

Miseon Kang is a professor in the department of architecture, Ewha womans university. Her research is mainly focused on housing, gender issues, planning processes and shared economy.

On September 18, 2017

Speaker

Jean-Charles Bazin, KAIST CT


Title

Computational Media: from Processing to Creation of Media Contents


Abstract

Computational Media refers to the analysis, processing and creation of media contents with computer algorithms, such as pictures, videos, VR/AR experiences, social networks data, and art materials. It encompasses several disciplines such as computer vision, computer graphics, machine learning, VR/AR, robotics and multimedia. Exploring their complementarities provides exciting research opportunities and opens a wide range of novel applications. In this talk, I will present some of our Computational Media projects: gaze correction in Skype calls, manipulation of facial performance videos, scene-space video processing, VR content stabilization, and video editing via motion data.


Bio

Prof. Jean-Charles Bazin is Assistant Professor at KAIST and Director of the Computational Media Lab. He holds a joint appointment at the Graduate School of Culture Technology (CT) and the School of Electrical Engineering (EE). He has been working on the analysis, processing and creation of media contents. A representative list of his research topics includes video editing, VR/AR, machine learning and audiovisual media processing. His work has been covered in the news by Discovery Channel, Gizmodo, Engadget, The Verge, TechCrunch, among many others. He published several papers in the premier venues of computer vision (TPAMI, CVPR, ICCV, and ECCV), computer graphics (TOG, SIGGRAPH, SIGGRAPH Asia and Eurographics), robotics (IJRR) and multimedia (MobileHCI and ICME).

Before joining KAIST, he was an Associate Research Scientist at Disney Research Zurich, and at the same time an Adjunct Lecturer at ETH Zurich, Switzerland (2014-2016). Before working at Disney Research, he was a Postdoc and a Senior Research Associate at ETH Zurich (2011-2014), and a Postdoctoral Fellow at the University of Tokyo, Japan (2010-2011). He received an MS in Computer Science at Universite de Technologie de Compiegne, France (2006) and a PhD in Electrical Engineering, at KAIST, South Korea (2011).

On September 25, 2017

Speaker

심은수(Eunsoo Shim), Samsung Electronics


Title

Challenges for Evolution of Connected Car


Abstract

Experts predict that between 2025 and 2035,people will be able to sleep in a car. In the morning, the car will recognize your face and automatically open the door. Simply tell the car your destination, then the car will brief your schedule, read your favorite book, or you can just take a nap, all while the car drives herself.

To realize this future connected car, we have a lot of technological challenges. An autonomous car needs to locate itself and detect vehicles and pedestrians in a split second with limited sensor hardware such as camera and radar in a diverse conditions, even under low illuminations and rainy and snowy weather.

To recognize the user, a convenient and reliable user authentication method such as face recognition is required which withstands spoofing attempts. An intelligent voice interface for safe driving could only be possible with extreme noise robustness and flexible understanding of natural language.

Successful implementation of these functions will have huge computational needs leading a challenge of designing innovative processor architecture. In this talk, we would like to overview the some of the key technical challenges, recent developments across academia and industry, and describe some of our efforts to address these challenges.


Bio

The speaker has been with Samsung Electronics since 2011 as a vice president and currently is the head of Software Solution Lab, Samsung Advanced Institute of Technology. He has been leading research & development of a wide range of artificial intelligence technologies including speech recognition,autonomous driving, and neural processor. His lab developed noise-resilient top-notch speech recognition technologies that are deployed in Samsung products since 2015. His lab also developed the robust face recognition algorithm deployed in Samsung’s latest products. He earned Ph.D. and MS in EE, respectively in 2004 and 1998 from Columbia University in NYC and MS and BS in Physics in 1991 and 1989 respectively from Seoul National University.

On October 23, 2017

Speaker

곽재용(Jae-young Kwak), Film Director


Title

영화 엽기적인 그녀가 만들어지던 과정


Abstract

영화 한 편이 준비되고 만들어지는 과정을 통해 영화의 산업적, 예술적인 측면을 이해하도록 한다.


