AGH University of Krakow, Poland (Previous name: AGH University of Science and Technology)
Faculty of Electrical Engineering, Automatics, Computer Science and Biomedical Engineering
Department of Biocybernetics and Biomedical Engineering
Field of Biocybernetics, Cognitivistics and Artificial Intelligence
This electable course includes 14 lectures, 14 laboratory classes, and 14 project classes and starts in fall semester of the academic year 2020/2021.
What is this course about?
This course is a continuation of the Computational Intelligence course and is focused on advanced models and advanced optimization techniques to raise the performance of the developed models as well as to give broader knowledge how to construct advanced knowledge-based CI and AI solutions. It is intended to give students a deep knowledge and experience about popular solutions and efficient neural network models as well as to learn how to construct and train intelligent learning systems in order to use them in everyday life and work. During the course, we will deal with the popular and most efficient models and methods of neural networks, fuzzy systems and other learning systems that enable us to find specific highly generalizing models solving difficult tasks. We will also tackle with various CI and AI problems and work with various data and try to model their structures in such a way to optimize operations on them throughout making data available without necessity to search for them. This is a unique feature of associative structures and systems. These models and methods will allow us to form and represent knowledge in a modern and very efficient way which will enable us to mine it and automatically draw conclusions. You will be also able to understand solutions associated with various tasks of motivated learning and cognitive intelligence. We will focus on the optimization and performance of the developed models. We will also try to develop advanced and complex knowledge-based AI and CI systems, e.g. linguistic systems (chatbots), speech recognition systems, or self-adapting systems adaptable to different computational tasks.
We could try to take part in various competitions, e.g. Kaggle!
Today, computational intelligence (CI), machine learning (ML), data mining (DM), and artificial intelligence (AI) achieved already a lot! We have many interesting methods, models and computational tools which let us adapt CI models to many practical tasks. The CI professionals can earn even a six-figure salary for their work, but it is only a tip of the iceberg! We can still achieve a great development in artificial intelligence can be still far more intelligent than we meet today. Today, we rather develop intelligently adapting computational tools than real artificial intelligence! Artificial Intelligence (AI) can do much more and without so much effort from the people's point of view. It must be embodied and equipped with some smartly motivational mechanisms and associative properties (known from our brains) which let it achieve an adequate level.
Lectures will be supplemented by laboratory and project classes during which you will train and adapt the models learned during the lectures on various data using self-developed applications and existing modern programistic tools like RapidMiner, TensorFlow, Keras and Jupiter Notebook. Your hard work and practice will enable you not only to obtain expert knowledge and skills but also to develop your own intelligent learning system implementing a few of the most popular and efficient CI methods.
Expected results of taking a part in this course:
Broad knowledge of neural networks, associative and fuzzy systems as well as other intelligent learning systems.
Novel experience and broaden skills in construction, adaptation and training of neural networks and fuzzy systems.
Ability to construct intelligent learning systems of various kinds, especially deep learning solutions using modern Python 3 and libraries like TensorFlow, Keras, Jupiter Notebook and other open CI tool and environments.
Good and modern practices in modeling, construction, learning and generalization using developed and other tools like RapidMiner, TensorFlow, Keras, Jupiter Notebook and other open CI tools and environments.
Own intelligent learning system to use in your life or work.
Satisfaction of enrollment to this course.
Students are invited to broaden their excellent projects from the project classes to their Master's theses and even scientific papers with help and contribution of the lecturer.
However lectures are generally not obligatory every student who enrolled in this course is invited to take a part in possibly all lectures which will be very intensive and full of news! The presented topics during lectures will be practically accomplished in the laboratory and project classes, so be conscious that the absence in the lectures will cause serious difficulties in the implementation of methods and accomplishing of tasks during the laboratory and project classes that will be evaluated after achieved results, implemented methods, and their quality. This course will be as much interesting, innovative, and valuable for you as the frequency and numerous presence of the participating students. Thus, all participating students are asked to organize their weekly schedule in such a way to be able to take a part in all course activities including lectures! Moreover, your ideas, whishes, and proposals can be also included and used in this course. We can do this course fascinating only together!
We start our classes with laboratory classes, followed by project classes. During the laboratory classes, you will implement the selected and the topmost practically promising tasks in the programming language you like. You can use computers and programming languages installed in the laboratory (Python, C#, C++, Java) or can use your computers together with your compilers and programming languages you prefer. During the project classes, you will implement the selected project that will substantially expand the chosen laboratory class task or will be a new one after your selection from the list of the proposed projects. However, we can also consider your proposals that will be ambitious enough.
Detailed information will be presented at the beginning of this course, i.e. during the first lecture, as well as during the first laboratory classes and the first project classes.
Nikola K. Kasabov, Time-Space, Spiking Neural Networks and Brain-Inspired Artificial Intelligence, In Springer Series on Bio- and Neurosystems, Vol 7., Springer, 2019.
A. Horzyk and J.A. Starzyk, Fast Neural Network Adaptation with Associative Pulsing Neurons, IEEE Xplore, In: 2017 IEEE Symposium Series on Computational Intelligence, pp. 339-346, 2017. - presentation, movie Iris-4, movie Iris-12
A. Horzyk, Human-Like Knowledge Engineering, Generalization and Creativity in Artificial Neural Associative Systems, Springer Verlag, AISC 11156, ISSN 2194-5357, ISBN 978-3-319-19089-1, ISBN 978-3-319-19090-7 (eBook), DOI 10.1007/978-3-319-19090-7, Springer, Switzerland, 2016, pp. 39-51.
A. Horzyk, Innovative types and abilities of neural networks based on associative mechanisms and a new associative model of neurons – the invited talk and paper at the International Conference ICAISC 2015, Springer Verlag, LNAI 9119, 2015, pp. 26-38, DOI 10.1007/978-3-319-19324-3_3.
