Text analytics and machine learning and large language models

The course aims to provide students with both theoretical and practical knowledge of contemporary text analytics, machine learning techniques, neural network architectures, and Large Language Models (LLMs), with particular attention to their applications in Computational Social Science. Students will be introduced to the evolution of computational approaches to text analysis, from traditional bag-of-words representations and topic modelling techniques to advanced embedding methods and Transformer-based architectures that underpin contemporary LLMs.
The course will cover the implementation and use of machine learning and neural architectures on a variety of textual datasets, enabling students to understand how different models behave across different research tasks and analytical settings. Particular emphasis will be placed on practical experimentation through programming activities in both R and Python, using tools such as TensorFlow, Keras, and the Hugging Face ecosystem.
Students will learn how to preprocess textual data, implement supervised and unsupervised machine learning algorithms, validate and evaluate model performance, and critically assess model robustness. The course will also introduce methods for both local and global model interpretation, enabling students to understand, explain, and evaluate algorithmic decision-making processes in computational social science applications.
In addition, students will explore semantic representation techniques through static and dynamic word embeddings, with a special focus on encoder-based architectures and their role in modern natural language processing pipelines. They will also learn how to access, fine-tune, evaluate, save, and deploy pretrained models using the Hugging Face ecosystem, including adaptation to downstream tasks and domain-specific datasets.
The course will further address the growing importance of interpretability, explainability, and transparency in AI systems. Students will explore methods for validating and interpreting model outputs in order to critically assess reliability, fairness, and potential biases in machine learning systems applied to social science research.

By the end of the course, students will be able to:
- Understand the theoretical foundations of Text Analytics
- Understand the theoretical foundations of Machine Learning algorithms
- Apply explainability and interpretability techniques to assess and validate machine learning outputs and validate the results for computational social science applications
- Understand the theoretical foundations and differences between Recurrent Neural Networks, Convolutional Neural Networks, and Transformer architectures
- Explain the principles underlying Large Language Models and their relevance to Computational Social Science research
- Develop practical skills for integrating machine learning and LLMs into computational social science workflows and research projects
- Critically evaluate the limitations, biases, and ethical implications of LLM-based systems.