Deep Learning and Computer Vision

Description

During the course, students will get deep knowledge about Deep Learning. The course covers basic essentials about Deep Learning gradually covering more complicated topics. Practical application, use cases and problems that can be solved with Deep Learning will be discussed in the course.

Students will learn how to build successful projects in Deep Learning, what data requirements and metrics are needed to get the best results. They would learn how to set up a development cycle of projects and models improvement pipeline. After the course, students would understand what is convolution and the way the convolutional neural network works as well as how to build convolutional neural networks and apply it to image data.

 

Also, they would know the difference between balanced and unbalanced datasets, overfitting and underfitting problems, the way how to determine such a problem and effective ways for solving it.  

 

Learning outcomes:

Students would get fundamental knowledge in Deep Learning basics and all necessary building blocks for advancing their level of proficiency in the future. Moreover, they would learn how to build and ship deep learning products as well as how to detect potential problems and potential way to solve them

 

Course structure

Part 1.

  • What is Deep Learning
  • Practical Application of Deep Learning
  • Building blocks of Neural Networks
    a. Perceptron
    b. Activation Functions
    c. Neural Network
    d. From Binary to Multiclass classification
    e. Training Neural Networks
    f. Learning Rate
    g. Loss Function
  • Overfitting and Underfitting. Regularization in Deep Learning

 

Part 2

  • Introduction to CNN
  • Learning Visual Features
  • Layers in CNNs
  • Optimization. Gradient Descent
  • CNNs for Classification
  • Transfer Learning
  • Quick Architectures overview
  • Fine Tuning
  • From Supervised to Unsupervised models:
    a. Intro to Deep Generative Models
    b. Intro to Autoencoders and VAEs
    c. Intro to GANs
    d. Intro to Domain Transfer

 

Part 3.
Practical Workshop. Building Object Classification Model using FastAI

 

Exams & certification:
After the successful completion of the course, the participants will get a certificate.

Language: English

 

Price: UAH 24 000

 

__________________________________________________________________________

Course outline:

 

April 3 2020 (12 academic hours)

  • What is Deep Learning
  • Practical Application of Deep Learning
  • Building blocks of Neural Networks
  • Overfitting and Underfitting. Regularization in Deep Learning

 

April 4 2020 (12 academic hours)

  • Introduction to CNN
  • Learning Visual Features
  • Layers in CNNs
  • Optimization. Gradient Descent
  • CNNs for Classification
  • Transfer Learning
  • Quick Architectures overview
  • Fine Tuning

 

April 5 2020 (12 academic hours)

  • From Supervised toUnsupervisedmodels
  • Practical Workshop.Building ObjectClassification Modelusing FastA

 

Contact us to get more details:

Viber, WhatsApp +380 67 441 01 11, [email protected]