Program

Core-sets in BIG-DATA - My Personal View

I will survey the notion of core-sets and its use in Big-Data. I survey some recent results on constructing core-sets, and point to some connections to related other problems and further research.

Short Bio:

Ilan Newman is a Professor in the Computer Science Department at the University of Haifa. His research interest is Theoretical Computer Science, in particular: combinatorial property testing and finite metric spaces. He is also interested in concrete complexity models: circuit complexity, decision trees, communication complexity. His publications can be found on DBLP and Google Scholar.

An Introduction of Burrows-Wheeler Transform and Its Variants

Burrows-Wheeler transform (BWT) was proposed in 1994 as a means to perform lossless compression on text strings. Soon after that, Ferragina and Manzini (FOCS'00; JACM'05) discovered the string matching power of BWT, and this result, together with independent work by Grossi and Vitter (STOC'00; SICOMP'07) and by Sadakane (ISAAC'00; JALG'03), subsequently started the field of compressed text indexing. Nowadays, BWT gains much popularity in bioinformatics areas, for a couple of famous applications such as BWA, Bowtie, and SOAP2, use BWT as the core in their design.

In this talk, we will give an introduction of BWT and some of its variants, and show how they work, and how they can be used to derive interesting results, both in theory and in practice. No background knowledge is required.

Short Bio:

Wing-Kai Hon is a Professor in the Computer Science Department at National Tsing Hua University. He received his doctorate from the University of Hong Kong. His research interests are: String Matching, Indexing, Compression, External-Memory, Combinatorial Optimization. His teaching areas are: Discrete Math, Algorithms, Data Structures, Theory of Computing, Randomized Algorithm. His publications can be found on DBLP and Google Scholar.

Fast sampling with quantum annealing devices

Fast sampling of Boltzmann distributions is a bottleneck in the application of many machine learning methods.

Quantum annealers are potential candidates for accelerating this process. I will discuss variations on the standard annealing procedure that allow more effective sampling of quantum Boltzmann distributions, and demonstrate an application in a model with strong long range correlations (Nature 560 pp 456--460).

I will contrast the observed dynamics with quantum Monte Carlo, a classical method to sample the same distributions.

Short Bio:

Dr. Jack Raymond is an optimization algorithms researcher at D-Wave Systems Inc., in Burnaby, Canada. He completed his PhD in applied mathematics at Aston University, UK. His publications can be found on DBLP and Google Scholar.

Applications of Network Science for Business and Marketing

In recent years, network science has been increasingly applied into web data analysis, such as video, image and text. Most of network analysis methods can widely pursue network distance, community detection and node importance tasks. Due to the nodes in these kinds of networks mostly contain the common attributes and share the same neighbors, we identify them as homogeneous networks, also including real-world networks such as social network, citation network and collaborative network. Therefore, we present an explanatory study on the applications of network science into the fields of business and marketing, and demonstrate two case studies. One is the supply chain management (SCM), and the other one is customer relationship management (CRM).

In the first case study, we redefine the supply chain as supply network, and analyze interactional relationships and topological characteristics between suppliers and customers in supply network rather than in traditionally linear supply chains. The experimental results deliver the efficient and effective support to supply network design and management.

In the second case study, we describe a novel analytic approach applying network science into customer relationship network (CRN) instead of CRM, which is constructed from the RFID data. The CRN can help us to understand the customer relationships from the network perspective, and the results provide us a structural and relational understanding of customer groups by a topological process.

Bio:

Yi Zuo is a professor of Maritime Big Data and Intelligent Ships Technology Institute in Dalian Maritime University. He received his B.S. degree in College of Information and Engineering of Northeast University in China in 2004, and received his M.S. degree and Ph.D. degree in School of Information Science of Nagoya University in 2009 and 2012, respectively. He is IEEE Technical Committee Co-Chair of the SMC and has been Program Committee member for several international conferences such as MSNDS, MISNC and APWC on CSE. He also served as invited reviewer for several international journals such as Computers & Operations Research, IEEE Trans. on SMC: Systems and Technological Forecasting & Social Change. He majors in Machine Learning and Data Science also including Evolutionary Algorithm. His primary researches are the application of Bayesian network for prediction and classification to resolve practical issues in real world. Recently, he briefly focuses on network science and social computing, and pursues an exploratory work on the applications of such novel techniques into the fields of business and marketing.

Storing Partitions in Sublinear Space

We introduce a data structure for representing a partition of an integer n, which uses O(sqrt(n)) bits of space. This is constant multiple of the information theoretic lower bound. Three types of operations access, bound, prefixsum are supported in constant time by using the notion of conjugate of a partition.

Data Science for Analysis of Path Data in Marketing

This is a report on the analysis of path data conducted by Kansai University Data Science Laboratory. The laboratory promotes research to explain consumer behavior through the analysis of path data in which a person’s movement is recorded in a spatial configuration. The first half of the report focuses on the complexity of consumer behavior and investigates the relationship between shopping path complexity and purchase behavior using shopping path data obtained from a supermarket. The method of analyzing consumer behavior using fractal dimension is introduced. The latter half of the report outlines an eye-tracking experiment, also conducted at the supermarket, and reports on the construction of a model that shows the relationship between gaze and product choice. Finally, the report describes future prospects for sensor device marketing that uses sensor information.