Siirry suoraan sisältöön

Fundamentals of Deep Learning (5 cr)

Code: TT00EO93-3001

General information


Timing

01.01.2021 - 31.12.2021

Number of ECTS credits allocated

5 op

Virtual portion

5 op

Mode of delivery

Distance learning

Unit

ICT ja tuotantotalous

Campus

Karaportti 2

Teaching languages

  • English

Seats

0 - 5000

Degree programmes

  • Tieto- ja viestintätekniikan tutkinto-ohjelma

Teachers

  • Virve Prami

Groups

  • ATX21TV
    NonStop virtuaaliopinnot vuosi 2021

Objective

Deep learning is a new area of machine learning that is concerned with algorithms inspired by the brain’s structure and functionality. Deep learning is evolving as one of the crucial practices in industries like manufacturing, hospitality, digital assistants, automotive, etc. This is an introductory course that provides a unique opportunity for the student to get familiar with the basic concepts of deep learning. After passing this course, the student will be familiar with different types of deep learning architectures and models and the intuitions behind them. In fact, the student gets acquainted with variations of the neural network algorithm, which are used for various types of data. Furthermore, the most critical concepts and techniques of deep learning in today’s industry have been discussed.
This course is 100% virtual, thanks to the comprehensive tutorial videos and content prepared for this course.
The student will pass this course after submitting the required quiz, assignments, and the final project.

Content

1. Introduction to Machine Learning:
The Concept of Learning – Data – Machine Learning – Supervised Learning – Unsupervised Learning – Applications
2. Introduction to Deep Learning:
What is Deep Learning? – Deep Learning Architectures – Deep Learning vs. Machine Learning – Artificial Neural Network vs. Biological Neural Network – History of Deep Learning
3. Feed-Forward Neural Networks:
A Single Neuron – Neural Networks – Training Neural Networks – Prediction and Evaluation – Bias and Variance – Regularization – Model Selection and Hyperparameters
4. Convolutional Neural Networks:
Motivation for Convolutional Layers – Convolutional Layer – Pooling Layer – Convolutional Networks – Analogy Between CNNs and Human Visual System – Popular CNN Architectures – Applications
5. Sequence Models:
Motivation for Sequence Models – Recurrent Neural Networks – Variations of RNN – Encoder-Decoder – Attention Mechanism – Applications
6. What’s More?
Transfer Learning – Autoencoders – Generative Adversarial Networks – Deep Learning Frameworks
7. Final Tasks:
Project – Self-study Essay

Location and time

Course can be done in own pace in TechClass environment.

Materials

Online.

Teaching methods

100% online Self-Study course.

- Tutorial Videos
- Exercises
- Quiz
- Project
- Self-study

Employer connections

N/A

Exam schedules

Online.

International connections

N/A

Completion alternatives

N/A

Student workload

Lectures = 40h
Exercises = 15h
Self-study = 40h
Quizzes = 10h
Project = 30h
Total = 135 hours

Content scheduling

Up to student her-/himself.

Evaluation scale

Hyväksytty/Hylätty

Assessment criteria, satisfactory (1)

- The student knows the basic definitions of Deep Learning.
- The student knows the general difference between Machine Learning and Deep Learning.
- The student knows the primary types of Deep Learning methods.
- The student is familiar with the concept of mapping that learning algorithms perform.
- The student is familiar with variations of Neural Network algorithms used for different applications.

Assessment criteria, good (3)

- The student is familiar with basic concepts of the Neural Network algorithm such as wight and neuron.
- The student knows the reasons for the advent of Deep Learning and its history.
- The student knows the intuitions behind Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs).
- The student is familiar with the steps of training and evaluating Neural Networks.
- The student knows the applications of CNNs and RNNs.

Assessment criteria, excellent (5)

- The student understands the concepts and intuitions behind the Neural Network algorithm.
- The student knows the concepts of regularization, overfitting, and hyperparameters selection.
- The student knows layers of CNNs.
- The student is familiar with the popular architectures of CNNs.

Assessment methods and criteria

Exercises 50%
Quizzes 25%
Project 25%