1	Course Title:	DEEP LEARNING BY PYTHON
2	Course Code:	IYS4220
3	Type of Course:	Optional
4	Level of Course:	First Cycle
5	Year of Study:	4
6	Semester:	8
7	ECTS Credits Allocated:	6
8	Theoretical (hour/week):	3
9	Practice (hour/week) :	0
10	Laboratory (hour/week) :	0
11	Prerequisites:	There aren' ant prerequerements
12	Recommended optional programme components:	None
13	Language:	Turkish
14	Mode of Delivery:	Face to face
15	Course Coordinator:	Prof. Dr. MELİH ENGİN
16	Course Lecturers:	Doç. Dr. Melih Engin
17	Contactinformation of the Course Coordinator:	melihengin@uludag.edu.tr
18	Website:
19	Objective of the Course:	Presenting the methods that can be used to learn high-level features obtained from different types of data by using deep architectures and showing how these methods can be applied for different purposes, from image recognition to robot control.
20	Contribution of the Course to Professional Development	History and theoretical advantages of deep learning, Basic artificial neural network architectures and learning algorithms that can be used for deep learning, Organization of Distributed Models, Optimization Techniques for Training Deep Models, Convolutional networks, Feedback and recursive networks, Autoencoders and Linear Factor Models, Learning by Representation , Deep Generative Models – Boltzman Machines.

21

Learning Outcomes:

1 Students can evaluate common deep learning methods for their effectiveness.; 2 Students can evaluate the advantages and disadvantages of the deep learning method that is considered to be used.; 3 Students can design and test basic deep learning solutions.; 4 Students determine and implement the appropriate deep learning architecture and algorithm for the envisioned solution.; 5 Students have knowledge about deep model editing and optimization methods.;

22	Course Content:

Week	Theoretical	Practical
1	Introduction – History and Theoretical Foundations
2	Mathematical Foundations: Linear Algebra, Probability and Information Theory
3	Artificial Neural Networks Basic Information
4	Feed Forward Deep Networks
5	Organizing Deep or Distributed Models
6	Optimization Techniques for Training Deep Models
7	Convolutional Networks
8	Convolutional Networks
9	Autoencoders and Linear Factor Models
10	Learning through Representation
11	Deep Generative Models – Boltzman Machines
12	Deep Generative Models – Boltzman Machines
13	Project Presentations
14	Project Presentations

23	Textbooks, References and/or Other Materials:	Hinton, G. E, Osindero, S., and Teh, Y. W. (2006). A fast learning algorithm for deep belief nets. Neural Computation, 18:1527-1554. Bengio, Y., Lamblin, P., Popovici, P., Larochelle, H. (2007). Greedy Layer-Wise Training of Deep Networks, Advances in Neural Information Processing Systems.
24	Assesment

TERM LEARNING ACTIVITIES	NUMBER	PERCENT
Midterm Exam	1	40
Quiz	0	0
Homeworks, Performances	0	0
Final Exam	1	60
Total	2	100
Contribution of Term (Year) Learning Activities to Success Grade		40
Contribution of Final Exam to Success Grade		60
Total		100
Measurement and Evaluation Techniques Used in the Course	Classic Quiz
Information	Quiz on Study Topics

25

ECTS / WORK LOAD TABLE

Activites	NUMBER	TIME [Hour]	Total WorkLoad [Hour]
Theoretical	14	3	42
Practicals/Labs	0	0	0
Self Study and Preparation	10	13	130
Homeworks, Performances	0	0	0
Projects	0	0	0
Field Studies	0	0	0
Midtermexams	1	4	4
Others	0	0	0
Final Exams	1	4	4
Total WorkLoad			180
Total workload/ 30 hr			6
ECTS Credit of the Course			6

26	CONTRIBUTION OF LEARNING OUTCOMES TO PROGRAMME QUALIFICATIONS
PQ1 PQ2 PQ3 PQ4 PQ5 PQ6 PQ7 PQ8 PQ9 PQ10 PQ11 LO1 1 3 4 5 3 4 1 2 5 3 5 LO2 2 4 5 5 3 4 2 2 5 3 5 LO3 2 4 5 5 3 4 2 2 5 3 5 LO4 3 4 4 5 3 4 0 4 5 3 5 LO5 5 5 5 5 3 4 5 5 3 5 3

LO: Learning Objectives

PQ: Program Qualifications

Contribution Level:

1 Very Low

2 Low

3 Medium

4 High

5 Very High