Wallenberg Autonomous Systems and Software Program
Assoc. Prof. Helena Lindgren, Dept of Computing Science, Umeå University
Phone +46 70 3657463, firstname.lastname@example.org
The aim of the research project is to develop socially intelligent software agents for human-agent collaboration. The fundamental challenge lies in how intelligent autonomous agents may collaborate with humans in decision making tasks to achieve goals, and in making prioritizations among potentially conflicting goals, needs, motivations, preferences and choices of actions, e.g. in medical situations where healthcare professionals diagnose or select treatment methods. This is also highly important in situations where a person aims to change unhealthy behaviour, or needs to take action in order to reduce risk in work situations.
To provide socially intelligent systems that humans can trust enough to collaborate with, algorithms for explaining automated learning, reasoning, and values of arguments and decision outcomes will be developed. Artificial intelligence-based methods for user modelling, user adaptation, and for the system to act in a socially acceptable way tailored to a situation will also be developed, partly by formalising theories about human behaviour. Methods that handle uncertain and incomplete information, as well as different types of values, norms, or utilities in such situations will be explored.
The successful candidate will be part of the multidisciplinary research group on User, Interaction and Knowledge Modelling (UIKM) at the department of Computing Science, which conducts foundational research in formal methods for reasoning and decision making, methods for user modelling, user adaptation and human-agent collaboration. Theories and methods in the fields artificial intelligence, human-computer interaction and cognitive science are developed. The group collaborates in its applied research with a number of research groups in the medical and health domains, and is providing a vibrant multi-disciplinary research environment. Please, visit the group's web page for more information.
Candidates are expected to have a background in computer science; a specialization in artificial intelligence or human-computer interaction is a merit. Candidates from related disciplines, such as cognitive science, with a very good understanding of fundamental concepts of computer science and good programming experience may also be considered. Experience from healthcare, or research projects relating to healthcare is also a merit. Since research is conducted in an international research environment, the ability to collaborate and contribute to teamwork, and a very good command of the English language, both written and spoken, are key requirements.
Please apply here.
Research group: http://www.cs.umu.se/forskning/forskargrupper/uikm/
Assoc. Prof. Kai-Florian Richter, Department of Computing Science, Umeå University
+46 90 786 68 31 , email@example.com
Autonomous systems, such as self-driving vehicles or robots in household or healthcare settings, are on the verge of becoming integral parts of our society and everyday lives. In recent years, there has been tremendous progress in making these systems truly autonomous, but considerably less thought seems to have been spent on how to interact with them. But in the end, this interaction will be crucial for their success. For example, we might rather like to tell a self-driving car to take us to “Grandma’s” or to “the gym” than to “22 Smith St” or to “57°33'N 24°15’W”. And if dozing off during the drive and waking up again, “on the highway” may be a valid but not particularly helpful answer by the car to our question of “where are we?” Likewise, when sending off an autonomous robot to do something for us, on its return simply stating “here” or “yes” may not be the kind of feedback we would expect.
In other words, while autonomous systems are autonomous, they are not self-sufficient. Sooner or later they need to interact with humans, which means they need to communicate (the results of) their processing in a way these humans can understand. This may be particularly challenging for those systems operating in environments where a user's knowledge about and experience with them is varying and possibly unknown, which would hold for the systems mentioned above, for example.
This PhD project will explore such interaction between human and autonomous system. It will focus on questions of how to establish and communicate meaningful spatial and temporal references, given that the last interaction may have been minutes or even hours ago. The aim is to develop methods for both understanding and producing such references. Among others, research in this project is expected to involve questions of relevance detection, reasoning, and keeping track of user interactions.
You will have a background in computer science, artificial intelligence, human-computer interaction, or computational linguistics. Candidates from related disciplines, such as cognitive science, with a good understanding of fundamental concepts of computer science and good programming experience will also be considered. You will be interested in developing prototype implementations as well as empirical evaluations in human-subject studies. You will also have very good knowledge of English, both spoken and written.
This PhD project will be part of the Spatial Cognitive Engineering group at the Department of Computing Science. The group is led by Assoc. Prof. Kai-Florian Richter and has strong international links. Research in the group addresses cognitive issues in communication processes between human and system, and in people’s use of assistance systems. To this end the group employs interdisciplinary approaches and methods from artificial intelligence, human-computer interaction, cognitive science, and geographic information science.
Please apply here.
Assoc. Prof. Martin Rosvall, Integrated Science Lab, Department of Physics, Umeå University
Many autonomous system rely on state-of-the-art machine learning algorithms. The best-performing algorithms are also the most opaque. This transparency problem¾the fact that we do not understand how they work¾causes a great challenge for verification. That is, for autonomous systems that are supposed to work in dynamic conditions, such as robots, self-driving cars, or smart analytics tools, we only know that they work in tested conditions. And as the autonomous systems acquire new knowledge, how do we know that the new machines still handle previously tested conditions? This challenge calls for more transparent machine learning algorithms that allow us to look under the hood and understand the machinery. This project seeks to integrate transparent but domain-specific inference techniques from network science with versatile but opaque machine learning algorithms and enable self-driving transparent analytics.
We are now seeking a PhD student to join our interdisciplinary group and develop innovative and transparent machine-learning algorithms for autonomous systems. This will allow us to take advantage of today's data explosion for revealing and taking advantage of dynamic patterns in complex systems with important applications in industry.
The Integrated Science Lab offers a thriving, creative environment for interdisciplinary research and plentiful opportunity to explore and realize other projects as well.
The successful candidate should hold a master degree (or equivalent) in mathematics, computer science, physics, or relevant field, and have excellent programming skills using modern languages. Experience in computational or mathematical modelling is an advantage. The candidate should have strong interest in interdisciplinary research, and must be highly motivated and have the ability to work independently as well as a part of the research group. The candidate must be fluent in both oral and written English.
Please apply here.