Machine learning on big data - free course from Open Education, training 5 weeks, from 5 to 14 hours per week, Date: December 3, 2023.

1. Preparing data for training

Today, machine learning works effectively when we have large amounts of labeled data. This week we will look at what data formats and markup exist and how this markup can be collected

2. Training classical models on big data

This week we'll learn how to train classical algorithms (linear models and decision trees) on big data.

3. Building recommender systems

We will look at how we can parallelize classic algorithms used in recommendation systems.

4. Analysis of large volumes of text information

Let's consider machine learning problems on texts. Let's talk about text preprocessing, and how to get a structured representation of text data using models such as word2vec and BERT.

5. Training deep neural networks

We will learn how to parallelize the training of modern neural networks, how Horovod and Parameter Server work inside, and talk about Transfer Learning.