Bachelor
2026/2027




Introduction to Neural Network
ID 1137584
Type:
Elective course (Fundamental and Applied Linguistics)
Delivered by:
Department of Applied Mathematics and Informatics
When:
4 year, 2, 3 module
Online hours:
20
Open to:
students of one campus
Instructors:
Alexander Demidovskij
Language:
English
ECTS credits:
6
Contact hours:
40
Course Syllabus
Abstract
The course covers fundamental and practical aspects of working with modern neural networks and Large Language Models (LLMs). Students will explore the entire lifecycle of language models: from neural networks inference and quality evaluation using standard benchmarks to attention mechanism analysis and parameter-efficient fine-tuning for downstream tasks.
Learning Objectives
- Be able to prepare a dataset for evaluating the quality of a language model
- Be able to perform language model inference
- Be able to evaluate the quality of a language model
- Be able to analyze the model architecture
- Be able to fine-tune the language model
- Know how to optimize a language model for its inference
Expected Learning Outcomes
- Be able to prepare the dataset in the form required for subsequent model quality assessment
- Is able to implement the model inference
- Is able to perform quality assessment of a large language model
- Is able to fine-tune a large language model
- Is able to perform optimization of a large language model
- Is able to analyze Transformer model architecture
Course Contents
- Modern tasks in NLP
- Datasets in the HuggingFace ecosystem
- Inference of large language models
- Evaluation of the quality of language models
- Architecture of Transformer models
- Fine-tuning language models with LoRA
- Optimization of language model inference
Assessment Elements
- Laboratory work №7
- Контроль знаний на семинарах
- Laboratory work №8
- Laboratory work №9
- Exam
Interim Assessment
- 2026/2027 3rd module0.18 * Laboratory work №7 + 0.4 * Exam + 0.12 * Laboratory work №9 + 0.18 * Laboratory work №8 + 0.06 * Контроль знаний на семинарах