We have already had our first contacts with the new iPhone 11 and iPhone 11 Pro, and we have already been able to see how some of the abstract concepts that Apple explained to us in the keynote of last September 10th are beginning to be introduced into our daily lives without us being aware of it. Words’ such as Machine Learning and Deep Fusion are concepts that we are tired of hearing these days but that we still do not know completely because we have remained on the surface of the concept. If you’re interested in this topic, in this post we’ll go a bit deeper into it.
If we stay at the base of the word, and in the mere translation of it, we could say that Machine Learning (ML) is only automatic learning, and I say only because many users stay in this layer, something dark and forbidden that impresses just by hearing it. Really the merit is not in the enormous capacity that nowadays our iPhone has to make huge and complex calculations in almost impossible times to calculate. The real merit lies in the brains of the developers and in the layer of the teams that manage the large amounts of data needed to make this technology work, which in broad terms works in a similar way to Big Data.
Any good programmer tries to teach the machine so that it always optimally solves a problem, but when the data and calculations required sometimes tend towards infinity, you have to use another strategy and that better than the machine makes mistakes and learns by itself . To do this, “heuristic” decisions are made that aim to give the software the ability to make decisions by “intuition “. It is something similar to the heuristic search of an antivirus, it is possible that a file presents anomalies that, although it does not appear as an infected file by comparison with another infected one, makes our antivirus software think that it could be and stores it in quarantine. In short, we teach the software to decide for itself , and although at first it is erratic, statistics tell us that little by little it will be almost as effective as a human being can be in making decisions.
The Machine Learning algorithms are designed so that with fewer resources, large volumes of data can be processed, and you can learn by yourself, something like the WOPR machine in the War Games movie.
It is important to understand this concept in order to be able to say that Apple has indeed innovated with the new iPhone 11 . It has innovated in the way it implements artificial intelligence in photography. Someone could say that apple is not the first company to implement ML in its photographic treatment, and in that we agree, but if it is the first to implement it as it has done, working with a large number of photos in real time, before and after pressing the shutter. From all this, it has been necessary to implement a tremendous processor A13 Bionic , which is capable of supporting these infinite calculations and in a minimum amount of time . For this, and no other obscure reasons, an iPhone XS will not be able to perform night mode as it does not have an operational level on its A12 Bionic processor.
As always, in the keynote of September 10th, Apple sinned of being modest and did not explain in an extensive way how the bestial A13 Bionic processor works. A powerful processor in itself does not add value to a device, and it also does not provide anything great algorithms if there is no brute force to move it. But as always, and here is the proof, Apple has managed to perfectly merge the hardware with its software . A processor like the one Apple has created should not be sold for the amount of operations it is capable of, but for how it integrates with the software it has to move. Once again, the absolute mastery of the signature of the bitten apple in the fusion of hardware and software is demonstrated.
That’s why we think we can say loud and clear that once again Apple innovated in how it does things , inventing algorithms that are impossible for common processors to execute. The custom design of a new processor that supports those calculation speeds is simply engineering mastery.