Transcript Machine learning

Introduction to
Machine Learning
Vũ Việt Vũ
Computer Engineering Division, Electronics Faculty
Thai Nguyen University of Technology
August, 2014
Outline
•
•
•
•
What is Machine Learning?
Application of Machine Learning
Research group in Machine Learning
Conclusion
What is the Machine Learning?
• Machine learning is a subfield of computer
science and artificial intelligence that deals
with the construction and study of systems
that can learn from data.
What is the Machine Learning?
AI = Artificial Intelligence
Problems of Machine learning
•
•
•
•
Supervised learning problem
Unsupervised learning problem
Semi-supervised learning problem
Active learning prblem
Supervised learning
Given a training data set (labeled data) T = {x1, x2,…,xn}; xi: vector
with m dimensions and xi belongs to a class Cv, v = 1…m.
Task: Build a classifier model to predict the label of a new data xnew.
T
xnew
Example
Face recognition
?
xnew
T: Traning data
Example (cont.)
CAR NUMBER
RECOGNITION
Text Picture
Recognition Result
What is the T here?
What is the xnew here?
Methods
-
Support Vector machine
Neural Network
Decision Trees,
K-nearest neighbor graph,…
Unsupervised learning
Unsupervised learning: The objective of unsupervised
learning is to discover structures in the data.
Example: clustering, outlier detection,...
How many clusters here?
Example
Methods for clustering
•
•
•
•
•
K-means clustering
Density-based clustering
Fuzzy-Cmeans clustering
Hierarchical clustering,
Graph-based clustering
Semi-supervised learning
Semi-supervised learning combines both labeled
and unlabeled examples to build the classifier
model or to discover structures in data.
Methods:
- Self trainning,
- Support Vector Machine,
- Graph-based methods
Active learning
Active learning is a special case of semi-supervised learning in
which a learning algorithm is able to interactively query the user
to obtain the outputs at new data points.
Machine
learning
algorithm
Input Data
Active
Learning
Output
Labeled data
Questions
Response
Users
(Experts)
[ Vu et al, ECAI2010
Vu et al, ICPR2010
Vu et al, Pattern
Recogntion’12]
14
Application of Machine learning
Computer vision
Object Recognition
Robotics (ASIMO, ...)
Natural language processing
Search engines (Google, Yahoo)
Medical diagnosis
Bioinformatics
Stock market analysis
Classifying DNA sequences
Speed and handwriting
recognition
Game playing
Software engineering
Adaptive website
Computational finance
Recommender systems
Research Group
of Machine Learning
•
•
•
•
•
•
•
Dr. Vu Viet Vu, Thai Nguyen University of Technology
Prof. Nicolas Labroche, France
Prof. Violaine Antoine, France
Prof. Le Ba Dung: Institute of Information Technology, Viet Nam
Dr. Vu Hai: Ha Noi University of Technology
Dr. Nguyen Thi Oanh: Ha Noi University of Technology
PhD student. Nguyen Manh Tuan: Institute of Information
Technology, Viet Nam
Theory: Unsupervised learning, clustering, active learning,...
Application: Image processing, object recognition
Publication
• Vu Viet Vu, Nicolas Labroche, and Violaine Antoine. Semi-supervised
graphe-based clustering. Submitted to Pattern Recognition Journal (ISI),
2014.
• Violaine Antoine, Nicolas Labroche, Vu Viet Vu. Evidential seed-based semisupervised clustering. Submitted to the 7th International Conference on Soft
Computing and Intelligent Systems and 15th International Symposium on
Advanced Intelligent Systems, Japan, 2014
• Vu Viet Vu, Nicolas Labroche, Violaine Antoine, and Le Ba Dung, Active
seeds selection with a k-nearest neighbors graph, In proceeding of the first
NAFOSTED Conference on Information and Computer Science (NICS'14),
Ha Noi, Viet Nam, pp: 386-395. Selected to publish in Advances in Intelligent
Systems and Computing, Springer. 2014.
Publication (cont.)
•
•
•
•
•
•
Vu Viet Vu, Nicolas Labroche, and Bernadette Bouchon-Meunier, Viet-Thang Vu, and
Nguyen Thi Thu Hien. Graph based Semi-supervised Clustering. Journal of Science and
Technology, Ha Noi University of Education, Viet Nam, March- 2013 .
Vu Viet Vu, Nicolas Labroche, and Bernadette Bouchon-Meunier. Improving Constrained
Clustering with Active Query Selection. Pattern Recognition 45(4): 1749-1758 [SCI],
ISSN: 0031-3203, 2012
Vu Viet Vu, Nicolas Labroche, and Bernadette Bouchon-Meunier. An Efficient Active
Constraint Selection Algorithm for Clustering. In Proc. of the 20th IEEE International
Conference on Pattern Recognition (ICPR-2010), Istanbul, Turkey, August, 2010
Vu Viet Vu, Nicolas Labroche, and Bernadette Bouchon-Meunier. Boosting Clustering by
Active Constraint Selection. In Proc. of the 19th European Conference on Artificial Intelligence
(ECAI-2010), Lisbon, Portugal, August, 2010
Vu Viet Vu, Nicolas Labroche, and Bernadette Bouchon-Meunier. Active Learning for SemiSupervised K-Means Clustering. In Proc. of the 22nd IEEE International Conference on Tools
with Artificial Intelligence (ICTAI-2010), Arras, France.10.2010
Vu Viet Vu, Nicolas Labroche, and Bernadette Bouchon-Meunier. Leader Ant Clustering with
Constraints. In Proc. of the 7th IEEE International Conference on Computing and
Communication Technologies (IEEE-RIVF-2009), Danang, Vietnam, July, 2009
Conclusion
• Developing new methods for machine learning
• Using machine learning methods for real
applications: image processing, pattern
recognition, speed processing,...
• The courses at TNUT: Artificial Intelligence,
Image Processing, Speed Processing, Algorithm
theory,...
Thank you for your attention!