- Новости
- Краткая информация
- Время и место
- Связь с преподавателями
- Результаты выполнения заданий
- Программа курса
- Список литературы
- Полезные ссылки
- Подведены итоги курса - см. детальную разбалловку
- Выложена десятая лекция и соответствующий ноутбук
- Запустили соревнование по машинному обучению https://www.kaggle.com/t/af6a791ec09847f390bddbb03a86cea3
- Выложены лекции с седьмой по девятую
- Выложено второе практическое задание. Дедлайн для получения полного балла - 17 мая 2019
- Второе теоретическое задание. Дедлайн для получения полного балла - 10 мая 2019
- Выложены лекции с четвертой по шестую и ноутбуки с пятой и шестой лекций
- Выложено первое практическое задание. Дедлайн для получения полного балла - 08 апреля 2019
- Третья лекция и ноутбуки с третьего занятия
- Вторая лекция, первое теоретическое задание (дедлайн для получения полного балла - 01 апреля 2019) и ноутбуки со второго занятия
- Первая лекция
- Первая лекция состоится в пятницу, 15 февраля, в 18:30 в аудитории 1225 (ГЗ МГУ).
В весеннем семестре 2019 года на механико-математическом факультете МГУ им. М. В. Ломоносова начинается чтение нового спецкурса по выбору студента, посвященного классическим алгоритмам машинного обучения.
Курс будет читаться на базе кафедры Математической Теории Интеллектуальных Систем под руководством д.ф.-м.н., профессора Бабина Д. Н. Курс будут читать к.ф.-м.н. Петюшко А. А. и к.ф.-м.н. Иванов И. Е.
Курс читается по пятницам, 18:30, ГЗ МГУ, аудитория 1225.
- Telegram-канал, в котором будут появляться все важные новости
- Обратная связь - по почте mlcoursemm@gmail.com
- Ну и всегда можно написать в issues :)
- Соревнование по машинному обучению
- Разбалловка, сроки проведения и правила участия можно найти тут
- Второе практическое задание
- Сроки сдачи: дедлайн 17 мая, если будет сдача после него - баллы умножаются на 0.5
- Разбалловка указана в самом ноутбуке
- Решения присылать на почту курса mlcoursemm@gmail.com
- Формат: ноутбук с Вашим кодом
- Второе теоретическое задание
- Сроки сдачи: дедлайн 10 мая, если будет сдача после него - баллы умножаются на 0.5
- Все остальные условия - как и в первом теоретическом задании
- Первое практическое задание
- Сроки сдачи: дедлайн 8 апреля, если будет сдача после него - баллы умножаются на 0.5
- Разбалловка указана в самом ноутбуке
- Решения присылать на почту курса mlcoursemm@gmail.com
- Формат: ноутбук с Вашим кодом
- Первое теоретическое задание
- Сроки сдачи: дедлайн 1 апреля, если будет сдача после него - баллы умножаются на 0.5
- Разбалловка: 1 балл за задачу, если в задаче подпункты - то по баллу за каждый подпункт
- Решения присылать на почту курса mlcoursemm@gmail.com
- Формат: лучше пдф-кой (сделать в ТеХе), но можно и просто сфотографировать решение на камеру
- Итоговая табличка: Ссылка на гуглдокс
Тема 1. Постановка задач машинного обучения. Понятие искусственного интеллекта.
Тема 2. Валидация алгоритмов, различные типы валидации.
Тема 3. Переобучение. Недообучение. Декомпозиция ошибки на смещение и разброс.
Тема 4. Методы классификации. Байесовский подход.
Тема 5. Введение в язык программирования Python. Обзор библиотеки scikit-learn.
Тема 6. Методы классификации. Метрические методы.
Тема 7. Стохастический градиентный спуск.
Тема 8. Методы классификации. Линейные классификаторы.
Тема 9. Введение в искусственные нейронные сети. Теорема Новикова.
Тема 10. Машины опорных векторов.
Тема 11. Решающие деревья и логические методы классификации.
Тема 12. Методы восстановления регрессии. Линейные модели.
Тема 13. Методы восстановления регрессии. Нелинейные модели.
Тема 14. Композиции алгоритмов машинного обучения. Композиции классификаторов.
Тема 15. Тестирование алгоритмов, метрики качества.
Тема 16. Задачи компьютерного зрения.
- Курс лекций по машинному обучению на http://www.machinelearning.ru от Воронцова К. В.
- Hastie, T. and Tibshirani, R. and Friedman, J. The Elements of Statistical Learning, 2nd edition, Springer, 2009.
- Bishop, C.M. Pattern Recognition and Machine Learning, Springer, 2006.
