Artificial intelligence is mostly about using technology to resemble the way we think and learn. Once we pick up a skill or gain some knowledge, we use it to reason out. Interestingly enough, we can use artificial intelligence to detect spam in the same way.
Natural-language processors
Simply speaking, natural-language processors are a group of complex algorithms. The job of each algorithm is to scan e-mail messages to discover the content of them.
Natural-language processors serve as a powerful artificial intelligence tools in fighting against spam. An AI mail-filtering software routes unsuspected traffic to corresponding the SMTP server(The server used for sending e-mail), and flags other messages as spam.
How Does The Natural-language Processing Technology Filter Out Email?
As stated above, it works similarly to how we judge if an e-mail is spam. We first skim a message, and reference keywords by their location in a sentence. Then we understand the difference between “Apple Inc.” is a company(Amazing computer!), and “Apple pie” is food. Similarly, natural-language processing technology combines all pieces of messages and analyze the words, sentences and paragraphs.
Traditional Way of Email Filtering By using Keyword Analysis
Consider the following e-mail:
“Hey Bob, the chicken breast looks good to eat so let's cook it out tonight. Also, it will be great if you can bring Mary. Call 408-637-2222. Oh, I post the pictures about the dinner we had last week, and here is the URL www.sharephoto.com/2342/dinner.”
A standard keyword analysis would flag the words: breast, look good, phone number and Web site URL, and classify the e-mail as spam but artificial intelligence can figure out the message is an invitation to dinner.
To Conclude
As far as I know, Artificial intelligence technique is the most natural and efficient technique to filter e-email.
It is good that you put small headers before each paragraph to help the reader understand what each paragraph is saying. Your definition about artificial intelligence was easy to understand: make technology think like humans. I thought it was interesting that you talked about e-mail spam instead of robots. Basically, using your example, by looking at keywords, normal analysis might show something as spam, but if it were artificially intelligent, it would know that it is not spam, so it goes in the inbox. Nice post!
ReplyDeleteHello! When I think of Artificial Intelligence, usually what comes to mind are video games, voice recognition, etc. However, I would never think of email-filtering. I like this idea of looking for key words for the AI to determine if the email is spam or not. However, I think this method is not totally reliable since anyone could use keywords in a way that AI recognizes it as spam. I once was expecting an important email from a certain source. All of the emails went to spam because of the keywords but it was not spam! In any case, I do believe that this usage of AI saves a lot of time.
ReplyDelete