How many times do you not type and your iPhone corrector corrects certain words by adding others that have nothing to do with it? And how many times, even when correcting, does the broker insist on a word you don’t want? It’s not a coincidence or a system failure, it’s just mechanics and artificial “intelligence” at work. At least the one explaining it is the creator of self-correction, Ken Kocienda.
One of the reasons the broker gives some suggestions is to fine-tune the auto-correction. But there is a reason behind the mechanism.
Does iPhone autocorrect counteract it?
The explanation appeared in the Wall Street Journal which conducted an interview with the creator of the automatic corrector, Ken Kocienda. It is very interesting to understand what is behind this tool that over the years has helped the planet not to make writing errors in the most varied applications.
Thanks to the automatic correction of our iPhone or iPad, the texts we write are free of spelling mistakes. In recent years, however, this same self-corrector seems to insist on substituting certain words for no apparent reason. But calm down, in fact, there is a reason, and also solutions in this regard.
According to Kocienda, it’s the shift to smarter systems that has led to autocorrect that doesn’t always work as expected. Just as Apple surprised the world with the original iPhone’s virtual keyboard, AutoCorrect has evolved and now draws on four sources of information that complement and in some cases neutralize each other to provide us with different results.
The main dictionary does not contain “insults”
There are rules for the autocorrector to receive input that is suggested to the user. First, a primary static dictionary is applied which, by the way, does not include bindings. Then a second “never auto-complete” dictionary, also static, which contains “swear words”.
Then we move on to a custom dynamic dictionary that learns from the words we write. And finally, a machine learning system that lets you learn words and phrases directly from the internet.
Thus, whatever the source of the information for the autocorrection, the correction system will try to propose the best alternative according to its primary dictionary and its modifications. Swear words are the best candidates to be replaced by others, but they are not always the only ones.
Fortunately, the system is smart enough to learn what we want to write. This is where the regular dictionary comes in, supplementing the main static dictionary with words we’ve written at least three times. However, as Kocienda recognizes, the artificial intelligence system that allows the broker to learn from texts on the web can lead to particular results. Sometimes you end up learning bad habits that have spread on the internet.
Can we help autocorrect?
The creator of the automatic corrector gave several examples where he himself had difficulties. Words that the system insisted on replacing and there was no way to correct the proposed correction itself. However, in these 15 years of existence of the iPhone, its keyboard software has evolved.
Thus, the user now has several ways to make the iPhone or iPad behave exactly as desired.
The simplest, which we have already mentioned, consists of writing a few words three times to include them in the dictionary. Another is to use text replacements in Settings > General > Keyboard > Text Replacements, because the words or phrases that we add to both phrase and quick function will be added to the dynamic dictionary.
Another option is to save certain keywords as contacts, which Joanna Stern, author of the WSJ article, mentions. Apparently, the reporter has saved all of her favorite slurs as contacts so she can quickly write them down whenever she wants.
Of course, as the author of the article says, we already depend on this help. She herself spent a week without the autocorrection system and noticed a greater difficulty not only in the speed of writing but also in the quality of spelling. Sometimes we want to write so quickly and so synthesized that without the corrector the mistakes would distort the meaning of the sentences even more. By the way, everyone remembers the generation of “tax bem amiga, tou xim…” of 2013!
These systems, in fact, cooperate with the user, because they learn and are resources that create in us certain good writing habits. Even if they sometimes plot against us!
We want to give thanks to the writer of this article for this incredible web content
Annoying AutoCorrect: Why Does iPhone Corrector Emphasize the Word We Don’t Want?
Check out our social media profiles as well as other pages related to themhttps://www.ai-magazine.com/related-pages/