- Краткая справочная информация по Python, NumPy, SciPy, SciKit-learn, Pandas, MatPlotLib, Jupyter Notebook: см. в соответствующей папке курса
- Python Programming in 15 min: Part1, Part2, Part3
- Python Programmin - A Modern Approach: Code, notebooks and slides
- Playground and Cheatsheet for Learning Python: github repo
- Homemade Machine Learning: github repo
- Machine learning: Курс Andrew Ng на площадке https://www.coursera.org
- News
- Short info
- Time and place
- Communication with teachers
- Task results
- Course program
- Bibliography
- Useful links
- Summed up the results - see the detailed table
- Uploaded the 10th lecture and corresponding notebook
- Launched a machine learning competition on kaggle
- Lectures are uploaded: 7th through 9th
- Uploaded the second practical task. Deadline for getting the Full Score - May 17, 2019
- Uploaded the second theoretical task. Deadline for getting the Full Score - May 10, 2019
- Lectures are uploaded: from 4th to 6th and notebooks with the 5th and 6th lectures
- Uploaded the first practical task. Deadline for getting the Full Score - April 08, 2019
- Third lecture and notebooks from the third lesson are uploaded
- Second lecture, first theoretical task (deadline for getting the full score - April 01, 2019) and notebooks from the second lesson are uploaded
- First lecture
- The first lecture will take place on Friday 15, February, at 6:30 pm in room 1225 (main bilding MSU).
In the spring semester of 2019 at the Faculty of Mechanics and Mathematics of Lomonosov Moscow State University begins reading a new special course of the student's choice, dedicated to classical machine learning algorithms.
The course will be taught on the basis of the department Mathematical Theory of Intelligent Systems under the guidance of Doctor of Physical and Mathematical Sciences, Professor Babin D.N. The course will be delivered by Ph.D. Petiushko A.A. and Ph.D. Ivanov I.E.
The lessons are to be taught on Fridays, 18:30, main bilding MSU, room 1225.
- Telegram-channel, where all important news will appear
- Feedback - by email mlcoursemm@gmail.com
- Well, you can always write in issues :)
- Machine learning competition
- The resulting table, timing and rules of participation can be found here
- Second practical task
- Timing: deadline on May 17, if there is a delivery after it - points are multiplied by 0.5
- Each task scoring is indicated in the notebook itself
- Send solutions to the course email mlcoursemm@gmail.com
- Format: notebook with your code
- Second theoretical task
- Timing: deadline on May 10, if there is a delivery after it - points are multiplied by 0.5
- All other conditions are the same as in the first theoretical task
- First practical task
- Timing: deadline on April 8, if there is a delivery after it - points are multiplied by 0.5
- Each task scoring is indicated in the notebook itself
- Send solutions to the course email mlcoursemm@gmail.com
- Format: notebook with your code
- The first theoretical task
- Timing: deadline on April 1, if there is delivery after it - points are multiplied by 0.5
- Scoring: 1 point for the task, if there are sub-points in the task, then 1 point for each sub-task
- Send solutions to the course email mlcoursemm@gmail.com
- Format: it is better pdf (done in TeX), but you can just take a picture of the solution with a camera
- Summary table: Link to googledocs
Topic 1. Machine learning problems statement. Artificial intelligence concept.
Topic 2. Validation of algorithms, various types of validation.
Topic 3. Overfitting. Underfitting. Decomposition of the error into bias and variance.
Topic 4. Classification methods. Bayesian approach.
Topic 5. An introduction to the Python programming language. An overview of the scikit-learn library.
Topic 6. Classification methods. Metric methods.
Topic 7. Stochastic gradient descent.
Topic 8. Classification methods. Linear classifiers.
Topic 9. An introduction to artificial neural networks. Novikov's theorem.
Topic 10. Support vector machines.
Topic 11. Decision trees and logical classification methods.
Topic 12. Regression. Linear models.
Topic 13. Regression. Non-linear models.
Topic 14. Compositions of machine learning algorithms. Compositions of classifiers.
Topic 15. Algorithm testing, quality metrics.
Topic 16. Computer vision tasks.
- Machine Learning Lecture Course on http://www.machinelearning.ru from Vorontsov K.V.
- Hastie, T. and Tibshirani, R. and Friedman, J. The Elements of Statistical Learning, 2nd edition, Springer, 2009.
- Bishop, C.M. Pattern Recognition and Machine Learning, Springer, 2006.
- Quick reference information on Python, NumPy, SciPy, SciKit-learn, Pandas, MatPlotLib, Jupyter Notebook: see in corresponding folder course
- Python Programming in 15 min: Part1, Part2, Part3
- Python Programming - A Modern Approach: Code, notebooks and slides
- Playground and Cheatsheet for Learning Python: github repo
- Homemade Machine Learning: github repo
- Machine learning: Course by Andrew Ng on the site https://www.coursera.org