The study module "Natural language processing for multimodal information processing" is mainly aimed at developing in-depth and highly specialized knowledge level competencies and skills for students studying both humanities, interdisciplinary STEM+-based and information technology study programs.
The objective of the study module is to provide students with the opportunity to learn to cooperate, offering and creating solutions to interdisciplinary, multimodal information processing challenges of different levels of complexity, related to language technology and limited definition, within the cycle of lectures, workshops and practical classes.
The aims of the study module are to give students the opportunity to practice their skills and improve their understanding of the added value that multimodal information processing methods and tools can provide in the creation and distribution of multimodal digital content, digital editing and publishing, interaction of textual information, definition of concepts, spatial and sequential relationships, in evaluating the meaning of emotion-related words in text, researching syntagmatic and paradigmatic meaning dimensions, selecting mood-related characteristics and contextual features, game development and localization, digital advertising and semiotics data management, extracting, analyzing, classifying and evaluating text semantic information, emotions and moods in identification, polarity analysis.

 

AVAILABLE FOR YOU TO STUDY:

 

Digital Semantics and Pragmatics (ETH727)

Study for FREE (you do not need to be a RTU student) on the MOOC platform! 
Lecturers: Tatjana Smirnova, Tatjana Hramova, Zane Seņko, Oksana Ivanova, Alīna Vagele-Kricina, Tatjana Menise, Marina Platonova
Along with a comprehensive overview of the fundamental issues associated with the retrieval, collection, organization, and processing of semantic and pragmatic data (semantic and thematic fields, meaning representation, meaning extension, conceptual mapping, and ontology building, discourse and truth-value analysis), students will get acquainted with the state-of-the-art in the area of natural language processing (NLP) and natural language generation (NLG), which will help them establish a comprehensive theoretical framework for performing a variety of NLP and NLG-related tasks. 
 
Students will study the foundations of compositional and distributional semantics, learn to analyse the topic structure and develop their competence to build semantic models, i.e., semantic networks, taxonomies, ontologies and knowledge graphs, and customize and apply existing ones. 
 

Multimodal Digital Semiotics (ETH728)

Study for FREE (you do not need to be a RTU student) on the MOOC platform!
Lecturers: Marina Platonova, Tatjana Menise, Tatjana Smirnova, Tatjana Hramova, Alīna Vagele-Kricina, Oksana Ivanova, Zane Seņko
The study course promotes students' awareness of various linguistic and non-linguistic semiotic systems and helps them develop a comprehensive understanding of the current trends in their change and development under the influence of digital technologies and media. Upon completion of the study course, students will advance their knowledge of various sign systems, textual interactions, conceptual relations, spatial relations, sequential relations, and syntagmatic and paradigmatic dimensions of signification. Students will develop advanced competence in creating and disseminating multimodal content via digital media, they will also establish a sound competence for the development, customization, and maintenance of digital semiotic resources.
The syllabus of the study course has been balanced to cover a range of theoretical and applied issues to demonstrate the added value of integrating language technology-instigated solutions in the development, customization, and maintenance of multimodal digital semiotic resources.
Through a series of lectures, workshops, and practical classes, students will learn to cooperate in proposing and creating solutions to interdisciplinary semiotic language technology-related challenges of various complexity level with limited definition associated with creation, edition and dissemination of digital content in different formats, and self-expression through digital means. Students will develop awareness of the added value that the methods and tools of Multimodal Digital Semiotics may ensure in multimodal digital content creation, digital editing and publishing, development and localization of games, digital advertising, and semiotic data management. 
 

Digital Sentiment Analysis (ETH729)

Study for FREE (you do not need to be a RTU student) on the MOOC platform!
Lecturers: Marina Platonova, Tatjana Hramova, Tatjana Smirnova, Zane Seņko, Oksana Ivanova, Alīna Vagele-Kricina, Tatjana Menise,
The study course is envisioned for post-graduate students with the basic knowledge of natural language processing (NLP) willing to advance their competence in sentiment analysis and textual data processing for a variety of applied industry-related tasks.
Students will learn to classify unstructured and semi-structured data to determine sentiment polarity (i.e., either positive, negative, or neutral) with the help of free and commercial tools that they will customize for their own research, learning, and occupational needs. Students learn to retrieve and select sentiment-related characteristics and contextual features using relevant models and to assess the impact of the emotion-related words on the overall sentiment of the analysed text. 
Students will develop skills to extract the semantic information from the text, to analyse, classify and evaluate it in terms of sentiment for improving customer experience and quality assurance purposes. Students will master speech tagging, noun phrase extraction, emotion detection, and sentiment analysis, and will address such notions as polarity, intentions and subjectivity by practically working with Python and its dedicated libraries for sentiment analysis, e.g., NLTK and TextBlob. In the practical assignments, students will implement lexicon-based sentiment analysis (e.g., using the VADER (Valence Aware Dictionary and Sentiment Reasoner) lexicon), use pre-trained models for sentiment identification (e.g., the RoBERTa model) and other solutions such as Matplotlib library for the visualization and evaluations of sentiment analysis results.