European project OpenWebSearch.EU

OpenWebSearch.EU is a project to develop and pilot the core for a European Open Web Index (OWI) and the foundation of an open and extensible European Open Web Search and Analysis Infrastructure (OWSAI) by bringing together strong European players, who jointly define, develop and pilot an open technological backbone for cooperative web search. The proposed pilot infrastructure will demonstrate, how search applications and web-based AI data products can be realized through cooperative crawling, analysis, storing and indexing of web content. The project will demonstrate the feasibility and potential of an open European web index and how it stimulates a competitive web search and web data product market. Therefore, the pilot aims to reach a technology readiness level (TRL) of 5.

Project members: Arjen de Vries, Djoerd Hiemstra, 1 PhD position, 1 PostDoc position

External Link:

NWA project LESSEN: Low Resource Chat-based Conversational Intelligence

LESSEN aims to make the chat technology accessible for languages and domains with relatively little training data and compute power. It focuses on developing data and compute efficient chat algorithms that can make optimal use of data and offer safe and transparent task-oriented conversational agents. Dr. Hasibi receives 500K funding and works on augmenting conversational training data and developing algorithms that enrich data-hungry models with rich structured information stored in Knowledge Graphs. The consortium is led by Prof. Maarten de Rijke (UvA) and consists of University of Amsterdam, Leiden University, University of Groningen, Amsterdam University of Applied Sciences, Radboud University, Achmea, Albert Heijn,, KPN, Rasa Technologies, Ahold Delhaize, and National police force.

Project members: Faegheh Hasibi, 2 PhD positions

External link:

NWA project OBSeRVeD - Odour Based Selective Recognition of Veterinary Diseases

When chickens in a farm become infected or have parasites, specific odours are produced. A cross-disciplinary team will combine innovative sensors, affinity layers, and machine learning to develop and test an electronic nose. This sensitive system can recognize a fingerprint of Volatile Organic Compounds and thus recognise specific diseases at an early stage, when (preventive) measures are most effective. In this project, veterinary health, industry, science professionals and societal organisations will collaborate towards developing a practically applicable poultry health monitoring system to improve chicken and public health, general welfare and reduce antibiotics/chemicals use and the environmental impact of livestock farming.

Project members: Tom Heskes, 1 PostDoc position

The Heat Is On (MOOI Grant)

Start date: 1 May 2021

For industries like agrofood, paper industry and specialty chemicals, some 40-80% of the COemissions is related to the energy that is needed for heat-driven processes like separation and drying. ‘The heat is on’ focuses on heat integration, a balancing act between heat generation and use/reuse. Due to the magnitude of energy consumption, a small change to achieve more energy efficiency in one part of the process, can have substantial effects. The process optimization is realized by making use of digital twins.

Project members: Thanh Tran, Tom Heskes

External link:

Detecting manmade structures from space

Start date: 1 May 2021

The ongoing developments in machine learning and remote sensing open up promising opportunities to fill this gap. In this Radboud AI voucher project, we will first focus on the detection of dams. Our goal is to build and evaluate a convolutional neural network based on satellite imagery. We will use existing datasets of dam locations as input information and to evaluate the performance of our network. Finally, we will apply the network across Europe to estimate the expected runtime and feasibility of a subsequent global application. We believe that in the future, our project can be a meaningful contribution to the automatic and efficient detection of manmade structures, leading to improved environmental assessments.

Project members: Konrad Mielke, Tom Claassen

External link:

Transfer Learning for Federated Search

Start date: 15 February 2020

Project members: Negin Ghasemi, Djoerd Hiemstra

Tine: Automatic grading and gamification for open questions

Start date: 1 February 2020

Project members: Hendrik Werner, Patrick van Bommel, Djoerd Hiemstra

EU-EFRO MARBLE (sMARt Borstkanker detectie met deep LEarning)

The objective of the 4-year project MARBLE is to develop and validate AI algorithms for breast cancer screening. In MARBLE, our industrial partner Screenpoint Medical, providing valorization, and the RUMC are directly involved.

Project members: Gijs van Tulder, Elena Marchiori, Yao Tong

NWO Crossover programme MOCIA (Maintaining Optimal Cognitive function In Ageing)

The objective of the 6-year project MOCIA is to develop and validate AI algorithms for identifying non-invasive modifiable risk and protective factors and for designing scoring tools to quantify risk of cognitive decline.

Project members: Elena Marchiori

TOP ZonMW Parkinson (Big Data for Personalized Medicine)

Project members: Luc Evers, Tom Heskes, Jesse Krijthe

Data2Person Multiple Sclerosis (Self-monitoring Based Management of Multiple Sclerosis)

Project members: Gabriel Bucur, Tom Heskes, Jesse Krijthe

TTW DrComics (Dose-response Curves for –OMICS data in soil quality assessment)

Project members: Yuliya Shapovalova, Tjeerd Dijkstra, Tom Heskes

NWA PrimaVera (Predictive maintenance for Very effective asset management)

Mariëlle Stoelinga leads an integral project on predictive maintenance of asset management. In this project, big data algorithms are used to better predict malfunctions in infrastructure and production resources and thus enable better planning of maintenance. This is called predictive maintenance.

Project members: Roel Bouman, Tom Heskes

NWA CORTEX (Center for Optimal, Real-Time Machine Studies of the Explosive Universe)


The CORTEX consortium of 12 partners from academia, industry and society will make self-learning machines faster, to figure out how massive cosmic explosions work, and to innovative systems that benefit our society. RU will contribute to this project by developing new self-learning algorithms.

Project members: Aleks Kolmus, Twan van Laarhoven, Tom Heskes

NWO TOP CHiLL (Causal Discovery from High-Dimensional Data in the Large-Sample Limit)


Existing algorithms for causal discovery from observational data are not very well suited to big data: small changes in the data or in the algorithmic details can lead to significantly different causal conclusions, in particular for data sets containing many different variables and even in the limit of a large number of samples. In this project we aim to tackle these issues through a much better mathematical understanding of the appropriate asymptotic statistics.

Project members: Gabriel Bucur, Konrad Mielke, Tom Claassen, Tom Heskes, Gido Schoenmaker, Nastaran Mohammadian Rad, Ruifei Cui