Yandex Announces Vega Update For Its Search Engine

Yandex has announced a significant algorithm update that features up to 1,500 improvements. And the Russian search engine is calling it the Vega update.

One of the significant changes in the update is the introduction of human feedback to the algorithm training. Also, the tech company says Vega brings the ability to double the size of its search index without harming search result speed.

In the announcement, head of Yandex Search, Andrey Styskin said:

“At Yandex, it’s our goal to help consumers and businesses better navigate the online and offline world. With this new search update, users across the RuNet are helping us do just that.”

Here’s a break down of the Yandex update.

How the Vega Update Improves Yandex

Here are four things you should know about the Vega Update.

1. Adding Human Input to Algorithm Training

Like Google, Yandex also uses quality raters – assessors – to test new algorithm changes.

However, the Russian search engine went a step further by employing experts to review the assessor’s work for improved accuracy. So, since these experts vouch and verify Yandex training data, it’ll presumably be more accurate.

“By training our machine learning algorithms with expert assessments, our search engine learns to rank relevant information higher in results thanks to the work of a highly qualified group of individuals,” says Yandex.

2. Predicting Search Queries and Pre-Rendering Results

Yandex says that it has been using a pre-rendering technology which predicts the user’s query and selects relevant search result while the user is still typing.

Although the company announced the pre-rendering technology in the context of the vega update, it has implemented it on Android devices since March. One significant advantage of the feature is that is speeds up the time it takes users to find answers to a query.

3. Crowd Sourcing Search Result Raters

Google hires contractors that are familiar with its quality raters guideline to judge search results. Yandex, on the other hand, is relying on its crowdsourcing platform, Yandex. Toloka to do the same.

Compared with Google’s method, Yandex’s crowdsourcing is less controlled. However, the Russian search engine provides rater guidelines to improve rating accuracy.

4. Expanding Search Index with Clustering

In the past, Yandex would search through an entire index for an answer to a query. Well, not anymore.

The Russian search engine now group pages into topical clusters. So, instead of searching the entire index for an answer, the Yandex now seeks the most topically relevant clusters.

Thanks to this clustering technology, Yandex was able to double its index to 200 billion pages without impacting the speed search result speed.