ARTIFICIAL INTELLIGENCE: From Alan Turing’s Imagination to 'Straight-Through-Processing' in Seven Decades

When Alan Turing came up with the notion of artificial intelligence in his 1950 paper, “Computing Machinery and Intelligence” he nailed the concept, but technologically the world wasn’t ready. Turing saw that if humans could use available information and apply reasoning to make decisions which solved problems, then surely machines could be programmed to do the same. 

Of course, he was right, and back in the 50s accessibility to technology was the only thing holding him back. Computers were incredibly expensive, painfully slow and had crippling storage limitations in comparison to what we have today. Machines could be told what to do, but we weren’t yet in a position where they could be expected to remember what to do.

Fast-forward seven short decades and we have everything we need and more, to automate all our mundane and repetitive business admin tasks. This has given us high confidence in Straight Through Processing (STP) - the scenario where machines have become confident enough to bypass human involvement for certain tasks all together. As a result artificial Intelligence has gained momentum in the mainstream, and will continue to do so. Exciting times! 

AI experts and experienced users of AI have realised that there is a balance to be had in terms of trying to hand everything over to machines, and having ‘human-in-the-loop’ for machine learning, for optimum results.

At Celaton, for our inSTREAM Intelligent Document Processing Platform, we combine of a wide range of technologies. Including some of the very latest tools such as the ‘Entity Recognition Model’ or NED (Named Entity Recognition) which applies Natural Language Processing to understand large volumes of unstructured human language – with the ability identify it and extract it. But we also use tools that have been around a little longer, such as OCR (optical character recognition), Intelligent Character Recognition, and Machine Learning. We even use binary trees as part of random forest classifiers and decision loops where appropriate.

The point is that the ability to expertly select and integrate available as well as up-and-coming technologies to increase confidence in ‘Straight-Through-Processing’ is what buyers of AI solutions want in 2022.

inSTREAM is the AI platform of choice for document processing for the likes of Currys, ASOS and Capgemini.