Title: Age of AI: Artificial Intelligence, Agglomerative Intelligence, Adaptive Intelligence, and Ambient Intelligence

Speaker: Dr. Hsiao-Wuen HON
               Corporate Vice President, Microsoft
               Chairman, Microsoft Asia-Pacific R&D Group
               Managing Director, Microsoft Research Asia

Date:     19 November 2015 (Thursday)
Time:     3:00 p.m. – 5:00 p.m.
Venue:   Rm 513, William M. W. Mong Eng. Bldg

Abstract:
The building of a computer that can intelligently carry out tasks that require "human level intelligence" has been a goal for computer scientists since the 1950's. Since that time, tasks that were considered to exhibit human level intelligence such as understanding natural language, carrying out conversational speech, playing master level chess, and interpreting images have all seen a great deal of improvements and many have seen real world applications. In this talk, I will provide an overview of the progress of artificial intelligence, highlight some recent milestone results in image recognition and natural language understanding, and predict the progress of AI in research and application over the next 5-10 years.

---

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

Speaker: Chin-Yew LIN
               Principal Research Manager
               Microsoft Research Asia

Date:     19 November 2015 (Thursday)
Time:     3:00 p.m. – 5:00 p.m.
Venue:   Rm 513, William M. W. Mong Eng. Bldg

Abstract: 

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.

 

 

About Speaker:

Dr. Lin is a Principal Researcher and Research Manager of the Knowledge Computing group at Microsoft Research Asia. His research interests are knowledge mining, natural language processing, problem solving, question answering, and automatic summarization.

Recently, his main research directions are: (1) developing a knowledge computing framework for real world applications and services including automatic acquisition of semantic knowledge, machine reading for semantic indexing, and automatic understanding of user intents; and (2) developing big social data analytics platform and services – Project Soul. Building on experiences learned from Project Soul, his team is developing technologies to automatically learn social interaction knowledge from large-scale real world online data and transform unstructured and semi-structured web data into structured data to enable semantic computing. The goal is to enable context-aware interactive knowledge-enriched applications powered by intelligent data in the cloud.

 

He developed automatic evaluation technologies for summarization, QA, and MT. In particular, he created the ROUGE automatic summarization evaluation package. It has become the de facto standard in summarization evaluations. ROUGE has been chosen as the official automatic evaluation package for Document Understanding Conference since 2004.

Before joining Microsoft, he was a senior research scientist at the Information Sciences Institute at University of Southern California (USC/ISI). He was the program co-chair of ACL 2012 and program co-chair of AAAI 2011 AI & the Web Special Track. He is an Action Editor of Transactions of ACL and a member of the Editorial Board of Computational Linguistics.

 

Title: Video Understanding Meets Deep Learning

Speaker: Dr. Tao MEI
               Lead Researcher
               Microsoft Research Asia

Date:     19 November 2015 (Thursday)
Time:     3:00 p.m. – 5:00 p.m.
Venue:   Rm 513, William M. W. Mong Eng. Bldg

Abstract:
The recent advances in deep learning have boosted the research on video analysis. For example, convolutional neural networks have demonstrated the superiority on modeling high-level visual concepts, while recurrent neural networks have been proven to be good at modeling mid-level temporal dynamics in the video data. We present a few recent advances for understanding video content using deep learning techniques. Specifically, this talk will focus on: 1) translating video to sentence with joint embedding and translation, which achieves the best to-date performance in this nascent vision task, 2) first-person video highlight extraction with a pairwise deep ranking model, and 3) action recognition with a multi-granular spatiotemporal architecture which achieved rank 2 in CVPR THUMOS 2015 video classification challenge.