The Intelligent and Interactive Systems group (Prof. Justus Piater and Dr. Matteo Saveriano) is looking for two talented Ph.D. candidates.
We would like to be able to equip our robots with almost arbitrary skills in unconstrained environments. However, programming robot skills by hand is time consuming, and supervised and exploratory learning do not easily scale to complex and diverse tasks. Open-ended learning requires a combination of exploratory learning, explicit teaching, and the ability to reuse knowledge.
The goal of this project is to enable robotic devices to acquire knowledge about the task execution from human observation and to build a large scale database of robotic tasks. Modern computing technologies, like noSQL, graph databases and cloud computing, are leveraged to populate and query the knowledge base and plan tasks on-line. A large knowledge base will allow the robot to execute a multitude of tasks and, more importantly, can potentially allow the robot to reuse its knowledge in different scenarios in order to learn novel behaviors by self-practice. The Ph.D. candidate will also try to overcome the current imitation learning paradigm where the human is the only source of knowledge by adopting a bidirectional paradigm where the robot uses already acquired knowledge to assist the human during the teaching.
This is a university assistant position that includes minor teaching requirements.
Most current work in robot learning focuses on learning trajectories or on mapping perception to action. However, tasks are characterized by the achievement of specific post-conditions; it is these the robot should understand, along with means to achieve and verify them. The objective of this research is to leverage explicit human teaching and other means to focus the attention of the robot to the crucial aspects of its perception/action space that define the task, allowing it to acquire complex skills in open-ended scenarios.
This position is part of the new EGTC Euregio IPN OLIVER - Open-Ended Learning for Interactive Robots, with Prof. Angelika Peer (U. Bolzano) and Prof. Nicu Sebe (U. Trento).
Applicants must have
Applications must state the position(s) of interest (1, 2 or both), and must include
They should be sent as PDF attachments by e-mail to both Justus.Piater@uibk.ac.at and Matteo.Saveriano@uibk.ac.at.
Applications must be received by July 15, 2019. The ideal starting date is September 1st, 2019.
The University of Innsbruck dates back to 1669 and celebrates its 350th anniversary this year. It offers a complete set of academic curricula and currently counts 28000 students. Founded in 2001, our young Department of Computer Science is highly productive in diverse research domains, and is internationally very well connected.
Innsbruck is home to 35000 students who imprint a distinctive, international student atmosphere upon this lively city of 130000. Beautifully located in the Tyrolean Alps, on the Inn river and surrounded by summits of up to 2718m, Innsbruck offers outstanding opportunities and quality of life all around the year.