The creation of superintelligence is thought to be the biggest leap ever made since the Industrial Revolution. For this reason, Turing School in 2016. has developed an Artificial Intelligence course to introduce this technology from an economic and technological point of view. It is a Level 2 program designed for students with basic programming knowledge who seek to further their study in computer science. This programming course is taught at Erasmus Rotterdam and Oxford Universities. We have adapted it according to Lithuanian study programs, so you have the opportunity to study it here - in Vilnius!
The course will begin with the design aspects of the Python programming language and its procedural, objective, and functional programming languages. Learning will take place using Jupyter Lab - a modern programming environment that allows you to create interactive Jupyter analysis books in a browser. Although the browser will be used, the calculations will be performed using the most powerful graphics card cluster in the Baltic States for artificial intelligence.
We will deepen our knowledge of machine learning, which is the intersection between data science and artificial intelligence. We will talk about different types of machine learning: supervised and non-supervised and tasks: classification and regression. Using the Python sci-kit learn, students will create artificial intelligence models to determine whether or not a received SMS message is promotional.
We will start talking about artificial neural networks, their training procedure, and how they can be used for a variety of machine learning tasks. To build multilayer neural networks, we will assimilate the advanced Keras API of the Tensorflow 2.0 Deep Learning Library, with which we will develop models that can recognize images of dogs and cats.
We will talk about convolutional networks, their biological similarity to the mammalian visual system, and the computer resources needed to develop models of this type of deep neural network. We will formulate recognition (discrimination) tasks and create models capable of discovering a human face in a photograph.
Probably one of the most interesting areas of application of deep learning is the synthesis of images with generative opposing networks that allow the generation of original images in the style of various artists. Using these technologies, we will create images of Vilnius in the style of Mikalojus Konstantinas Čiurlionis!
Much of the data: text, sound, or the human genome can be represented as sequences: consecutive elements in a certain order. To learn to recognize different sequences, we will look at recursive neural networks and use them to recognize the sentiments of Twitter messages.
In the last month of learning, we will dedicate ourselves to the branch of machine learning - an incentive for learning, with which we will create agents who learn to perform various tasks by interacting and getting to know the environment around them. It is this branch of artificial intelligence, together with deep learning, that allowed the creation of the famous Go game agents - AlphaGo and AlphaZero. During this month’s hands-on exercise, we will teach computer programs to play ATARI games using the OpenAI gym simulation environment.
At the end of the course, we will work on an individual (or group) project according to a chosen or self-invented topic.
Presentations of final theses. A Turing School certificate attesting to the completion of the taught program is awarded.