Event

Summarize today's lecture? When were neural networks invented? An artificial intelligence that can answer these questions is a long dream of mankind. With the help of machine learning and in particular large language models m(LLMs), we are seeing the first programs today that can solve these problems. This lecture provides the skills and knowledge to develop solutions to these natural language processing problems using state-of-the-art methods.
After an introduction to the challenges of natural language processing, the different tasks in natural language processing are discussed. A focus of the course is on methods from the field of deep learning and the new possibilities through large language models (ChatGPT). First, sequence classification tasks such as sentiment analysis are dealt with. Then methods of sequence labeling are discussed, such as those used to recognize proper names or determine part-of-speech tags. Afterwards the lecture will discuss sequence-to-sequence methods. These models are used in many natural language processing tasks, such as machine translation, automatic summarization, and automatic question answering.
In this course, the important challenges in the development of systems are covered: the representation of words, neural architectures to model language, methods to train complex models and finding the most likely output.