2025/2026
Глубинное обучение
Статус:
Маго-лего
Где читается:
Факультет компьютерных наук
Когда читается:
1, 2 модуль
Онлайн-часы:
48
Охват аудитории:
для всех кампусов НИУ ВШЭ
Преподаватели:
Ратников Федор Дмитриевич
Язык:
английский
Контактные часы:
48
Course Syllabus
Abstract
The proposed course is dedicated to the methods of "deep learning" — neural network methods of machine learning that have sparked a surge of development in several applied fields. Primarily, the course aims to develop students' skills in solving practical problems using deep neural networks.
In recent years, deep learning methods have firmly established themselves in various applied areas, such as:
- Computer Vision: recognizing visual patterns, segmentation, synthesizing images from text descriptions;
- Text Processing: sentiment analysis, question-answer systems, machine translation, large language models;
- Speech Processing: recognition, synthesis, classification of accents, tones, etc.;
- Robotics: optimizing behavior strategies, Reinforcement Learning.
In this course, students are briefly introduced to the applications of neural networks in all these
Learning Objectives
- The proposed course is primarily aimed at developing practical skills in students and providing them with experience in solving applied problems using neural networks. This includes training, fine-tuning, analyzing, and effectively applying neural network models.
Expected Learning Outcomes
- General skills in building and training neural network models.
- Using neural networks for computer vision, working with convolutional and ViT-like neural networks
- Using neural networks for text analysis, including mechanisms of attention.
- Fine-tuning models trained on general data (transfer learning).
- General knowledge in several other areas of neural network application
Course Contents
- week01_backprop
- week02_autodiff
- week03_convnets
- week04_finetuning
- week05_interpretability
- week06_nlp
- week07_attention
- week08_llm
- week09_generative
- week10_speech
- week11_rl
- Темы по выбору
Assessment Elements
- week01_backprop
- week02_autodiff
- week03_convnets
- week04_finetuning
- week05_interpretability
- week06_nlp
- week08_llm
- week09_generative
- week10_speech
- week11_rl
- week07_attention
Interim Assessment
- 2025/2026 2nd moduleFinal = min(10, round(0.1*(ДЗ_1+ДЗ_2+ДЗ_3+ДЗ_4+ДЗ_5+ДЗ_6+ДЗ_7+ ДЗ_8+ДЗ_9+ДЗ_10+ДЗ_11)))
Bibliography
Recommended Core Bibliography
- 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
Recommended Additional Bibliography
- Ian Goodfellow, Yoshua Bengio, & Aaron Courville. (2016). Deep Learning. The MIT Press.