Projects for PhD studies at KTH

Project KTH1: Security and Privacy of Autonomous Systems in the Home

Project KTH2:  Reinforcement learning endowed multi-robot planning and control under temporal logic tasks

Project KTH3: Learning Dynamical Systems

Project KTH4: Embedded Optimization for Real-Time Machine-Learning

Project KTH5:  Principled Integration of Logic Reasoning and Deep Learning

Project KTH6:  Autonomous Time-Critical Cloud

Project KTH7: Trustworthy Internet of Things

 

Link To Application Portal: 7 WASP PhD student positions

Application deadline: June 15 2017

 

Project KTH1: Security and Privacy of Autonomous Systems in the Home

Supervisor:  Sonja Buchegger, School of Computer Science and Communication, KTH

Contact info:  Sonja Buchegger, buc@kth.se,  http://www.csc.kth.se/~buc/

Short Project description: Smart-home technology and the Internet of Things increasingly incorporate autonomous systems that can make decisions on our behalf – both to increase the utility and to decrease the complexity for the user. This becomes especially important in home environments for independent life which allow people with reduced activity ranges due to age, illness, or disabilities to live in their own homes longer. Yet such autonomous systems raise concerns about security and privacy. Quantitatively, this means more (sensor) data as well as more interconnected devices leading to increased system complexity. Qualitatively, the collected and inferred data is increasingly personal. This project will analyze these concerns, derive requirements, and, based on cryptographic approaches, develop solutions toward balancing functionality, security, privacy, and performance such that the networked systems can be useful and trustworthy. The latter is a prerequisite for the adoption of such autonomous systems; they must be protected from adversaries that, once having forced access, could change the functionality or leak private data.

 

Project KTH2:  Reinforcement learning endowed multi-robot planning and control under temporal logic tasks

Supervisor: Dimos Dimarogonas, School of Electrical Engineering, KTH

Contact info: Dimos Dimarogonas, dimos@kth.se,  https://www.kth.se/profile/dimos, http://people.kth.se/~dimos/

Short Project description: We will consider the problem of distributed task and motion planning for multi-robot systems in unknown and dynamic environments. Robotic tasks in this project will be considered to be of a non-standard nature from a classic control theory viewpoint, since they will be defined in the form of temporal logic/language based formulas from formal verification. Such logics, such as Linear Temporal Logic and Signal Temporal Logic (abbr. LTL/STL) impose specifications to robots that can be seen as a combination of Boolean, time/temporal and state-space/spatial constraints such as “in case of detecting an intruder, return to the base station within 5 minutes and switch on alarm”.  Based on the limited sensing and communication capabilities of the robots, we plan to develop distributed reinforcement learning tools for appropriate control policy adaptation that will (i) exploit inter-robot communication to collaboratively learn the common workspace model, and furthermore (ii) iteratively adapt and improve the parameterized local policy to optimally accommodate the updated model and the uncertainties due to other robots’ behaviors. The research will blend elements from machine learning, distributed control and formal verification towards a novel approach to flexible multi-robot task planning and control in unknown and dynamic environments.

 

Project KTH3: Learning Dynamical Systems

Supervisor: Håkan Hjalmarsson,  School of Electrical Engineering, KTH

Contact info: Håkan Hjalmarsson, hjalmars@kth.se, https://www.kth.se/profile/hjalmars

Short Project description:  Learning dynamical systems is an area closely related to machine learning, cyber-physical systems as well as real-time big data analytics, and it provides backbone algorithms for digitalization of industry and society. Among others, it is core technology in autonomous systems with applications such as smart buildings, self-driving vehicles, and self-learning robots. In this project we focus on three key themes: Fundamental techniques concerns learning parsimonious models in a statistical and computationally efficient way. Active and on-line learning concerns how to improve data-efficiency by actively controlling the excitation of the system in a sequential manner. Dynamical networked systems addresses issues of relevance to learning of interconnected dynamical systems, a field rapidly increasing in importance thanks to the fast development of 5g communication technology and the Internet-Of-Things paradigm.

