The Consortium

PTB ’s technical role in this project will focus on the development of statistical estimation procedures for smart measurements using ML and on the implementation of the developed methods for sHSI. They were the first to demonstrate compressed sensing for sHSI. They also have a track record in uncertainty evaluation and large-scale inverse problems and PTB is member of JCGM WG 1. PTB will coordinate this project and will participate in all WPs (leading WP4 and WP6). They have experience in coordinating previous EMRP and EPM projects related to uncertainty evaluation methods (e.g. EMRP JRP NEW04 Uncertainty and EPM 22DIT01 ViDiT) and participated in many EMRP/EMPIR/EPM JRPs including IND10 Form, IND09 Dynamic, IND52 xDReflect, ENG63 GridSens, 15HLT05 PerfusImaging, 15SIB01 FreeFORM, 17NRM05 EMUE, 18HLT05 QUIERO, 20IND08 MetExSPM, 20IND09 PowerElec, 20IND12 Elena, 22HLT06 MAIBAI. 

CMI’s technical role in the project will focus on development of hyperspectral Scanning Probe Microscopy methods. CMI is a national metrology institute of the Czech Republic. Department of Nanometrology has experience with SPM methods for more than 20 years and is active in all its aspects from instrumentation up to data processing. In the project, CMI leads WP3.

NPL’s technical role in the project will focus on the development of methods for uncertainty evaluation and their validation, and the production of a prototype for the photocurrent mapping of semiconductors application. They currently lead WPs focused on ML uncertainty evaluation in two EPM projects (22HLT01 QUMPHY and 23NRM03 BioAirMet). They led the writing of the AI and ML topic in the MATHMET Strategic Research Agenda, and they are a member of JCGM WG1. They have used compressed sensing to achieve megapixel-resolution PCM of semiconductors for the first time.

TUBITAK aims to ensure the reliability of all measurements conducted in Turkey. TUBITAK’s technical role in the project is mainly to implement the algorithms developed in the project using the LIBS infrastructure as a use case and to develop a fast and smart measurement procedure.

VTT’s technical role in the project is to participate to the development of generic approaches for use of ML in metrology and specially to apply these methods in the hyperspectral imaging scatterometer application (HIS). VTT has extensive experience in nanoscale metrology, microscopy, diffractometry, and scatterometry for characterizing micro-/nanostructure surface features directly or indirectly. VTT has long experience in utilising ML methods for various applications.

Electro has expertise in advanced materials and measurement. The company’s Technical Director has over 35 years of experience in applied R&D in academic and industrial environments. Electroscience’s develops novel metrology tools for characterising piezoelectric and ferroelectric materials. Their capabilities extend to the use of ML of large and sparse materials property datasets for correlation analysis for touch-sensitive and haptic piezo technologies and in the analysis of in situ metrologies developed under EMPIR Advent and OpMetBat for optical, electro-mechanical and structural correlation.

Finden has experience in characterisation and analysis services to industry across sectors that involves big data collected at large scale facilities (predominantly synchrotrons) They regularly publish on both material systems and methodological developments including ML/DL. Finden’s main role in the project will be coding developed approaches for their release within open-source tomographic reconstruction platforms, such as CIL and LION. Also, with support from UCAM, Finden will investigate the applicability of developed ML-based regularisation procedures to X-ray scattering computed tomography.

JCMwave has a long-standing background in the development and application of ML tools for solving inverse problems as well as simulation tools for rigorous simulations of Maxwell’s equations with an emphasis on nanooptics applications. JCMwave contributes to WP1, WP2, and WP4 by providing simulated data for sHS and supporting the analysis of data from the hyperspectral imaging scatterometer.

Random Red has expertise in digital technologies and their application, with an emphasis on the digital transformation of metrology. Random Red has experience in collaborative R&D projects, such as NGI Trublo Metroracle project, NGI TrustChain MorphMetro, EPM projects 22DIT02 FunSNM and 23NRM01 SBS Uncert, and through IT projects related to data engineering, ML and cyber-security. Random Red will contribute to advancing digital metrology by aligning with industry standards and digital transformation trends, ensuring that project outcomes meet the evolving needs of both sectors – metrology and information technology.

RWTH represented by the Laboratory for Machine Tools and Production Engineering (WZL), focuses in its technical role on the ML-based solving of inverse problems and the corresponding uncertainty estimation as well as the ensuring of FAIR data and state of the art research data management. Under the topic of (measurement) uncertainty evaluation, statistical data analysis and ML-based modelling the WZL is performing research in various German and international research projects. WZL is member of the VDI-WG 1.20 and VDA-WG 5 which are responsible for different guidelines in measurement uncertainty determination and the proof of suitability of inspection processes. WZL will participate in all WPs except WP3 (leading WP5). The RWTH participated in different EMRP/EMPIR/EPM JRPs including IND53 LaVa, 20IND02 DynaMITE and 23IND12 ADAM.

SOLEIL is the French national synchrotron radiation facility. The participating lab from SOLEIL is the SMIS infrared spectromicroscopy beamline that has expertise in a wide scientific and instrumentational scope of infrared experiments from solid state physics to biomedical research as well as experience in developing machine learning tools for spectroscopy data analysis. The SOLEIL team will participate in work packages WP1, WP4 and WP5 to contribute with training and testing data to the ML model development, the design and definition of measurement protocols their execution as well as report writing and reviewing for the relevant work packages.

UCAM, represented by the Cambridge Image Analysis group (CIA) at the Department of Applied Mathematics and Theoretical Physics, will focus in its technical role on the development of ML-based approaches for solving inverse problems, especially on learned regularisation and learned denoisers in a variational and plug&play framework, respectively. The CIA has been contributing to the ML revolution within inverse problems with seminal papers on learned regularisation and learned inversion approaches with mathematical guarantees, as well as task-adapted reconstruction within an ML for inverse problems framework. These have been published in top mathematical venues (e.g. Acta Numerica) and ML conferences (NeurIPS, ICML, ICLR). UCAM will participate in all WPs (leading WP1).

UNISA is ideally linked to the oldest academic institution in the Old Continent: the Schola Medica Salernitana. Founded in the eighth century AD, the School reached its peak between the 10th and 13th century. The measurement group has strong expertise on the MEMS characterization, uncertainty evaluation, neural network, machine learning and uncertainty assessment in Artificial Neural Networks. In the project UNISA will contribute to WP1, WP2 and WP5.

INM (Instituto Nacional de Metrología). Is the National Institute of Metrology of Colombia. INM provides metrological traceability to calibration and testing laboratories in Colombia. INM has connections with industry and other productive sectors in Colombia. Furthermore, INM actively participates in the different Working Groups (WPs) of the Sistema Interamericano de Metrología (SIM) - Inter-American System of Metrology, interacting with other metrology institutes in Latin America; in this way, INM represents an opportunity for the project to make an impact beyond European borders and strengthen INMs advancements in ML for uncertainty evaluation. INM has participated in the European Union project LA/2019/407-085 Quality for Competitiveness – Reducing Quality Gaps of Regional MSMEs in Colombia.