Welcome

The DSAIG seminar series aim to bring together faculty members and students across different disciplines within the University to discuss research topics that are related to Data Science and its applications. It serves as a forum to foster interdisciplinary research and to uncover new possibilities for research collaboration in the Data Science field.

Monthly seminars will be organized with focuses on specific research topics related to Data Science and its applications, and given by renowned experts of the field with a broad range of academic backgrounds including Business, Engineering, Science, and Social Science.

For enquiries, please contact us at cb-dsaig@cityu.edu.hk.


Next Seminar

Dr. Antoni B. CHAN

Next Seminar

SIMPLIFYING MIXTURE MODELS WITH THE HIERARCHICAL EM ALGORITHM

By Dr. Antoni B. CHAN

Department of Computer Science
City University of Hong Kong

Date
5 June 2018 (Tuesday)
Time
12:30pm-2:00pm
Venue
Room 14-221, 14/F, Lau Ming Wai Academic Building, City University of Hong Kong
Language
English
Registration
Register Here

(A light sandwich lunch will be provided starting from 12:00noon. Please confirm your attendance.)

 

Abstract
We propose a hierarchical EM algorithm for simplifying a finite mixture model into a reduced mixture model with fewer mixture components. The reduced model is obtained by maximizing a variational lower bound of the expected log-likelihood of a set of virtual samples. We develop four applications for our mixture simplification algorithm: recursive Bayesian filtering using Gaussian mixture model posteriors, KDE mixture reduction, belief propagation without sampling, and clustering hidden Markov models. For recursive Bayesian filtering, we propose an efficient algorithm for approximating an arbitrary likelihood function as a sum of scaled Gaussian. Experiments on synthetic data, human location modeling, visual tracking, vehicle self-localization, and eye gaze analysis show that our algorithm can be widely used for probabilistic data analysis, and is more accurate than other mixture simplification methods.

Biography
Dr. Antoni B. Chan received the B.S. and M.Eng. degrees in electrical engineering from Cornell University, Ithaca, NY, USA, in 2000 and 2001, respectively, and the Ph.D. degree in electrical and computer engineering from University of California at San Diego (UCSD), La Jolla, CA, USA, in 2008. He was a Visiting Scientist with the Vision and Image Analysis Laboratory, Cornell University, from 2001 to 2003, and a Post-Doctoral Researcher with the Statistical Visual Computing Laboratory, UCSD, in 2009. In 2009 he joined the Department of Computer Science, City University of Hong Kong, Hong Kong, and is currently an Associate Professor. His research interests include computer vision, machine learning, pattern recognition, eye-gaze analysis, and music analysis. Dr. Chan received the National Science Foundation Integrative Graduate Education and Research Training Fellowship from 2006 to 2008, and an Early Career Award from the Research Grants Council of the Hong Kong Special Administrative Region, China, in 2012. He is currently a senior area editor for IEEE Signal Processing Letters, and was an area chair for ICCV 2015 and 2017.


Forthcoming

Dr. Antoni B. CHAN
SIMPLIFYING MIXTURE MODELS WITH THE HIERARCHICAL EM ALGORITHM
By Dr. Antoni B. CHAN (CS)
We propose a hierarchical EM algorithm for simplifying a finite mixture model into a reduced mixture model with fewer mixture components ... [ more ]
 
By Zhenzhen WANG
A Lightening Introduction to Parallel Computing: Using Apache Spark as an Example
By Zhenzhen WANG
Traditionally, many problems are so large and/or complex that it is impossible to solve them on a single computer, especially given limited computer ... [ more ]
 
Dr. Geoffrey TSO
Development of two consumer satisfaction and confidence indexes
By Dr. Geoffrey TSO (MS)
We shall cover the development of two consumer indexes with local impacts in Hong Kong. The Hong Kong Consumer Satisfaction Index (HKCSI) is a leading performance indicator for HK businesses ... [ more ]
 
