Just like humans, artificial intelligence has to perceive – i.e. to sense/measure/observe the real world. It also needs to communicate with humans both at the input and at the output.
Without a connection to the real world, AI is of no use. It needs to see the world around as well as to understand our language and talk to people properly, carrying a meaningful and purposeful dialogue.
The Natural Language Processing course offers a broad and yet detailed and technical overview of the several important areas in NLP such as language modelling, word and sentence embeddings, information retrieval or machine translation. To handle the softness of the observations and uncertainty in the data, NLP has been using statistical methods and (deep) neural networks ever since. To allow for expressing what we, users of language, feel as rules, symbolic methods have always been at hand, including exact handling of ambiguities. For many NLP tasks, data are plentiful and easy to process (plaintext!) and yet immediately demonstrate the complex fabric of the language and our communication. Be sure to get your hands dirty with the data, in the NLP course.
This Dialogue Systems course introduces the architecture of spoken dialogue systems, voice assistants and conversational systems (chatbots). It discusses the main components of dialogue systems (speech recognition, language understanding, dialogue management, language generation and speech synthesis) and shows alternative approaches to their implementation.
The Computer Vision Methods course addresses some key challenges in visual perception including how to find images with similar semantic and not always the same visual content (image retrieval), how to pair images (image matching) and how to follow moving objects (visual tracking). The course also discusses some machine learning approaches used in the present computer vision world.