Bio

Kwak Jae-yong (born 22 May 1959) is a South Korean film director and screenwriter. He studied physics at Kyung Hee University. He achieved success with his debut film Watercolor Painting in a Rainy Day in 1989, but the failure of his next two movies led to eight years of unemployment before a comeback with the smash-hit film My Sassy Girl in 2001. He is known for his limitless fondness of love stories set in a mix of different genres.

On October 30, 2017

Speaker

박기호(Kiho Park), Photographer


Title

Just Do It ... Go Ahead Take a Chance ... You Will Love It!


Abstract

When was the last time you did something that you always wanted to do but couldn’t because you were too afraid to do it? As a child, I took chances all the time: I rode my brother’s bicycles and crashed into a incoming car, I jumped over between two houses, but ended up falling below and broke my leg, I played with matches and ended up almost burning up my house. Although most of those foolish actions angered my parents and left indelible memories, I still cherish those crazy events and want to pass it on to my kids.

In the world of fine arts, musicians, architects, painters, sculptures create various mater pieces to enlighten our existence. I strongly believe, those creative thinkers are often risk takers, unafraid of any unknown consequence. That willingness to challenge is embedded in our early childhood. Unfortunately, as we grow older, instead of harnessing those qualities, we throw way that creative silliness. Now it’s time to be in the mind of child again to enlighten our mind.


Bio

Photographer, Kiho Park was born in Seoul, Korea in 1960 and moved to the United States in 1973.  While majoring in photography at Rhode Island School of Design (RISD), he met Bruce Davidson, a renowned documentary photographer, and worked with him on several projects including National Geographic.  Under Davidson’s guidance, he learned that a photographer must learn to create stories in pictures with poetic honesty, just as a good writer would with his pencil.  For his documentary graduate project about South East Asian refugees, instead of exhibiting his artwork in a gallery setting, Kiho chose to re-create the living conditions of refugees by displaying his photographs at a Cambodian family’s house, sparking dialog about their experiences.

After graduation in 1986, he worked as a photo director at a small design firm in Cleveland, Ohio.  But after seeing the political turmoil taking place in his homeland, he moved to Seoul as a freelance photographer for Newsweek.  Upon arrival in Seoul, Kiho was greeted by an endless stream of rocks and Molotov Cocktails being thrown at him, as students and despairing citizens confronted the riot police in downtown Seoul.  For the next twenty years Kiho remained in Asia as an Asian specialist for Time, Forbes, Fortune, and Business Week.

On November 06, 2017

Speaker

오상록(Sang Rok Oh), Korea Institute of Science and Technology


Title

4차 산업혁명과 데이터기반 스마트 농업


Abstract

초연결, 초지능으로 대표되는 4차 산업혁명 시대가 도래하고 있다. 이에 따라 4차 산업혁명 시대의 대표적인 기술들인 사물인터넷(IoT), 클라우드 컴퓨팅, 빅데이터, 인공지능 기술들이 융합을 통하여 기존의 시장이나 산업의 영역을 혁신하고 있거나 새로운 시장, 산업들이 나타나고 있다. 농업도 예외는 아니다. 과거 경험 위주의 농업에서 데이터를 이용한 스마트 농업으로 빠르게 변화하고 있으며 이러한 변화는 농업이 기존의 먹거리를 제공하는 단순한 1차 산업에서 제조업과 서비스산업을 포함하는 6차 산업으로 변하고 있다. 본 강연에서는 이러한 변화의 배경에 대하여 분석하고 4차 산업혁명 시대의 신성장동력으로서의 데이터기반 스마트 농업에 대하여 토론하기로 한다.


Bio

오상록 박사는 현재 한국과학기술연구원 강릉분원장을 맡고 있으며, IT융합 및 로보틱스 관련 연구를 하고 있다.

On November 13, 2017

Speaker

송창현(Chang Song), NAVER Corp.


Title

Ambient Intelligence R&D at Naver Labs


Abstract


Bio

Chang Song received the B.S. degree in computer science, from Iowa State University, and the M.S. degree in computer science from Purdue University in 1996. He started his career at DEC (Digital Equipment Corp.) in UNIX filesystem and later worked on VFS and Cluster filesystem. He continued his career at HP, Microsoft, and Apple. He then moved to NAVER and later became the CTO of Naver in 2015. He founded NAVER LABS (NAVER spin-off) to realize his next dream. His research interests include cloud service infrastructure, system performance analysis and improvements, next generation interface, maps, machine learning, autonomous machine, product design, AR/3D, and more.

On November 20, 2017

Speaker

Daisuke Inoue, National Institute of Information and Communications Technology


Title

Overview of NICTER Project - Practical Cybersecurity R&D Project in Japan


Abstract

To fight against highly sophisticated cyberattacks, continuous research on Cybersecurity technologies and their seamless deployments are key issues. NICT, a sole national research institute in ICT field in Japan, is researching cutting-edge Cybersecurity technologies such as a large-scale darknet (unused IP addresses) monitoring system NICTER, a darknet-based alert system DAEDALUS, and an anti-APT platform NIRVANA-KAI. In this talk, we present these technologies with real-time visualizations, and also show some case studies of their practical deployments in Japan.


Bio

Daisuke Inoue received his B.E. and M.E. degrees in electrical and computer engineering and Ph.D. degree in engineering from Yokohama National University in 1998, 2000 and 2003, respectively. He joined Communications Research Laboratory (CRL), Japan, in 2003. CRL was relaunched as National Institute of Information and Communications Technology (NICT) in 2004, where he is currently the director of Cybersecurity Laboratory. He received several awards including the best paper award at the 2002 Symposium on Cryptography and Information Security (SCIS 2002), the commendation for science and technology by the minister of MEXT, Japan, in 2009, the Good Design Award 2013, the Asia-Pacific Information Security Leadership Achievements 2014, and the award for contribution to Industry-Academia-Government Collaboration by the minister of MIC, Japan, in 2016.

On November 27, 2017

Speaker

손수엘(Sooel Son), KAIST


Title

Web Attacks on Common Vulnerabilities


Abstract

We live in the era of mobile and Web technology. The more Web applications process private/sensitive information, the riskier Web vulnerabilities become. In this talk, I will address common developer mistakes that cause Web vulnerabilities as well as their detection methods. In the second half of the talk, I will present a new mobile attack vector that leverages popular mobile advertising libraries, which was presented NDSS 2016. I will demonstrate how malicious ads can infer sensitive information about users by accessing external storage, which is essential for media-rich ads in order to cache video and images. I will conclude the talk with recommendations for Android app developers to mitigate the threat.


Bio

Sooel Son is an assistant professor at KAIST. His research interests cover Web security and privacy as well as software analysis. He worked in Google security/privacy research and advertising identity/privacy teams. He received a Ph.D. (2014) from The University of Texas at Austin.

On December 04, 2017

Speaker

이미란(Miran Lee), Microsoft Research Asia


Title

Sharing the Latest Research of Microsoft Research


Abstract


Bio

Miran Lee is a Director and principal research program manager of Microsoft Research Outreach Group at Microsoft Research responsible for academic collaboration in Korea and Asia-Pacific region.

Lee joined Microsoft Research Asia in 2005 as university relations manager to build long-term and mutually beneficial relations with academia. She is based in Korea, where she engages with leading research universities, research institutes, and relevant government agencies. She establishes strategies and directions, identifies business opportunities, designs various programs and projects, and manages budget. She works with students, researchers, faculty members, and university administrators to build strong partnerships, and works closely with the research groups at Microsoft Research, focusing on research collaboration, curriculum development, talent fostering, and academic exchanges. She has successfully run a number of global and regional programs such as Gaming & Graphics, Web-Scale NLP, Machine Translation, eHealth, SORA (Software Radio), Kinect, and Microsoft Azure for Research.

Prior to her current role, Miran Lee co-founded Smart Systems, which specializes in IT outsourcing services in Illinois, United States. As CEO of Smart Systems, she successfully led the business with more than 100 percentage annual growth. From 1993 to 2002, she worked at British Telecom Korea in various positions ranging from systems engineer to account director to vice president. Lee also worked at Samsung SDS, where she was responsible for International VAN (Value Added Network) businesses and led the International VAN business team. She started her business career as a system developer at General Electric Information Services, where she developed email, EDI, and in-house applications.

Miran Lee was an adjunct professor in the Telecommunication Department at Anyang University for two years (2001–2002) and she earned her MS in Engineering from Ewha Womans University.