Seminars

There will be two main types of seminars within the course:

tutorials and exercises on the topics covered by lectures, algorithm implementation in Java,
consultations on assignments - individual analysis of the solved problem and discussions on the progress made and further steps towards completing the task.

There are also a midterm test and oral presentations of the assignment results planned on certain weeks.

Students use a development platform of their choice to implement the programs for their assignment (typically C/C++, Java, Matlab or Mathematica).

An expected average home preparation time is 5 hours per week.

Seminar	Date	Topic	Materials
1.	25.2.	Survey of successful applications of neural networks	ann_examples-2016.pdf
2.	3.3.	Data mining	JavaNNS, JavaNNS (Win), data.zip, JavaNNS (Mac), snnsv4.2.manual.pdf, Quick Guide to javaNNS RapidMiner RapidMiner download Glass data Three-way data splits
3.	10.3.	Successful applications of evolutionary algorithms, EA individual project assignment	applications_of_eas_2015.pdf, individual_projects_ea_2016.pdf, individual projects ANNs
4.	17.3.	Evolutionary models (steady-state, generational, etc.) (Java)	CheatSheet, sga_source_codes.zip, solution
5.	24.3.	Consultations on assignment, implementing MLP evaluation and back-propagation learning (Java)	Cheatsheet, Cheatsheet2, Sources, solution
6.	31.3.	Evolutionary algorithms: implementing and testing crossover/mutation operators (Java)	Sources, Cheatsheet, Solution, TSP references
7.	7.4.	Consultations on assignment
8.	14.4.	Recurrent ANNs: synchronous/asynchronous evaluation (Java)	Cheatsheet, Sources
9.	21.4.	Consultations on assignment
10.	28.4.	Midterm test.	example test, Test results
11.	5.5.	Consultations on assignment
12.	12.5.	multiobjective optimization: NSGA (Java)	Sources, cheatsheet_nsga2.pdf
13.	19.5.	Deadline for the program implementation, presentations of the program implementations
14.	26.5.	Deadline for the report submission, presentations of the program implementations, Credits