CRITICAL DESIGN AND PRACTICAL ETHICS FOR AI

This course will teach you how to build convolutional neural networks. You will learn to design intelligent systems using deep learning for classification, annotation, and object recognition.

This course deals with model-based testing, a class of technologies shown to be effective and efficient in assessing the quality and correctness of large software systems. Throughout the course the participants will learn how to design and use model-based testing tools, how to create realistic models and how to use these models to automate the testing process in their organisation.

The purpose is to give the students an overview of issues and methods for development and assurance of safety-critical software, including details of selected technologies, methods and tools. The course includes four modules: Introduction to functional safety; knowledge that give increased understanding of the relationship between Embedded systems / safety-critical system / accidents / complexity / development models (development lifecycle models) / certification / “the safety case”. Analysis and modelling methods; review of analysis and modelling techniques for the development of safety-critical systems. Verification and validation of safety critical software, methods and activities to perform verification and validation. Architectures for safety critical systems. Safety as a design constraint.  

Would you like to know what Industry 4.0 is about? Then this course is for you! In the course, we look at enabling technologies of Industry 4.0 from a human and industrial perspective. The course covers many topics and you will learn the basic terminology related to Industry 4.0 as well as insight and understanding of the Fourth Industrial Revolution and how it is set to affect industry and individuals.

Answer Set Programming (ASP) is a declarative programming paradigm designed within the field of Artificial Intelligence (AI), and used to solve complex search-problems. The declarative nature of ASP allows one to encode a problem by means of logic. In this way, unlike in imperative programming approaches, there is no need to design an algorithm as a solution for the given problem. In this sense, ASP is comparable with SAT-based encoding or constraint satisfaction problems. However, due to its stable-model semantics, ASP provides a richer representation language useful to handle uncertain situations more effectively for real world scenarios. The advantages of declarative programming together with non-monotonic nature of ASP in handling uncertainties have recently made ASP more attractive both for academia and industry. This course focuses on formalizing and solving various search problems in planning, scheduling and system configuration in ASP.

The aim of this course is that students will learn about the analysis, design, and programming of deep learning algorithms. The course is part of the programme MAISTR (hh.se/maistr) where participants can take the entire programme or individual courses. The course is for professionals and is held online in English. Application is open as long as there is a possibility of admission. The courses qualify for credits and are free of charge for participants who are citizens of any EU or EEA country, or Switzerland, or are permanent residents in Sweden. More information can be found at antagning.se. About the course Applied Deep Learning with PyTorch, 5 credits Who is this course for?This course provides the theoretical and practical aspects of deep neural networks. It is intended for students with a background in computer science and engineering. What will you learn from this course?Students will learn about the analysis, design, and programming of deep learning algorithms. The course has two modules: theory and practice. The theoretical content covers basic principles of multi-layer perceptions, spatio-temporal feature extraction with convolutional neural networks (CNNs), and recurrent neural networks (RNNs), classification and regression of big data, and generating novel data samples using generative models. The practical sessions cover the basics of programming with PyTorch. For instance, image classification and semantic segmentation using CNNs, future image frame prediction with RNNs, and image generation with generative adversarial networks. What is the format for this course?Instruction type: Teaching is in English and fully online. It consists of lectures, computer exercises, and project work. In the computer exercises, the student solves small problems using deep learning models. After programming various exercises, the participants will develop an advanced deep learning project. Participants will be encouraged to bring their own data. High-end GPU machines can be provided for the exercises and project.