Academy projects from ICT 2023 calls

Six projects at the Department of Computer Science received funding from the Academy of Finland ICT 2023 calls. The principal investigators of the projects are Professor Jussi Kangasharju, Assistant Professor Arto Klami, Associate Professor Teemu Roos, Dr. Tuukka Ruotsalo, Professor Sasu Tarkoma, and Professor Hannu Toivonen. Prof. Toivonen's project is collaboration with Professor Männistö. In addition, Professor Ville Mustonen obtained a project affiliated with Viikki Campus. Altogether, the Department was highly successful in these calls targeting research on computation, machine learning and AI, and industrial internet, respectively.

Prof. Kangasharju’s project Where’s My Data? Managing Data in Edge Computing Environments investigates distributed data management in emerging edge computing environments, in particular targeting Industrial Internet scenarios via two case examples, collaborative robotics and distributed smart microgrids. We will explore tradeoffs between various data management and placement options in edge computing, and provide insight into what kinds of edge and cloud computing architectures are most suitable for Industrial Internet applications.

Assistant Prof. Klami is partnering with Associate Professor Aki Vehtari at Department of Computer Science, Aalto University, and Assistant Professor Antti Honkela at Department of Mathematics and Statistics, University of Helsinki. The consortium project Reliable automated Bayesian Machine Learning (RAB-ML) develops theory and methods for assessing and improving the quality of distributional approximations required for Bayesian inference. The results will be implemented as part of leading probabilistic programming systems to ensure wide applicability.

Associate Prof. Roos’s group coordinates a consortium together with the Helsinki Institute of Physics (HIP) focusing on developing and applying new probabilistic machine learning techniques and their applications in quantum mechanics-based atomistic simulations. The integration "white-box" machine learning techniques into large-scale simulations of atomic migration has potential for breakthroughs in the design of complex alloyed materials that can withstand extreme conditions in, e.g., fusion reactors and particle accelerators.

Dr Ruotsalo’s project Neuro-adaptive Intention learning aims to develop computational methods that can estimate and predict human intentions directly from human brain-signals measured via EEG and other sensors. The target is to develop methods that enhance truly implicit communication between the human and the computing system. Methods are developed to encode the user’s intention, but also to build generative approaches that can create new unseen content based on implicit brain-feedback.

Prof. Tarkoma's project is titled SHINE: Self-healing Software-defined Industrial Networks. The SHINE project aims to design a toolbox for the optimization of SDN deployment decisions in industrial Internet of Things (IIoT) networks that take into account the intricate requirements for managing hardware heterogeneity and scalability.

Prof. Toivonen’s and Prof. Männistö’s project Cooperation-Aware Software and Creative Self-Adaptivity aims to develop models and architectures for intelligently self-adaptive, collaborative software components. We combine research in software architectures with research in artificial intelligence, more specifically computational creativity.

More information:

 

 

12.10.2017 - 13:34 Veli Mäkinen
12.10.2017 - 13:26 Veli Mäkinen