[ELI5] Why does autocorrect insist that the first letter of a misspelled word is more important than the rest of it?

Question

For example, if I spell “umportant”, it’s easy for us to recognise that it’s supposed to be “important”, but autocorrect insists that it’s something like “umbrella”, or I guess more logically “unimportant”, even though “important” is only 1 correction away.

​

These are real examples from my phone (Samsung Galaxy):

​

Wuick gets the suggestions Wicked, Which, Wucky, Whickham, Whicker, Wick, Wickets, Wicket, and Wickham. None of which are “Quick”, what I intended to write.

​

Nrown gets the suggestions Now, Nr own, Noon, Nowhere, Nr owner, Nr owns, and Nr owners. None of which are “Brown”.

​

Dence gets the suggestions Dance, December, Denied, Dancers, Decent, Dense, Dench, and Deuce. None of which are “Fence”.

​

It’s bothered me for years that it never ever picks up on a misspelt first letter.

​

Edit: I tried “umportant”, and it actually comes with 0 suggestions. Not umbrella, not unimportant, not even “important”. But “inportant” and “ikportant” and even “iqportant” are all recognised as “important”.

In: Technology

Answer 1

I just started coding. This is my small understanding.

As we type, autocorrect checks if the words are present in a library of words. If it’s not present, then it’s misspelled and needs to be edited. It can be edited by inserting, deleting, switching, and replacing letters until it forms words that are present in the library. Suggestions can be based on how frequently words are used globally. In the case of smartphones, the words we frequently use and select for autocorrect are counted locally.

Ive too felt that autocorrect used be better for certain things. However, I think AI has changed the needs for autocorrect. It’s no longer just about checking a dictionary for suggestions, or finding typos by comparing what keys are in proximity of each other and often get swapped. Now, with natural language processing, there are other factors such as: what words are commonly used together, what word and syntax patterns a person tends to use.

I think in order to gather better data and save on processing power, autocorrect has less focus on proximity typos and now more on word and sentence comprehension.