*박태준학술정보관 502호 4시 입니다.

 Clova AI: Future of NAVER and LINE

Biography

Jung-Woo Ha is the Clova AI Research Director and the leader of machine learning team of Clova, which is a collaboration project between NAVER and LINE, for an advanced AI-assistant platform. Ha got his BS and PhD degrees from Seoul National University. He is interested in deep learning-based computer vision, natural language processing, and signal modeling. which are core technologies of real-world AI applications that make global users happy. He has published his work on many top-tier conferences including NIPS, AAAI, KDD, CVPR, ICLR, and so on.

Abstract

Beyond a technology, now, artificial intelligence became an industry paradigm to enhance performances and create new values. NAVER, Korea’s No.1 internet company, has also developed diverse deep learning-based technologies such as image understanding, speech recognition, voice synthesis, recommendation, and natural language processing to its search and content services. Recent appearance of AI assistant platforms such as Amazon Alexa and Google assistant accelerates the movement toward AI-oriented industry. NAVER and LINE, a global messenger company with over 200 million MAUs, are developing a general purpose AI platform, called Cloud-based Virtual Assistant (Clova) as a collaboration project together. In 2017, Clova launched a mobile app in May, WAVE (AI-speaker) in Korea and Japan in August, and Friends (LINE character-based AI speaker) in October.

In this talk, I will present deep learning-based AI technologies and their applications to NAVER services. Also, I introduce the Clova platform and detailed technologies to make Clova services smarter, including StarGAN and Skim-RNN. Furthermore, we present a machine learning platform, NAVER Smart Machine Learning (NSML), to allow ML researchers to focus on their models and data, thus enhancing research efficiency.