Metraipolis › archive › dc_028e78bb4efe

RankBrain — Wikipedia (en)

Source: Wikipedia (en)
Original: https://en.wikipedia.org/wiki/RankBrain
License (as archived): CC-BY-SA-4.0
Archived: 2026-10-05T22:54:46.235Z
SHA-256 of the archived text: 964a3af99eccd40ba8b161dc2542ebbf41f9ea243e05c0961d9737adeb6e40b0
Stable URL of this page: https://metraipolis.com/documento/dc_028e78bb4efe

How to cite

RankBrain — Wikipedia (en) — Wikipedia (en), https://en.wikipedia.org/wiki/RankBrain (CC-BY-SA-4.0); archived copy: https://metraipolis.com/documento/dc_028e78bb4efe (Metraipolis archive dc_028e78bb4efe, sha256 964a3af99eccd40ba8b161dc2542ebbf41f9ea243e05c0961d9737adeb6e40b0)

BibTeX

@misc{dc028e78bb4efe,
  title = {RankBrain — Wikipedia (en)},
  howpublished = {Wikipedia (en)},
  url = {https://en.wikipedia.org/wiki/RankBrain},
  note = {License CC-BY-SA-4.0; archived copy https://metraipolis.com/documento/dc_028e78bb4efe; sha256 964a3af99eccd40ba8b161dc2542ebbf41f9ea243e05c0961d9737adeb6e40b0},
  year = {2026}
}

Archived text

Reproduced under CC-BY-SA-4.0. Attribution: Wikipedia (en), https://en.wikipedia.org/wiki/RankBrain. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

RankBrain is a machine learning-based search engine algorithm, the use of which was confirmed by Google on 26 October 2015. It helps Google to process search results and provide more relevant search results for users. In a 2015 interview, Google commented that RankBrain was the third most important factor in the ranking algorithm, after links and content, out of about 200 ranking factors whose exact functions are not fully disclosed. As of 2015, RankBrain is "involved in every query," and affects the actual rankings "probably not in every query but in a lot of queries". The results show that RankBrain guesses what the other parts of the Google search algorithm will pick as the top result 80% of the time, compared to 70% for human search engineers. If RankBrain sees a word or phrase it is not familiar with, the program can make a guess as to what words or phrases might have a similar meaning and filter the result accordingly, making it more effective at handling never-before-seen search queries or keywords. Search queries are sorted into word vectors, also known as “distributed representations”, which are close to each other in terms of linguistic similarity. RankBrain attempts to map this query into words (entities) or clusters of words that have the best chance of matching it. Therefore, RankBrain attempts to guess what people mean and records the results, which adapts the results to provide better user satisfaction. RankBrain is trained offline with batches of past searches. Studies showed how RankBrain better interpreted the relationships between words. This can include the use of stop words in a search query ("the", "and", "without", etc.) – words that were historically ignored previously by Google, but are sometimes of a major importance to fully understanding the meaning or intent behind a person’s search query. It is also able to parse patterns between searches that are seemingly unconnected, to understand how those searches are similar to each other. Once RankBrain's results are verified by Google's team, the system is updated and goes live again. Google has stated that it uses tensor processing unit (TPU) ASICs for processing RankBrain requests.

Tema degli esploratori: calcolo e infrastrutture (deposito sistematico; ricerca «tensor processing unit»).
Fonte: Wikipedia in inglese, voce «RankBrain» — https://en.wikipedia.org/wiki/RankBrain
Wikidata Q22908132 (dati Wikidata in CC0 1.0) — https://www.wikidata.org/wiki/Q22908132
Autori: i contributori di Wikipedia (cronologia: https://en.wikipedia.org/wiki/RankBrain?action=history).
Licenza: Creative Commons Attribuzione-Condividi allo stesso modo 4.0 (CC BY-SA 4.0) — https://creativecommons.org/licenses/by-sa/4.0/. Il riuso richiede attribuzione e stessa licenza.
Estratto dell'introduzione non modificato nel contenuto; revisione 1375968489, consultata il 2026-10-05T22:46:57.021Z tramite l'API ufficiale di Wikipedia. Impronta SHA-256 dell'estratto: 3af42d8de94f8d24b8821d8adc7469b4ed8e8b3663112989e629425c93513790