New publicly funded projects related to Flanders AI Research

With ICON (Interdisciplinary Cooperative Research) projects, imec, Flanders Make and VLAIO  want to bridge the gap between research results of the strategic research centers and their practical applicability within the Flemish business community.

VLAIO recently approved eight of these projects (site in Dutch), of which five relate to research of the Flanders AI Research Program.

The BOCEMON-project research how information on an office building (sensor information or user feedback) can be used to improve both user comfort and energy efficiency of the building.  The AI methods make use of Reinforcement Learning.
Reference: Grand Challenge 4, WP4, research groupIDLab UAntwerpen/imec.

The MultipLICITY-project will use AI-based real-time algorithms and highspeed cameras to control the 3D printing process.  The goal is to improve the quality of the 3D printed products, reduce waste and save energy. 
Reference: Edge AI, Grand Challenge 2 , research groups UGent-IPI/imec, UAntwerpen-VisionLab/imec, Flanders Make.

The OptiROutS project will use AI-based algorithms for optimal routing, taking into account also social aspects such as the societal cost of cut-through traffic. It will use Hybrid AI methods that combine simulation models with graph-based methods (GCN, spatio-temporal graph-based convolutional networks). 
Reference:  Grand Challenge 4, WP4;  research groupIDLab UAntwerpen/imec.

The FARAD2SORT project wil use AI to facilitate the automatisation of quality inspection and the use of robots:  facilitate the annotation of the data, the automatic retraining and speed optimalisation. This Flanders Make ICON project builds upon research for the automation of the AI flow and on edge/cloud deployment.
Reference: Grand Challenge 1 , Grand Challenge 2,  research groups UHasselt EDM, UGent / imec IPI.

The project TUPIC will use AI-methods such as Reinforcement Learning to facilitate the use of control algorithms in complex industrial processes.  The project will use “physics-based learning” methods .
Reference: Grand Challenge 1, WP 4, research group UGent EEDT-DC.