Dirbtinis Intelektas – pradedančių studijos

In this section we will learn how to handle numeric information in Python using Numpy library. We will about two of the most important data science concepts – code vectorization and broadcasting as well as Numpy array methods and operations.

In this part of the course we will learn how to use Pandas library to work with tabular data. We will learn how to create, write, read and index Pandas dataframes. We will also learn dataframe methods and how to use them for analysing and visualizing tabular data.

In this section we will learn the fundamentals of machine learning. We will focus on random forests – one of the most powerful and versatile machine learning algorithms. We will also learn how to explore your data, validate your models, handle missing values and other machine learning essentials.

In this section we will learn the basics of deep learning. We will learn about the types of neural networks, activation functions, loss functions and optimizers. We’ll also spend some time learning about the current applications of deep learning in artificial intelligence and why they are behind the current artificial intelligence revolution.

In this part of the course we will move our focus to structured data, which is extremely important in business, but often neglected in most of the deep learning courses. We will do a portfolio project classifying a binary variable.

In this section we will start tackling the most important and the most useful application of artificial intelligence – computer vision. We will concentrate our attention to convolutional neural networks. The main focus of this section are the portfolio projects: you will build image classifiers with vastly different architectures, formats and number of classes. While working on the projects you’ll learn the most advanced architectures, and will practice the most modern training methods.

In this section of the course we will be diving deeper into computer vision and build a reverse image search model capable of finding similar items to the one provided by the user. This project will help us understand the underlying meaning of the weights in the deep learning models and prepare us for the natural language processing and recommender systems sections.

Finally it is time to make some money! We will try to predict stock market movements using recurrent neural networks. While working on this portfolio project we will learn the differences between recurrent neural networks, long short-term memory networks and gated recurrent units, when to use each of those architectures and their strengths and weaknesses.

In this section we will learn how neural networks learn the representations of natural language. While natural language processing (NLP) is totally new to us, we will use the familiar recurrent neural networks to tackle this problem. We will learn the most important NLP concepts and use them to create two NLP portfolio projects.

In this section of the course we will build a recommender system. While not new, recommender systems saw a huge improvement in accuracy with the coming of the deep learning models. While working on the recommender systems we will learn about embeddings and collaborative filtering.

In this section we will return to computer vision once again. We will learn about generative deep learning models and create a convolutional neural network capable of generating images aka deep dreaming.

In this section we will focus on the advanced computer vision topics such as object detection and segmentation. You will learn how to build and apply state of the art computer vision algorithms.

During the final part of the course you will work on your capstone project. You will be able to apply everything that you learned during the course to create a great AI project. While you are working on the project we will also review your Github portfolio, LinkedIn profile and conduct mock interviews to prepare you for getting a job as a deep learning/machine learning/artificial intelligence engineer.