智能信息处理国际会议IIP

第十届智能信息处理国际会议IIP2018

发布时间:2022-08-01

10th IFIP International Conference On

Intelligent Information Processing

19-22, October, 2018 Nanning, China

Sponsored by

    International Federation for Information Processing, IFIP TC12

Co-Sponsored by

Chinese Association for Artificial Intelligence

Guangxi University  

Institute of Computing Technology, Chinese Academy of Sciences

Conference Organization

Steering Committee

Z. Shi (China) (Chair)              A. Aamodt (Norway)            D. Leake (USA)        

S. Vadera (UK)............................. ..   .B. An (Singapore)               X. Yao (UK)

General Chair

U. Furbach (Germany)             P. Yu (USA)                   X. Yao (UK)

Program Chairs

Z. Shi (China)                     E. Mercier-Laurent (France)         J. Li (Australia)

Organization Committee

Organization Chairs:  Cheng Zhong, Guangxi University, China

General Secretary:    Zuqiang Meng, Guangxi University, China

Keynotes Speakers

Advances in Transfer Learning

Qiang Yang 

Chair Professor at Computer Science and Engineering Department

Hong Kong University of Science and Technology

China.

Abstract.  Transfer learning aims to leverage knowledge from existing tasks to solve new tasks. In this talk, I will give an overview of recent advances of transfer learning and point to future works that both have practical significance and theoretical potential.

Grounding & Learning about Human Environments & Activities for Autonomous Robots

Anthony G Cohn

Director of Research and Innovation,

School of Computing,

University of Leeds, Leeds, LS2 9JT, UK.

Distinguished Visiting Professor at Tongji University

Abstract:  With the recent proliferation of human-oriented robotic applications in domestic and industrial scenarios, it is vital for robots to continually learn about their environments and about the humans they share their environments with. In this paper, we present a novel, online, incremental framework for \emph{unsupervised} symbol grounding in real-world, human environments for autonomous robots. We demonstrate the flexibility of the framework by learning about colours, people names, usable objects and simple human activities, integrating state-of-the-art object segmentation, pose estimation, activity analysis along with a number of sensory input encodings into a continual learning framework. Natural language is grounded to the learned concepts, enabling the robot to communicate in a human-understandable way. We show, using a challenging real-world dataset of human activities as perceived by a mobile robot, that our framework is able to extract useful concepts, ground natural language descriptions to them, and, as a proof-of-concept, generate simple sentences from templates to describe people and the activities they are engaged in.

Invited Speakers

Artificial Intelligence overview and impacts

Eunika Mercier-Laurent 

CReSTIC, University of Reims   Champagne Ardennes

FR-34160 SAINT DREZERY, France

E-mail: eunika@innovation3d.fr

Abstract:  The recent craze for AI and limitation to data, deep learning and chatbots cover only a very small part of AI patrimony. Facing various and difficult challenges requires knowing the whole spectrum in aim to select the best approach and techniques. Environmental impact and climate change can be easily faced by right AI and alternative thinking. Smart software (and hardware) conceived using eco-design approach have a potential to reduce our impact and bring a contribution to the Planet protection.

        Deep Learning based Image Interpretation

Lichen Jiao

School of Artificial Intelligence

Xidian University, Xi'an, China

Abstract:  With the development of sensor and data storage technology, the data acquisition becomes easier, but it brings “big data” problems, of which, Images are the most common information sources in daily life. Compared with other information sources, the images contain huge amounts of information, and its complexity, redundancy and other characteristics distinguish it from other information sources. The image processing is relatively difficult, and the human visual system has shown excellent capabilities in image processing, which attracting the attention of many researchers. The application of deep learning model in recent years has made a new progress in the study of deep neural networks and brought a new research boom.

Is knowledge Engineering out-of-date?

Yueting Zhuang

Dean of College of Computer Science

Zhejiang University, China

Abstract:  The world is now in the era of a new wave AI technology. Though, many of us still remembered the days when knowledge Engineering along with expert system was extremely hot, in such a state that is similar to deep learning or machine learning nowadays. This talk will first give a short survey of AI, especially the concept of knowledge Engineering, rule-based expert system, and so on, and then introduce the data-driven machine learning approaches used in systems like Wikipedia, Freebase, Google Knowledge Graph etc. It will conclude that knowledge engineering is NOT out-of-date. What indeed outdated is the method of knowledge acquisition. Finally it will introduce knowledge computing engine in order to support knowledge engineering. 

Effective Utilization of Genomic Data

Yadong Wang

School of Computer Science and Technology

Harbin Institute of Technology, China 

Abstract:  With the rapid development and wide application of high-throughput genome sequencing technology, a series of large scale international genomics study plans have been carried out. This makes an explosive and continuous growth of genomics data, and the in depth integration of genomics data and healthcare data, which triggers a “data revolution” in life science.

Nowadays, the effective use of genomics data has become an engine critical to the development of life science as well as other related fields such as healthcare, medicine, drug development, etc. Genomics data has large volume, various data structures and complex relationships, which makes it difficult to effectively analyze and utilize. State of the art genomics data analysis technologies can merely dig out 30-50% of the value of the data, i.e., the large potentials of the data cannot be fully realized. This has been one of the biggest challenges to genomics and bioinformatics.

The drawbacks of the existing analysis approaches, including (but not limited to) low sensitivity, low accuracy, low consistency, low efficiency, etc., are the bottlenecks to the effective use of genomics data. It is the main way to solve these problems by developing advanced bioinformatics algorithms, to continuously improve the quality and efficiency of data analysis. Centers for Bioinformatics of Harbin Institute of technology have made great efforts in recent years to develop a batch of innovative genomics data analysis algorithms and systems. These algorithms and systems substantially improve their performances for a series of fundamental genomics data analysis, such as sequencing read alignment, variant calling and genomics big data visualization. With these achievements, several technical bottlenecks have been breakthrough, which make large contributions to the effective use of genomics



附件下载: