Moduł oferowany także w ramach programów studiów:
Informacje ogólne:
Nazwa:
Basics of Machine Learning
Tok studiów:
2017/2018
Kod:
AMA-2-094-MZ-s
Wydział:
Matematyki Stosowanej
Poziom studiów:
Studia II stopnia
Specjalność:
Matematyka w zarządzaniu
Kierunek:
Matematyka
Semestr:
0
Profil kształcenia:
Ogólnoakademicki (A)
Język wykładowy:
Angielski
Forma i tryb studiów:
Stacjonarne
Strona www:
 
Osoba odpowiedzialna:
dr Morkisz Paweł (morkiszp@agh.edu.pl)
Osoby prowadzące:
dr Morkisz Paweł (morkiszp@agh.edu.pl)
Krótka charakterystyka modułu

The course will be devoted to practical applications of Python, classical machine learning algorithms, and deep learning algorithms.

Opis efektów kształcenia dla modułu zajęć
Kod EKM Student, który zaliczył moduł zajęć wie/umie/potrafi Powiązania z EKK Sposób weryfikacji efektów kształcenia (forma zaliczeń)
Wiedza
M_W001 Student knows and understands basic machine learning methods MA2A_W08 Projekt,
Egzamin
M_W002 Student has detailed knowledge about algorithms and their computational complexity within the topics considered in the course MA2A_W07, MA2A_W08 Projekt,
Egzamin
M_W003 Student knows basic concepts of Python programming language MA2A_W08, MA2A_W12 Projekt
Umiejętności
M_U001 Student can gain information from the literature, data bases and other sources and apply gained knowledge to solving practical problems. MA2A_U13, MA2A_U16 Projekt
M_U002 Student knows simple classification, regression, and clustering machine learning methods and is able to choose properly and use them in practice MA2A_U19, MA2A_U16, MA2A_U12 Projekt,
Egzamin,
Aktywność na zajęciach
Kompetencje społeczne
M_K001 Student is able to work in team, understands the necessity of systematical work with any long term projects. MA2A_K03 Projekt,
Egzamin
M_K002 Student can evaluate presented solution of a problem and track missing elements or errors MA2A_K02, MA2A_K01 Projekt,
Egzamin
Matryca efektów kształcenia w odniesieniu do form zajęć
Kod EKM Student, który zaliczył moduł zajęć wie/umie/potrafi Forma zajęć
Wykład
Ćwicz. aud
Ćwicz. lab
Ćw. proj.
Konw.
Zaj. sem.
Zaj. prakt
Zaj. terenowe
Zaj. warsztatowe
Inne
E-learning
Wiedza
M_W001 Student knows and understands basic machine learning methods - - - - + - - - - - -
M_W002 Student has detailed knowledge about algorithms and their computational complexity within the topics considered in the course - - - - + - - - - - -
M_W003 Student knows basic concepts of Python programming language - - - - + - - - - - -
Umiejętności
M_U001 Student can gain information from the literature, data bases and other sources and apply gained knowledge to solving practical problems. - - - - + - - - - - -
M_U002 Student knows simple classification, regression, and clustering machine learning methods and is able to choose properly and use them in practice - - - - - - - - - - -
Kompetencje społeczne
M_K001 Student is able to work in team, understands the necessity of systematical work with any long term projects. - - - - + - - - - - -
M_K002 Student can evaluate presented solution of a problem and track missing elements or errors - - - - + - - - - - -
Treść modułu zajęć (program wykładów i pozostałych zajęć)
Konwersatorium:
  1. Discussion seminar (in lab)

    The seminar will be a combination of lectures on practical applications of Python, classical machine learning algorithms, and deep learning algorithms, presentations of students and discussions on practical problems arising in implementation of programs solving particular problems. Mathematical roots of each algorithm together with the complexity estimate will be presented. Topics considered during seminar include (but are not limited to):

    1. Python programming language (including libraries: numpy, pandas, scikit learn, keras).

    2. Discussion of problems and methods related to classification, regression, and clustering.

    3. Problem of over- and under-fitting, parameters fine tuning, models validation.

    4. Usage of decision trees based algorithms, e.g. random forest, XGBoost

    5. Deep learning algorithms (dense layers, convolutional layers, batch normalization, dropout, LSTM)

    6. Differences and difficulties of supervised, semi-supervised, and unsupervised learning.

    Students will be given some problems for individual solving and discussion during seminar. Each student will pick a project, i.e. a problem including some freely available data set and well defined purpose, for machine learning algorithms application. Student will construct his own solution based on the obtained knowledge and the project will be assessed.