Dr. LI Xi
One picture is worth 253 characters – Using photo mining to understand the role of a camera in online word of mouth
By Dr. LI Xi (MKT)
The last few years have witnessed an explosive growth in the number of photos attached to reviews in online communities. In this paper, using two datasets from TripAdvisor and Yelp, we empirically ... [ more ]
 
 

Previous

Dr. Lishuai LI
Characterize Operations in Air Transportation Systems Based on Large-scale Operational Data
By Dr. Lishuai LI (SEEM)
Many operations in air transportation systems, such as flying an airplane or air traffic management, are performed by human following specific procedures ... [ more ]
 
Dr. Hailiang CHEN
Vertical Integration, Movie Screening, and Theatrical Performance
By Dr. Hailiang CHEN (IS)
In Mainland China, vertical integration between producers/distributors and exhibitors is quite common in the motion picture industry ... [ more ]
 
Prof. Jun WANG
Neurodynamics-based Parallel Data Selection in the Era of Big Data
By Prof. Jun WANG (CS)
In the present information era, huge amount of data to be processed daily. In contrast of conventional sequential data processing techniques, parallel data ... [ more ]
 
Dr. Junhui WANG
A Smooth Collaborative Recommender System
By Dr. Junhui WANG (MA)
In recent years, there has been a growing demand to develop efficient recommender systems which track users' preferences and recommend potential items of interest to users. ... [ more ]
 
Dr. LI Xin
Collaboration Process Pattern & Teamwork Performance: A Graph Mining-Based Methodology
By Dr. LI Xin (IS)
It is well documented in management literature that characteristics of collaboration processes strongly influence team ... [ more ]
 
Prof. Frank CHEN & Dr. Eman LEUNG
A Data-driven Approach to the Integrated Elderly Care
By Prof. Frank CHEN & Dr. Eman LEUNG (MS)
The global trend today is to shift the care burden away from the acute care and long-term care, which are ... [ more ]
 
Dr. Huazhong ZHAO
Status Leaders, Hubs and their Social Influence
By Dr. Huazhong ZHAO (MKT)
The prevalence of internet technology and social media has provided marketers new opportunities to identify influencers and utilize their social influence to affect consumer decisions. We introduce status leaders ... [ more ]
 
Dr. Kevin SUN
Mining Triage Notes to Predict Emergency Department Admissions
By Dr. Kevin SUN (MS)
Emergency department (ED) overcrowding is a worldwide problem that undermines hospitals’ ability to provide timely care to patients who need it urgently. One of the ... [ more ]
 
Dr. Qingpeng ZHANG
Prediction of chronic disease occurrence using EHR data: A pilot study in Hong Kong
By Dr. Qingpeng ZHANG (SEEM)
Chronic diseases are major causes of morbidity and mortality worldwide. Many chronic diseases are preventable through effective disease management and prevention programs. In addition, some ... [ more ]
 
Prof. Jonathan ZHU
Mining Time Information in Digital Data
By Prof. Jonathan ZHU (COM)
Time is the single common denominator across all types of digital data from the internet, mobile networks, and IoT (internet of things). However, the rich information embedded in timestamps of digital data has often been taken for granted. As such, this gold ... [ more ]
 
Prof. Jeff HONG
Understanding Demand Uncertainty
By Prof. Jeff HONG (EF/MS)
Poisson processes are often used to model arrival (demand) processes in many business and engineering fields. In this talk, I will show you the results from a sequence of studies that we have done trying to understand the uncertainty behaviors in demand processes. The examples include phone calls ... [ more ]
 
Prof. Leon ZHAO
Blockchain as Trust Platform for Cross-boundary Dataflow
By Prof. Leon ZHAO (IS)
In the era of big data, more and more companies have become digitized. Further, under globalization initiatives such as the One-Belt One Road effort, business processes are being automated nowadays cross companies, leading to cross-boundary data flow. However, conventional information ... [ more ]
 
 

Privacy Policy · Copyright · Disclaimer · Accessibility · Contact Us
© 2018 City University of Hong Kong. All Rights Reserved.