Machine learning. Advanced - free course from Otus, training 5 months, Date: December 4, 2023.

Miscellaneous / by admin / December 08, 2023

You will master advanced machine learning techniques that will allow you to feel confident in leading Middle/Senior positions and cope even with non-standard tasks.

You will expand your range of tools available for work. Moreover, even for topics such as Bayesian methods and reinforcement learning, which are usually taught exclusively in the form of theory, we selected real working cases from our practices.

A separate module is dedicated to working in production: setting up the environment, optimizing code, building end-to-end pipelines and implementing solutions.

Versatile project assignments
During the course, you will complete several practical assignments to consolidate your skills on the topics covered. Each assignment is a hands-on data analytics project that solves a specific machine learning application.

Who is this course for?
For analysts, programmers and data scientists practicing machine learning. The course will help you expand your capabilities and move further along your career path.

instagram viewer

After completing the course you will be able to:
Set up the environment and write production code ready for implementation
Work with AutoML approaches and understand the limitations in their use
Understand and be able to apply Bayesian methods and reinforcement learning to relevant problems
Solve non-standard problems arising in recommender systems, time series and graphs

I started at school with a soldering iron in my hands. Then there was the ZX Spectrum. I went to university to major in engineering. There is a lot of interesting things in mechanics, but in 2008 interest in IT took over: computer networks -> Delphi -> PHP -> Python. There have been experiments with other languages, but I want to write in this language. Participated in projects to automate business processes using neural networks (Maxim taxi ordering service), and develop information systems in medicine. Worked with GIS systems and image processing using Python. In teaching, the position is: “If someone cannot explain something complex in simple words, it means they are not very good at it yet.” understands.”Education: Kurgan University, Department of Security of Information and Automated Systems, Ph.D. Graduated in 2002 Kurgan State University with a degree in “Multi-purpose tracked and wheeled vehicles.” In 2005 he defended his dissertation on continuously variable transmissions. Since then, he has been officially employed at the university (KSU). Teacher

Works as a data analyst at the hedge fund Meson Capital. Engaged in the construction of various models that predict behavior on the stock market. Before that, I spent more than 9 years solving business problems based on machine...

Works as a data analyst at the hedge fund Meson Capital. Engaged in the construction of various models that predict behavior on the stock market. Before that, he spent more than 9 years solving business problems based on machine learning in companies such as Alfa Bank, SberMegaMarket, HomeCredit, LPSU MIPT, building models of computer vision, natural language processing and time rows. He is a guest lecturer at MIPT, where he teaches his own course “Practical ML.” Valentin completed his master’s degree at MIPT. His interests include implementing and building infrastructure for data-driven solutions. Teacher

Experienced developer, scientist and Machine/Deep learning expert with experience in recommender systems. He has more than 30 scientific publications in Russian and foreign languages, and defended his PhD thesis on the analysis and forecasting of time series. Graduated from the Faculty of Computer Science at the National Research University Moscow Power Engineering Institute, where in 2008. received a bachelor's degree, a master's degree in 2010, and a candidate of technical sciences in 2014. Even before starting work on his dissertation, I became interested in data analysis and, when implementing my first significant project, I went from an ordinary programmer to the head of the development department. For about 10 years he taught related disciplines at the National Research University Moscow Power Engineering Institute, being an associate professor of the department. Leads Data Science teams developing projects in the field of NLP, RecSys, Time Series and Computer Vision Teacher

Advanced Machine Learning. AutoML
-Topic 1.Production Code of the project using the example of a classification/regression problem, Virtual environments, dependency management, pypi/gemfury
-Topic 2. Practical lesson - Code optimization, parallelization, multiprocessing, pandas acceleration, Modin for Pandas
-Topic 3.Advanced Data Preprocessing. Categorical Encodings
-Topic 4.Featuretools - are you going to come up with features for me?
-Topic 5.H2O and TPOT - are you going to build models for me?

Production
-Topic 6. Practical lesson - Construction of end-to-end pipelines and serialization of models
-Topic 7.REST architecture: Flask API
-Topic 8.Docker: Structure, application, deployment
-Topic 9.Kubernetes, container orchestration
-Topic 10. Practical lesson on working in production: deploying Docker to AWS

Time series
-Topic 11. Feature extraction. Fourier and Wavelet transformation, Automatic Feature generation - tsfresh
-Topic 12.Unsupervised approaches: Time series clustering
-Topic 13.Unsupervised approaches: Time series segmentation

Recommender systems. Ranking task
-Topic 14. Recommender systems 1. Explicit feedback
-Topic 15. Recommender systems 2. Implicit feedback
-Topic 16. Ranking task - Learning to rank
-Topic 17. Practical lesson on recommender systems. Surprise!
-Topic 18.Q&A

Graphs
-Topic 19. Introduction to graphs: basic concepts. NetworkX, Stellar
-Topic 20. Graph analysis and interpretation. Community Detection
-Topic 21.Link Prediction and Node Classification
-Topic 22. Practical lesson: Haters on Twitter

Bayesian Learning, PyMC
-Topic 23.Introduction to probabilistic modeling, a posteriori estimates, sampling
-Theme 24.Markov Chain Monte-Carlo (MCMC), Metropolis–Hastings
-Topic 25. Bayesian AB testing
-Topic 26.Generalized linear model (GLM) - Bayesian regressions, derivation of posterior estimates of coefficients
-Topic 27. Practical lesson on GLM
-Topic 28. Bayesian trust network: practical exercise
-Topic 29. Practical lesson on logit regression

Reinforcement Learning
-Topic 30.Introduction to Reinforcement Learning
-Topic 31.Multi-armed bandits for optimization of AB testing, from theory - straight into battle
-Topic 32. Practical lesson: Multi-armed bandits in ecommerce: search optimization
-Topic 33.Markov Decision Process, Value function, Bellman equation
-Topic 34.Value iteration, Policy iteration
-Topic 35. Practical lesson: medical case Markov Chain Monte Carlo
-Topic 36.Temporal Difference (TD) and Q-learning
-Topic 37.SARSA and Practical Lesson: Financial Case TD and Q-learning
-Topic 38.Q&A

Project work
-Topic 39. Consultation on the project, choice of topic
-Topic 40.Bonus: Finding Data Science Jobs
-Topic 41.Protection of design work

Machine learning in practice

An introductory practical course on machine learning. The full cycle of building a solution is considered: from selecting the initial data (“.xlsx file”) through building a model and explaining to the end customer the features of the data and the specifics of the received result. Theoretical sections - classification, regression, predictions, ensembles - are given in overview mode, to the extent necessary for the correct construction and understanding of the examples being analyzed.

41 500 ₽

More details To the course website

24 hours

Tags cloud

Miscellaneous

Rating

Views

Comments