Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Friday October 25 2019, @05:06PM   Printer-friendly
from the when-it-misunderstands-let-me-do-a-literal-search dept.

Google is currently rolling out a change to its core search algorithm that it says could change the rankings of results for as many as one in ten queries. It's based on cutting-edge natural language processing (NLP) techniques developed by Google researchers and applied to its search product over the course of the past 10 months.

In essence, Google is claiming that it is improving results by having a better understanding of how words relate to each other in a sentence. In one example Google discussed at a briefing with journalists yesterday, its search algorithm was able to parse the meaning of the following phrase: "Can you get medicine for someone pharmacy?"

The old Google search algorithm treated that sentence as a "bag of words," according to Pandu Nayak, Google fellow and VP of search. So it looked at the important words, medicine and pharmacy, and simply returned local results. The new algorithm was able to understand the context of the words "for someone" to realize it was a question about whether you could pick up somebody else's prescription — and it returned the right results.

The tweaked algorithm is based on BERT, which stands for "Bidirectional Encoder Representations from Transformers." Every word of that acronym is a term of art in NLP, but the gist is that instead of treating a sentence like a bag of words, BERT looks at all the words in the sentence as a whole. Doing so allows it to realize that the words "for someone" shouldn't be thrown away, but rather are essential to the meaning of the sentence.

https://www.theverge.com/2019/10/25/20931657/google-bert-search-context-algorithm-change-10-percent-langauge

I can't help wondering how this is handled in other languages. Chinese, for instance, where (I've read that) a word might have several meanings, but context defines the word.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 0) by Anonymous Coward on Friday October 25 2019, @06:21PM (3 children)

    by Anonymous Coward on Friday October 25 2019, @06:21PM (#911775)

    the other 90 they let go to shit.

  • (Score: 3, Informative) by ikanreed on Friday October 25 2019, @06:54PM (2 children)

    by ikanreed (3164) Subscriber Badge on Friday October 25 2019, @06:54PM (#911795) Journal

    Honestly, there's a specific subset they've been improving: topical searches, where you search a vague idea and usually find what you want in the first result.

    And there's a different subset they've actively punished: when you know what kind of information you want exactly and want to explicitly filter your results to not include similar concepts and just want to find an exact phrase dammit.

    • (Score: 2) by darkfeline on Friday October 25 2019, @10:57PM

      by darkfeline (1030) on Friday October 25 2019, @10:57PM (#911898) Homepage

      > just want to find an exact phrase dammit.

      This is very much a case of "be careful what you wish for" and "users don't know what they want".

      Doing the equivalent of grep over the entire Web will only get you billions of Viagra spam websites, to say nothing of the fact that that's infeasible to index.

      --
      Join the SDF Public Access UNIX System today!
    • (Score: -1, Flamebait) by Ethanol-fueled on Saturday October 26 2019, @12:00AM

      by Ethanol-fueled (2792) on Saturday October 26 2019, @12:00AM (#911919) Homepage

      Install, for example, Firefox and Brave. Independently search for something specific and watch Jewgle give you wildly different results despite identical searches. Jewgle and other deepstate globalist bullshit Is out to jew you mentally if not financially. That is its nature.