  2. Discussion seminar (in lab)

    The seminar will be a combination of lectures on practical applications of Python, classical machine learning algorithms, and deep learning algorithms, presentations of students and discussions on practical problems arising in implementation of programs solving particular problems. Mathematical roots of each algorithm together with the complexity estimate will be presented. Topics considered during seminar include (but are not limited to):

    1. Python programming language (including libraries: numpy, pandas, scikit learn, keras).

    2. Discussion of problems and methods related to classification, regression, and clustering.

    3. Problem of over- and under-fitting, parameters fine tuning, models validation.

    4. Usage of decision trees based algorithms, e.g. random forest, XGBoost

    5. Deep learning algorithms (dense layers, convolutional layers, batch normalization, dropout, LSTM)

    6. Differences and difficulties of supervised, semi-supervised, and unsupervised learning.

    Students will be given some problems for individual solving and discussion during seminar. Each student will pick a project, i.e. a problem including some freely available data set and well defined purpose, for machine learning algorithms application. Student will construct his own solution based on the obtained knowledge and the project will be assessed.

Nakład pracy studenta (bilans punktów ECTS)
Forma aktywności studenta Obciążenie studenta
Sumaryczne obciążenie pracą studenta 112 godz
Punkty ECTS za moduł 4 ECTS
Udział w konwersatoriach 30 godz
Wykonanie projektu 20 godz
Samodzielne studiowanie tematyki zajęć 60 godz
Egzamin lub kolokwium zaliczeniowe 2 godz
Pozostałe informacje
Sposób obliczania oceny końcowej:

Ocena końcowa (OK) jest oceną z zaliczenia konwersatorium (K) otrzymaną na podstawie projektu:
OK = K.
Istnieje też możliwość zaliczenia przedmiotu z egzaminem. W tej opcji ocena końcowa (OK) jest średnią arytmetyczną oceny z Egzaminu (E) oraz oceny z konwersatorium (K):
OK = 1/2 x E + 1/2 x K.

Wymagania wstępne i dodatkowe:

Passed two semesters of programming courses – including object oriented programming language – such as C++, passed course of introduction to numerical analysis.

Zalecana literatura i pomoce naukowe:

1. S. Raschka, Python Machine Learning, Packt Publishing Ltd, 2015,
2. Trevor Hastie, Robert Tibshirani and Jerome H. Friedman. The Elements of Statistical Learning, Springer, 2001,
3. A. Courville. I. Goodfellow, Y. Bengio. Deep Learning. MIT Press, 2016.

Publikacje naukowe osób prowadzących zajęcia związane z tematyką modułu:

1. Kusiak J., Morkisz P., Oprocha P., Sztangret Ł., On aggregation of stages in multi-criteria optimization of chain structured processes, ICAISC 2016, Lecture Notes in Artificial Intelligence, cop. 2016, pp. 411-419,
2. Kusiak J., Morkisz P., Oprocha P., Sztangret Ł., Multi-criteria optimization strategies for production chains, MATEC Web of Conferences, 2016 vol. 80, pp 1-6
3. Kusiak J. Morkisz P., Oprocha P., Pietrucha W., Sztangret Ł., Validation of optimization strategies using the linear structured production chains, Applied mathematics and computer science: proceedings of the 1st international conference on Applied mathematics and computer science : 27–29 January 2017, AIP Publishing, 2017, S. 020067-1–020067-6
4. Marzec M., Morkisz P., Wojdyła J., Uhl T., Intelligent predictive maintenance system , Proceedings of SAI intelligent systems conference (IntelliSys) 2016, Vol. 1, Springer International Publishing, cop. 2018. — (Lecture Notes in Networks and Systems, 794–804

Informacje dodatkowe:

Przedmiot dopuszcza wersję “z egzaminem” – 4 ECTS