 

Project KTH4: Embedded Optimization for Real-Time Machine-Learning

Supervisor: Mikael Johansson, School of Electrical Engineering, KTH

Contact info: Mikael Johansson, mikaelj@kth.se, https://www.kth.se/profile/mikaelj

Short Project description: An increasing number of our daily decisions are guided by computers that sense, infer and act on the data that they observe. Many of these autonomous decision-making tasks are naturally posed as optimization problems. Examples include finding the best parameters (“training”) of a deep neural network, deriving an optimal decision rule for trading of volatile assets, or determining the best action of an autonomous vehicle under dynamic constraints and uncertain observations. Real-time optimization is becoming a critical technology for making better, faster and more informed autonomous decisions. In this project, we will develop advanced optimization algorithms that are able to run in real-time on embedded hardware. We aim at novel classes of optimization algorithms that can deliver optimal or near-optimal solutions with limited memory, processing and energy resources. Particular attention will be given to algorithms that can exploit emerging multi-core/multi-GPU embedded platforms. The project will blend applied mathematics for algorithm development, implementation of these algorithms on emerging hardware architectures, and application of real-time optimization to autonomous decision-making.

 

Project KTH5:  Principled Integration of Logic Reasoning and Deep Learning

Supervisor: Hedvig Kjellström,  School of Computer Science and Communication, KTH

Contact info: Hedvig Kjellström, hedvig@csc.kth.se, http://www.csc.kth.se/~hedvig/

Short Project description: Machine Learning methods based on Deep Neural Networks (DNN) have been tremendously successful in the last few years. The success is especially prominent in domains where large volumes of training data can be acquired, such as Computer Vision and Speech Recognition. However, in domains where the goal is to infer complex causal chains, the needed amount of training data grows rapidly. An example is human-robot collaboration, where the robot might want to infer the goal of a sequence of actions performed by a human, in order to plan its own actions.

On the other hand, humans are able to learn complex models from very few examples. We argue that the key difference between human learning and DNN learning is that humans employ logic reasoning, and knowledge about intuitive physics and intuitive psychology in their learning. Such models, which will combine DNN with logic reasoning and probabilistic models in a principled manner, would enable learning of much more complex models from much less training data – a major breakthrough in Deep Learning!

 

Project KTH6:  Autonomous Time-Critical Cloud

Supervisor:  Dejan Kostic,  School of Information and Communication Technology, KTH

Contact info: Dejan Kostic, dmk@kth.se, https://people.kth.se/~dejanko/

Short Project description:  In a number of time-critical societal applications, besides providing high reliability and throughput a very important property the cloud infrastructure has to provide is guaranteed low latency for delivering data. This feature is sorely lacking today, with the so-called tail-latency of slowest responses in popular cloud services being several orders of magnitude longer than the median response times. Unfortunately, simply using a network infrastructure with ample bandwidth does not guarantee low latency because of problems with congestion at the intra-and inter-data center level, sub-optimal routing, server overload, etc. The goal of this PhD position is to develop the necessary technology for time-critical cloud services. Examples include advanced software-defined control of the network, highly streamlined network functions virtualization, and geo-distributed data storage systems. The student will work as a part of the team developing autonomic solutions for time-critical data transmission, which will be evaluated on the test case in intelligent transport systems.

 

Project KTH7: Trustworthy Internet of Things

Supervisor: Panagiotis Papadimitratos, School of Electrical Engineering, KTH

Contact info: Panagiotis Papadimitratos, papadim@kth.se, https://www.kth.se/profile/papadim

Short Project description: A wireless revolution has been materializing, transforming our environments and processes into ‘intelligent’ ones through an Internet of Things (IoT) and increasing levels of autonomous operation. The benefits include, for example, ubiquitous medical services, smart energy production and distribution, automated logistics chains, adaptable buildings, more effective tactical operations, more efficient and safer transportation. We have every reason to embrace the IoT and make our lives and businesses easier. But to do so, we need to ensure our systems are secure and privacy preserving. Otherwise, critical processes could come to a halt, individuals could be hurt and their privacy be comprised. This is exactly the motivation for this line of work. We are looking for candidates that can contribute high-quality research towards the design, implementation, evaluation, and analysis of secure and privacy-preserving IoT systems. Skills and background on systems, or formal methods, or information theory are required, along with a solid understanding of technology; combined profiles are a plus. The objective of this work is to contribute to the next wave of security and privacy solutions for the IoT, strengthen the theoretical foundations, instantiate security and privacy for key applications, and catalyze the adoption of resilient, trustworthy IoT systems.