Course contents: The course focuses on methods of computational Natural Language Processing (NLP), emphasizing the use of machine learning and deep learning techniques. Students will experience the full pipeline of text modeling: from word representation using embeddings to sentence understanding and language generation via transformer models and large language models (LLMs). The course combines theoretical knowledge with hands-on applications and includes critical analysis of ethical issues in NLP. 

At the end of the course the student will be able to:

• Represent words and phrases using count-based, distributed, and contextual models (e.g., TF-IDF, Word2Vec, BERT).
• Analyze and implement basic machine learning techniques for text processing (Naive Bayes, SVMs, neural networks).
• Apply and fine-tune modern Transformer models (BERT, GPT) across various NLP tasks (classification, QA, summarization).
• Develop dialogue systems and text generation applications using large language models (LLMs).
• Design multimodal applications combining language and vision (e.g., image captioning).
• Identify and assess ethical issues related to natural language processing and LLMs (bias, hallucinations).

Assessment: Course evaluation is based on short assignments and/or a midterm progress assessment during the semester, with a total weight of up to 40%. The final semester project (code and technical report) accounts for 60%. These percentages may vary (by up to ±10%) from year to year. To pass the course, students must achieve a passing grade both in the final project and in the overall grade. The assignments may be accompanied by an oral examination.