|
Research Interest
|
1) Data
Mining
|
|
2) Statistical
Machine Learning
|
|
3) Social
Network
|
|
4) Multimedia,
Image Related Topic
|
|
|
|
Research Experience
|
|
|
Center-Piece Subgraph
|
Proposed
a fast solution to center-piece subgraph: given Q
queries nodes, we want to find the nodes and the resulting subgraph with some budget b that have strong connection
to all or most of query nodes.
|
|
Fast Random Walk with Restart
|
Proposed
a fast solution to do random walk with restart.
|
|
Learn from Graph
|
Mainly
focusing on 1) investigating transductive
inference for multi-modality by graph model; 2) a unified framework for transductive inference on graph model
|
|
Image Retrieval
|
Mainly
focusing on 1) automatic keyword annotation for large image database; and
2) unified learning in image retrieval. Proposed 1) a manifold-ranking
based keyword propagation method; 2) an optimization method based on graph
model to unify low-level feature, on-line relevance feedback and off-line
long term learning.
|
|
Blur Detection
|
Proposed
an effective and computing efficient algorithm to 1) judge whether or not a
given digital image is blurred and 2) determine to what extent it is
blurred. The proposed technology might be adopted and transferred to
product in Microsoft Research Asia.
|
|
Image Quality Assessment
|
Proposed
a unified learning-based framework for No-Reference image quality
assessment. The proposed framework contributes to No-Reference image
quality assessment mainly in three parts: 1) providing a general solution
for distortion measurement; 2) making a first attempt for holistic quality
evaluation; and 3) unifying the above two.
|
|
Anomaly Detection in Network
|
Mainly
focusing on applying fusion technology (such as boosting, multi-view
learning, etc.) into these fields. Proposed 1) Kernel principle component
classifier for anomaly detection 2) Introducting
boosting technology into anomaly detection.
|
|
Internet Traffic Prediction
|
Introduced
and designed some specific versions of boosting algorithm for self-similar
and non-linear Internet traffic prediction.
|
|
Simulated Annealing Algorithm
|
Proposed
two improved versions of simulated annealing algorithm for Travel Salesman
Problem (TSP)
|
|
|
|
|
Research Projects Co-operated with others
|
|
1. GMine: an iterative graph mining and visualization
tool: with José Fernando Rodrigues Júnior
|
|
2.
Improved SVMactive in Image Retrieval: with Jingrui He
|
|
3.
Pseudo Relevance Feedback in Web Image Retrieval: with Jingrui
He
|
|
4. Automatic Peak Number Detection in Image
Symmetry Analysis: with Jingrui He
|
|
|
|
|
|
|
|
|