1 day

Version 1.0

The course is designed for managers and executives who would like to learn how Big Data and Artificial Intelligence can change and improve their business. The course discusses the key provisions of modern technologies and immediately gives specific examples of the use of such solutions. The course is designed for a wide audience, regardless of the business or organization the participants work in.

There are no prerequisites for participants. The course is conducted for a single customer group online or at the customer's premises.

Before the start of the course, the customer can send preferences which areas of Artificial Intelligence are particular interest to him. This will be taken into account when submitting the material. 

Recommended group size is from 6 to 16 attendees.

The duration of the course is 5 hours plus approximately 1 hour for Q&A. Classes take one day and are held at a convenient time for the customer. Usually, these are morning classes from 9 am to 3 pm. Time, date and schedule are determined by the customer.

Topic 1: Big Data and AI.

Data Mining and Machine Learning

· What is machine learning and data mining.

· How it can help business.

· How machine learning projects work: CRISP-DM

· Examples of machine learning projects. The CRISP-DM methodology.

· How the learning process works: test, training and validation samples.

· Machine learning metrics.

· Types of tasks: unsupervised (clustering, anomaly detection) and supervised (classification, regression, ranking).

· Algorithms: k-nn, linear models, decision trees, ensembles of models, neural networks.

Big and Small data

· How big data helps businesses. In which cases and small data is enough

· Volume/Variety/Velocity.

· Which tasks need a lot of data.

· A breakthrough in neural networks, which allowed using a large amount of data to improve the quality of models.

Organization and management of AI projects

· How teams are organized to conduct ML projects.

· Agile approaches in AI, typical team composition and roles in it.

· Which tasks can AI help solve, and which ones can't.

Topic 2: Interpretation of metrics

Gigerenzer's case about metrics on the example of a mammogram.

· How to look at metrics correctly and draw conclusions.

· Description: We are trying to guess from the task what conclusion we can make based on the results of the test.

· Bayesian approach to statistics

Solving a case about anomalies at an industrial plant

· How to organize an anomaly detection project correctly.

· Description: We come up with a common approach in teams.

· We are discussing which metrics are important here (a reference to the paragraph above).

Topic 3: Machine Learning Projects

Case of a marketing campaign for the sale of bicycles

· How to build machine learning projects, use historical data and uate the economic impact.

· Description: We have the data on historical bicycle sales via direct mail.

· We are planning the next campaign based on potential customers and we want to maximize net profit.

· We need to come up with a strategy for mailing.

Ranking in nettiauto.com: how to organize a project

· How to plan ranking projects and understand their feasibility.

· Description: In the process of solving, we learn the following ideas: baseline algorithm, degrading A/B tests for understanding value, basic approaches to solving the ranking problem