Uniwersytet Jagielloński: Biologicznie Inspirowane Paradygmaty Obliczeniowe

Syllabus Sesja Zimowa
Rok Akademicki
2015/2016

Typ zajęć: Wykład, 12 godzin, Laboratorium, 24 godziny
Punkty ECTS: 6
Zaliczenie wkładu: Na podstawie punktów za prezentacje wyników eksperymetów (6 ustnych referatów i pisemnych sprawozdań)
Zaliczenie ćwiczeń: Na podstawie punktów za projekty (6 dwu-tygodniowych projektów)
Wymagania wstępne: Zaawansowana umiejętność programowania (algorytmy i struktury danych) oraz znajomość analizy matematycznej i algebry liniowej

Skrócony Opis Kursu po Polsku

Jedna z dróg do zapewnienia postępu w dziedzinie informatyki jest poszukiwanie nowych metod i technik, które mogą umożliwić niekonwecjonalne podejścia do rozwiazywania problemów trudnych – lub wręcz niemożliwych – do rozwiazania tradycyjnymi metodami analitycznymi. Celem kursu jest zapoznanie studentów z kilkoma paradygmatami obliczeniowymi dla ktorych inspiracja sa naturalne zjawiska i systemy biologiczne takie jak: ludzki mózg, a szczególnie kora mozgowa; ewolucja i genetyka; układ immunologiczny; inteligencja roju; i sztuczne zycie. W kazdym z kilku blokow tematycznych studenci po przenalizowaniu teorii zaimplementuja symulatory i przeprowadza doswiadzenia w ich wykorzystaniu w wybranych zastosowaniach. Studenci przedstawia wyniki badań w formie ustnych referatów i pisemnych raportów.

Efekty kształcenia:

Po pozytywnym ukończeniu kursu jego absolwent:

E1 Zna i rozumie różnicę w analitycznym i heurystyczym podejściu do rozwiazywania problemów
E2 Potrafi zastosować biologicznie inspirowane metody rozwiazywania problemów
E3 Potrafi wybrać odpowiednie metody w zależności od rodzaju problemów
E4 Rozumie biologiczne fundamenty stosowanych paradygmatów
E5 Potrafi przeprowadzić analityczna ewaluację stosowanych metod
E6 Potrafi opublikować rezultaty eksperymentow w formie technicznych sprawozdań
E7 Potrafi efektywnie używać środowiska IPython Notebook z bibliotekami NumPy, SciPy, i matplotlib

Literatura Wymagana:

Bieszczad, Andrzej – Biologically-Inspire Computing – Lecture Notes – wolna od opłat electroniczna publikacja dla studentów.

Literatura Uzupełniajaca:

1) de Castro, L. N. – Fundamentals of Natural Computing: Basic Concepts, Algorithms, and Applications – Chapman & Hall/CRC – 2007 – ISBN: 978-1-58488-643-9
2) Lamm, E., and Unger, R. – Biological Computation – Chapman & Hall/CRC – 2011 – ISBN: 978-1-4200-8795-6
3) Floreano, D., and Mattiussi, C. – Bio-Inspired Artificial Intelligence: Theories, Methods, and Technologies – The MIT Press – 2008 – ISBN: 978-0-262-06271-8

Course Description in English

From the onset of Computer Science, computers were likened to brains, and even called “electronic brains”, in spite of the fact that a von Neumann architecture is not rooted in Biology. Nevertheless, over the years, researchers looked at biological systems for inspirations to create artificial intelligence, the Holy Grail of Computer Science. While the goal of creating artificially intelligent systems is evasive, many ideas from that quest proved very interesting, and what’s more important, found practical use. Neural Networks are used from primitive echo cancellation devices in telephones to sophisticated controllers of modern car engines and in data mining credit rating system. Evolutionary techniques found applications in engineering design, pattern classification, data mining, optimization, and elsewhere. Least cost routing can be approached with methodologies based on Swarm Intelligence, and Artificial Immune Systems may assist in network intrusion detection and medical diagnosis amongst other things.

This course is a comprehensive introduction to several of such “soft computing” paradigms.

Course Objectives

To assure progress in the field of computing, it is important to continue explorations of novel ideas, methodologies and techniques that may allow to address problems in unconventional ways that — quite often — are the only solutions. Biologically-inspired Computational Intelligence uses sub-symbolic approaches that are contrasted with methods of more traditional Artificial Intelligence that employ analytical systems using symbols to construct problem solutions. The distinctive characteristics of the techniques that the students will study in this course is that they do not use symbols for knowledge representation. In fact, often it is not even possible to locate particular elements of the knowledge in the system as it is distributed amongst its components. In spite of that, analytical methods can be, and are, used to examine the inner workings of such systems, test their limits, and explore their parameters to maximize the odds for obtaining practical results.

The objective of this course is to equip the students with skills in using unconventional methodologies that can be used to solve problems that are hard, or outright impossible, to treat with traditional analytical means.

After taking this class the students will be able to:

  • discuss the history of Biological foundations of a selection of computing paradigms,
  • design, implement, and utilize systems that use Biologically-inspired means to solve problems,
  • incorporate sub-symbolic problem solving techniques in traditional methodologies, and
  • explore Biology for further inspirations.

Course Outline

  • Cellular Automata
  • Neurosolver
  • Evolutionary Techniques
  • Swarm Intelligence
  • Artificial Immune Systems

Course Location

Instytut Informatyki i Matematyki Komputerowej Pokój 1064/1065
Wydział Informatyki i Matematyki Uniwersytetu Jagiellońskiego w Krakowie

Instructor

Prof. Andrzej (AJ) Bieszczad
California State University Channel Islands

Office: TBD
Email: aj.bieszczad@csuci.edu
Phone: TBD

Instructor Communication Policy

I will respond to your inquiry within 24 hours Mon-Fri. If I do not reply in this timeframe, please assume I did not receive your email and contact me again. Manage your communications diligently.

Supporting Texts

  1. de Castro, L .N., Fundamentals of Natural Computing, Chapman & Hall/CRC, 2007, ISBN: 978-1584886433.
  2. Lamm, E., and Unger, R., Biological Computation, Chapman & Hall/CRC, 2011, ISBN: 978-1420087956.
  3. Floreano, D., and Mattiussi, C., Bio-Inspired Artificial Intelligence: Theories, Methods, and Technologies, The MIT Press, 2008, ISBN: 978-0262062718.
  4. Instructional Approach

This course will have five study blocks. Each block will start with a lecture explaining the topic, Then, the students will research, solve, implement, experiment, and report a practical problem employing the techniques from the theory.

Detailed submission instructions will be included with the assignments, but in general they need to include complete code for the implementation and a research paper that describes the experiments with abundant use of visualization (e.g., visualization of data, progress of the simulation, graphs for evaluating and comparing, results, etc.).

IPython Notebook with SciPy, NumPy, and Matplotlib must be used in implementations and experiments by all students.

Research papers must be in form of an editable IPython Notebook and be of a quality appropriate for a scientific conference.

Grading

Each course block will be graded based equally on:

  • the quality of the research papers,
  • the presentations, and
  • the code.

The final grade from the course will be an average of the grades from each block.

Academic Honesty

The university, the course, the labs and the instructor are here for the students so they can acquire sufficient knowledge to open a window of opportunity for them in their future careers. Any academic dishonesty limits the student’s chances to succeed.

Students with Disabilities

Students with disabilities needing accommodation should discuss required accommodations with the instructor.

Subject to Change

This syllabus and schedule are subject to change in the event of extenuating circumstances.