Please read these sentences:
Popele lkie to use scoail meida scuh as Tiwettr and Fbooacek
Penjoroctis form the Dmeeprtant for Eiadtcuon sseuggt the ppiul puolotapin in Elgnand wlil saor by asomlt a milloin in the nxet dadece.
You’ll probably experience that you can read many of the words. Though, some words may be too challenging.
Before our algorithm, it was unknown how the brain could solve this challenge. With our algorithm, we can explain how the brain could find the correct spelling. That’s within the rather limiting elements of brain architecture, such as the extremely slow conduction speed of neurons.
In addition, we can demonstrate it through our partially programmed algorithm. It resolves any of the 13,643 words loaded with 99% accuracy. To solve this challenge, it only needs three layers of the model’s equivalent to what brain researchers know as minicolumns.
In turn, this provides us one of more reasons as to why our algorithm is truly brain based. Hence, our confidence in providing results too challenging to achieve with today’s A.I. approaches.