- Business
- About Commercialisation
- Market Engagement
- Business Development
- Spin-outs
- Licensing
- Business Areas
-
NICTA Short Courses
- Free Seminars
- Upcoming courses
- Radar Week
- Course Locations
- Newsletter
- Courses by Category
-
Courses by Title
- Airborne Radar
- Advanced Electronic Warfare
- Advanced Topics in Digital Signal Processing
- Application Scalability
- Automotive Radar Systems
- Breakthroughs and New Concepts in Radar
- Business Aspects of Broadband Wireless Technologies and Mobile Commerce
- C++ for Engineers
- Complex Digital System Design Using SystemC
- Concepts of EW and IW
- DSP Based Carrier and Timing Recovery Techniques in Digital Modems
- DSP with FPGAs
- Fundamentals of EW: Overview
- Fundamentals of Kalman Filtering
- Fundamentals of RF System Design and Simulation
- GPS Receivers: Applications and Inside the Box
- Introduction to Sonar Signal Processing
- Linux for Embedded Developers
- Link 22
- Microwave and RF Circuit Design
- Modern Digital Receivers
- Multiuser Communications
- Satellite Communications
- Software Engineering Technology for Improving Safety of Medical Processes
- Sensor Networks
- Signal Analysis Techniques
- 6 Degrees of Freedom Modelling and Simulation of Aerospace Vehicles
- SOA_Performance
- Surveillance Radar Systems
- Systems Engineering Introduction
- Systems Engineering Management
- Systems Engineering
- Tactical and Strategic Missile Guidance
- Telecommunications and Teletraffic Theory
- Tracking and Data Fusion
- Using Handel-C with DK
- Validation Techniques for Finding Software Bugs
- VMF
- Breakfast Seminars
- Contact Short Courses
- Presenters
Machine Learning Using Graphical Models
| Presenter: | Dr Alex Smola and DR SVN Vishwanathan, NICTA |
| Dates: | please enquire |
| Venue: | please enquire |
About Introduction to Machine Learning with Kernels
The two-day short course “Introduction to Machine Learning with Kernels” gives an overview on practical methods for classification, pattern recognition, regression, and novelty detection. In other words, it introduces simple algorithms to solve artificial intelligence problems and to perform nonparametric estimation tasks. The advantage of the algorithms is that they are (a) simple to implement, (b) very general (c) easy to understand and (d) they lead to state of the art performance. In the course, I will be presenting applications to bioinformatics, pattern recognition, database cleaning, and document analysis.
The course is intended for students, academics and industrial engineers seeking to gain a concise overview over modern methods in machine learning. Prerequisites are basic knowledge of linear algebra (matrices, eigenvalues, matrix inverse) and statistics (normal distribution, conditional probability).
Course Topics
- Introduction: concepts of machine learning, how to use it to solve problems in pattern recognition, classification, data cleaning; Bayes rule, inference.
- Density estimation: Parzen windows, Silverman's rule, classification, Watson Nadaraya estimator.
- Perceptron and kernels: linear separation of sets, kernels, examples, preprocessing, nonlinear separation.
- Support vector machines: geometrical properties, how to solve classification problems, convex optimization.
- Applications to text categorization and bioinformatics: datasets, special kernels, efficient computation.
- Implementation: efficient optimization, existing toolboxes, online learning.
- Novelty detection: database cleaning, basic idea, convex optimization problem, applications.
- Getting good results in practice: nu trick, median trick for kernel adjustment, scaling data before optimization, interpreting the kernel matrix, smoothness and regularization.
Course Text
Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond (Adaptive Computation and Machine Learning) by B. Schlkopf and A. J. Smola.
For further information please contact:
Anne-Marie Eliseo
Industry Education Manager
phone: +61-8-8302-3928
email: industryeducation@nicta.com.au
