Overview

This module will give students an advanced level of understanding of the language based artificial intelligences that are transforming our world of work and life. The content is cross-disciplinary and accessible to students with varied prior knowledge.

Classes involve teaching progressively theoretical aspects of applying the latest artificial intelligence approaches to automated understanding, transformation, and generation of multi-modal natural language. Practical components have a strong element of students gaining experience of working with libraries used in industry by leaders in the field such as Explosion.ai, HuggingFace, Meta and OpenAI to develop new models. As the field advances at an ever-accelerating pace, students will also acquire skills to continuously research, evaluate and apply new developments. The module will also make students familiar with multilingual technologies, in particular, in the more challenging context of developing technologies for lesser resourced languages such as Welsh.

Topics will include the fundamentals of natural language processing (NLP) and machine learning methods and architectures before progressing to Machine Translation, Speech to Text, Text-to-Speech and Conversational AI which encompasses using large language models and generative AI.

This module introduces students to a series of applied, theoretical and practical aspects of language technologies, which will help them to gain an advanced level of understanding of the complex skills required to develop automated handling of natural language data. Classes involve practical components with a strong Information Technology element, project-based exercises, group work and discussions. Students will also acquire the skills and knowledge necessary across different areas of language technologies, including the development of certain techniques and resources for background research, terminology management, text analysis and professional skills. The module will also make students familiar with aspects of the developments of language technology in Wales.

Representative topics will include: 1. Introduction to NLP methods 2. Principles of Machine Translation 3. Speech technology: text to speech 4. Speech technology: speech recognition 5. Conversational AI and deep learning 6. Overview of rule-based, statistical and neural network methodologies

Assessment Strategy

-threshold -50%>Student has achieved the minimum acceptable standard of understanding and/or knowledge in all the learning outcomes. Student can demonstrate aminimum level of understanding of the basic concepts and be able to apply them to data with some degree of accuracy. The answer must show evidenceof some background study. -good -60%>Student has achieved a better-than-average standard of understanding and/or knowledge in all learning outcomes, and has a clear and accurate understanding of concepts; ability to apply concepts to data critically and thoughtfully; evidence of wide reading and clear and accurate reference to source materials, including primary sources from current literature; mostly free from misunderstanding and errors of content and from irrelevant material. -excellent -70%>Student has achieved a thorough standard of understanding and/or knowledge in all learning outcomes; or student has demonstrated an exceptional level of achievement in one or more learning outcomes together with a good overall standard: student has achieved a thorough understanding of the subject, both in terms of content and theory; student is able to apply concepts clearly and accurately; substantial evidence of critical and original thought and analysis; clear, logical argument; evidence of an ability to make new links between topics and/or a new approach to a problem; high level of communicative competence; free from misunderstandings, oversimplifications and irrelevant material; evidence of extensive reading and engagement with various primary sources, with clear and accurate references to source material.

Learning Outcomes

• Assess the quality of language technology approaches recommend solutions for diverse language groupings such as lesser resourced languages

• Combine and synthesize information from various sources in order to summarize, evaluate and judge various approaches in the field of Language technologies

• Formulate a question or hypothesis, collect appropriate data and plan and conduct analyses designed to answer that question or test the hypothesis in the field of language technologies

• Integrate a range of linguistic and technicalogical concepts for analysis of language technologies in various monolingual and multilingual settings.