Бакалавриат
2019/2020
Введение в глубинное обучение
Статус:
Курс по выбору (Экономика и статистика)
Направление:
38.03.01. Экономика
Кто читает:
Департамент статистики и анализа данных
Где читается:
Факультет экономических наук
Когда читается:
3-й курс, 3 модуль
Формат изучения:
без онлайн-курса
Язык:
русский
Кредиты:
2
Контактные часы:
2
Программа дисциплины
Аннотация
The course starts with a recap of linear models and discussion of stochastic optimization methods that are crucial for training deep neural networks. Learners will study all popular building blocks of neural networks including fully connected layers, convolutional and recurrent layers. Learners will use these building blocks to define complex modern architectures in TensorFlow and Keras frameworks. In the course project learner will implement deep neural network for the task of image captioning which solves the problem of giving a text description for an input image.
Цель освоения дисциплины
- The goal of this course is to give learners basic understanding of modern neural networks and their applications in computer vision and natural language understanding.
Планируемые результаты обучения
- Student has basic knowledge about linear models and stochatic optimization methods
- Student write his or her first deep network
- Student builds blocks of deep learning for image input.
- Student generates, morphs and searches images with deep learning.
Содержание учебной дисциплины
- Introduction to optimizationCourse intro. Linear regression. Linear classification. Gradient descent. Overfitting problem and model validation. Model regularization. Stochastic gradient descent. Gradient descent extensions.
- Deep Learning for imagesMotivation for convolutional layers. Our first CNN architecture. Training tips and tricks for deep CNNs. Overview of modern CNN architectures. Learning new tasks with pre-trained CNNs. A glimpse of other Computer Vision tasks.
- Unsupervised representation learningUnsupervised learning: what it is and why bother. Autoencoders. Autoencoder applications: image generation, data visualization & more. Natural language processing primer. Word embeddings. Generative models. Generative Adversarial Networks. Applications of adversarial approach.
- Introduction to neural networksMultilayer perceptron (MLP). Chain rule. Backpropagation. Efficient MLP implementation. Other matrix derivatives. What is TensorFlow. Our first model in TensorFlow. What Deep Learning is and is not. Deep learning as a language.
Элементы контроля
- Выполнение заданий по теме 1
- Выполнение заданий по теме 2
- Выполнение заданий по теме 3
- Выполнение заданий по теме 4
- Контрольная работа
Промежуточная аттестация
- Промежуточная аттестация (3 модуль)0.2 * Выполнение заданий по теме 1 + 0.2 * Выполнение заданий по теме 2 + 0.2 * Выполнение заданий по теме 3 + 0.2 * Выполнение заданий по теме 4 + 0.2 * Контрольная работа
Список литературы
Рекомендуемая основная литература
- Kelleher, J. D. (2019). Deep Learning. Cambridge: The MIT Press. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=2234376
Рекомендуемая дополнительная литература
- Taweh Beysolow II. (2017). Introduction to Deep Learning Using R. Springer. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsrep&AN=edsrep.b.spr.sprbok.978.1.4842.2734.3