Artificial Intelligence at Mail.ru Service

    Good day, dear Habr Community!
    I thought for a long time and tediously about what my first post here will be.

    And today I decided to talk about artificial intelligence and stolen (stolen) accounts Or rather, how the first can come to the rescue and fight the second.

    All names and addresses mentioned in this article are fictitious. Any coincidence with real people is accidental.



    Background


    It just so happened that I have been a user of Mile.Ru for many many times over .
    Being engaged in offers of friendship, I faced a situation of frank porn spam, both immediate and postponed.
    (We are talking about accounts that advertise all kinds of services that are not in special demand with me.)

    Observations of such accounts showed
    that each of them has its own characteristics, among which it is possible to distinguish Users whose display name and mailing address do not match each other, for example:
    1. Semenova Ekaterina (kaleria.petrova@mail.ru)
    2. Vasilyeva Daria (xxxxxx @ mail domain)



    Solution (currently manual):

    1. Check accounts against the list of mutual friends. If there are friends on this list who enjoy my trust (in principle, they do not keep spammers in friends lists), then you can add such an address to your friendship list.
    2. Check accounts for synonymous usernames in the mailing address, for example: Marina Petrova (marishka @ mail domain, m.petrova@ mail domain, mary @ mail domain and their combinations)

    Since laziness is the engine of progress (I think fellow programmers will agree with me), it was decided to make a timid attempt to cry over the development of an Artificial Intelligence algorithm that can solve this problem with minimal human involvement.

    I invite you to a discussion.
    UPD: All the necessary technologies on Mail already exist.

    Also popular now: