Unsupervised Neural Networks
Prof. Giansalvo Cirrincione, from the University of Picardie Jules Verne, Amiens, France and the University of South Pacific, Suva, Fiji Islands, will teach the course 01SHCRV Unsupervised neural networks (didattica di eccellenza) – 6 ECTS Credits.
The course will begin November 20th in SALA RIUNIONI CLARKE first floor of DET
This course is the natural continuation of the lessons on "neural networks and pattern recognition", in the sense that the latter deals with the supervised neural networks and the former with the unsupervised ones. However, it is self-contained, i.e. it does not require any knowledge about neural networks.
At first, the fields of application of these networks are considered: in particular, vector quantization, clustering and dimensionality reduction. Then, the learning is analyzed according to the way weight vectors are updated: hard competitive and soft competitive. For each case, the most important neural networks are presented.
The second part of the course describes the most principled neural techniques for vector quantization and dimensionality reduction. It also considers the case of non-stationary input data (evolving data streams), which has many applications, e.g. in diagnostics and text processing and recognition.
The last part deals with the main approaches to clustering, in particular in the case of hierarchies of clusters and bi-clusters, which are a very important tool in bioinformatics.
For information regarding the course content and the detailed schedule, see the attachment.
If interested, add it to personal career plan and send a registration e-mail to email@example.com