This workshop will be divided into three main sections. The first one will focus on practical issues involved in the processing of printed and digital newspapers in SS&H research. Participants will learn about the most common difficulties related to the digitization and analysis of periodicals, such as discontinuous articles, multi-column layouts, mixing of different texts on a single page, among others. The session will demonstrate why automatic processing of such materials is challenging and in which situations manual work is still inevitable.

Subsequently, examples of tools and methods supporting data extraction from newspapers will be presented, including solutions that enable the extraction of specific fragments from larger texts. An example of using a tool, such as PressMint and/or ChronoPress for structuring and selecting data from large press corpora will also be shown.

The second part of the workshop will introduce Słowosieć (plWordNet), the largest relational dictionary in the world. The session will present the characteristics of relational description and its potential applications in linguistic, historical, and sociological research. Particular attention will be paid to semantic fields related to place names, archaic vocabulary, and marked or evaluative language. The workshop will also cover the Senselink tool. Next, basic tools for browsing the resource and generating lists of words and semantic relations between them will be demonstrated. The workshop will additionally refer to Open Multilingual Wordnet as a multilingual framework for linking lexical resources across languages.

Finally, we will present the use of the PLLuM family of models for academic research. PLLuM has been trained to a significant extent on Polish-language data and performs well in the Polish cultural context. The session will introduce the possibilities of fine-tuning models for specialized tasks. In particular, it will discuss the use of Retrieval-Augmented Generation (RAG) approaches for analyzing large collections of documents and files.