Machine Learning in Practice

Machine Learning in Practice is an in-depth course designed to help you bridge the gap between theoretical knowledge and real-world application of machine learning. This course is ideal for data scientists, developers, analysts, and managers who want to understand the practical aspects of implementing machine learning models in business environments.

The course begins with an overview of tasks that are best solved by machine learning, emphasizing the importance of having a domain expert rather than expecting non-specialists to learn everything on the job. You’ll gain insights into the kinds of problems ML can solve more effectively than traditional methods and understand the pitfalls of relying on generalists.

You will then dive into the crucial steps of preparing, cleaning, and exploring data. This module covers how to derive meaningful insights from the data provided by businesses and the roles of domain analysts versus data scientists in this process. You'll learn the importance of prioritizing tasks and understand what steps are necessary for successful data preparation.

The course also provides comprehensive training on model evaluation, focusing on both business and technical metrics. You’ll learn how to use metrics for classification, regression, and clustering tasks, understand the significance of cross-validation, and avoid common pitfalls like overfitting.

Further, you’ll explore model optimization techniques, including parameter management, trait selection, and the use of ensembles. This module offers practical tools and strategies for finding the best parameters and methods to optimize your models.

The final module focuses on neural networks, covering foundational concepts such as gradient descent and backpropagation and more advanced topics like generative networks, convolutions, recurrency, and techniques for improving model performance like batch normalization and dropout.

With a balanced approach between theory (5 hours) and practice (7 hours), this course ensures you gain both the knowledge and the hands-on experience needed to apply machine learning effectively in your projects.

Upon completing this course, participants will:

• Understand which tasks are best suited for machine learning and the importance of domain expertise.
• Prepare, clean, and explore data to uncover insights and ensure data quality.
• Evaluate machine learning models using appropriate business and technical metrics, including cross-validation and error matrices.
• Optimize machine learning models by managing parameters, selecting traits, and employing ensemble methods.
• Implement neural networks with a solid understanding of key concepts like gradient descent, backpropagation, and advanced neural network techniques.

This course offers a well-rounded mix of theoretical instruction and practical exercises, allowing you to develop a comprehensive understanding of how to apply machine learning in practice. With 12 hours of content split evenly between theory and practice, you’ll be equipped to tackle real-world machine learning challenges with confidence.

Upon completion of the "Machine Learning in Practice" course, trainees will be able to:

• Identify tasks that are best solved by machine learning and understand the risks of assigning them to non-specialists.
• Effectively prepare and explore data, distinguishing between the roles of domain analysts and data scientists.
• Evaluate machine learning models using business-relevant metrics and avoid common pitfalls like overfitting.
• Optimize machine learning models by selecting the right parameters, traits, and employing advanced techniques like ensembles.
• Implement and fine-tune neural networks for a variety of practical applications.

1. Task overview (theory - 1 hour)
   Tasks that are better solved by machine learning. What will happen if, instead of a Data Scientist, you hire a non-specialist in a given domain (just a developer/analyst/manager), expecting that they will learn everything in the process?
2. Preparing, cleaning, and exploring data (theory - 1 hour; practice – 1 hour)
   How to gain insight into data provided by business (and find whatever order in it at all). Processing steps. What can and should be done by domain analysts, and what should better be done by a Data Scientist. Priorities in solving a specific task.
3. Model evaluation (theory - 1 hour; practice – 2 hour)
   Business metrics and technical metrics. Metrics for tasks of classification and regression, error matrix. Internal and external metrics of clustering quality. Cross-validation. Overfitting.
4. Optimization (theory - 1 hours; practice – 2 hours)
   What makes one model better than another: parameters, traits, and ensembles. Parameter management. Traits selection practice. Overview of tools for searching the best parameters/traits/methods.
5. Neural networks (theory 1 hours; practice – 2 hours). Gradient descent, backpropagation, generative neural networks, convolutions, recurrency, batch normalization, dropout, activation functions, batching

Total: theory 5h, practice 7h