What a mix-up
Wilhe rianedg, our bairn deos a gaert job. For crroect utidndsanenrg of a txet the crrtahaces of the wdors can be ralmndoy miexd up. One can unnaertdsd the text aslomt whioutt any pmbelros if the fisrt and last chcrrtaaes are in thier ogirainl piosntois. Of crouse, the txtes are becinmog more coeclapitmd and hdearr to raed the legnor the centioand wrods are. (Borrowed from here and slightly amended)
It`s fascinating
Did you manage to read the text? It’s fascinating, isn’t it? I even found myself taking a closer look at some words to see whether the letters really were mixed up or not. Do you always notice every single spelling mistake when you’re proofreading a text, especially one you wrote yourself? I bet you don’t! The more you read, the better trained your brain becomes, so that it recognizes whole words regardless, based on their meaning. It no longer matters so much whether all of the letters are in the right order as long as the first and last characters are in their original positions.
Data analysis in the alphabet jumble
Why am I telling you all of this? Our brain is a true expert when it comes to pattern recognition. The temptation to try and emulate this ability is enormous, and it’s no wonder that the human brain was chosen as a model for machine learning. Learning from data to identify patterns and laws. A veritable deluge of information abounds in the process industry. This data can be more or less structured, and the quality of it can vary considerably. Sometimes it’s easy to recognize a pattern, as with the text above, while sometimes it can be quite tricky:
Dynscorae vtopnireen sima ot dttcee nda eratt a seesida rylae no. Ti nucseldi qseethcuni cuhs sa cncare nad bronenw nesrngcie. *
The answer is: “Absolutely”
It may take a bit more effort but if you practice hard enough, you’ll be able to understand even texts like this one better too. You could be forgiven for asking yourself in this context whether there’s any point in having spelling rules in the first place if you can still read this kind of jumble. The answer is: “Absolutely!” After all, if no-one pays any attention to correct spelling, a text will be far more difficult to read and assimilate:
Wil riandg, ur brn dus a grt jb.
This same principle can be applied to your process data: if the quality isn’t up to scratch, it will be an uphill struggle to recognize a pattern. Clearly, though, data analyses – for example pattern recognition using machine learning techniques – hold considerable potential for extracting information and generating added value for plant operation. That’s why this is such an important topic in the age of the digital transformation. And it’s why I’m looking forward to ACHEMA 2018 and to meeting you there. Come visit us at our booth in Frankfurt (Hall 11.1 / Stand E16) and talk to us about the opportunities and challenges of this exciting new trend.
Hold on – we haven’t finished yet!
That last sentence probably would be the last one in 99% of all blog posts. However, there are exceptions to the rule (in case you hadn’t noticed). We learn and adapt flexibly to new challenges – often without our doing, unfortunately not without headaches and (usually) very fast. It helps to celebrate small successes as you go along and to recognize that something isn’t working before it has time to become a problem. That’s what I particularly like about Yokogawa’s new strategy for responding to the increasingly rapid changes in process automation. Synaptic Business Automation (SBA) is structured in such a way that you can adapt to new requirements fast, and in a manner that adds value, in cooperation with your customers. That’s valid for data analyses and it’s equally valid for PAT or Module Based Automation. What is there to gain from keeping your eyes fixed firmly on your goal, never daring to look to the left or right, only to realize in the end that it would never have happened if you’d asked for advice once or twice on the way…
* Secondary prevention aims to detect and treat a disease early on. It includes techniques such as cancer and newborn screening. (German original taken from Wikipedia), source: Fehler-Haft.de
