Receive regular emails from NPL to get a glimpse of our activities and see how our experts are informing and influencing scientific debate

Machine learning

Webinar: machine learning for industrial measurements

A key feature of the Factory of The Future vision is that large networks of various types of sensor will be used to collect data throughout the factory, with the data being used to drive automated decision-making to optimise product quality and minimise production costs. This transition, to what is commonly known as Industry 4.0, will generate large volumes of complex interconnected data and one of the key challenges is to harness this data to make decisions in a principled and trustworthy manner.

Machine learning (ML) is a branch of artificial intelligence that can be used to analyse large volumes of complex data. ML is particularly strong at identifying patterns, connections, and common features in data sets, and can often find links in data sets that physical intuition would not provide. ML models are often a “black box” in the sense that the model inputs and outputs are known, but the inner workings of the model are not well understood. Therefore, it is challenging to assess confidence in the outputs of ML models – a crucial requirement for the digital transformation of manufacturing.

In response to this challenge, NPL is hosting a webinar on the introduction to machine learning for industrial measurements. The webinar will give an overview of how ML can be applied to industrial applications and how metrology principles of uncertainty, calibration and traceability are needed to ensure confidence in machine learning outputs.

As part of an EMPIR project, NPL is working with other NMIs, including PTB, LNE, VSL, INRIM and IMBiH, and external partners, including the Universities of Strathclyde and Cambridge, ZEMA and SPEA, to establish a metrological framework to support the complete lifecycle of measured data in industrial sensor networks. The framework will ensure transparency, comparability and sustainable quality of measured data, processing methods and measurement results.

Find out more and register here.

23 Feb 2021