Russian search giant Yandex is formalizing its partnership with CERN by joining its Openlab program. The latest collaborative effort sees Yandex offering up its proprietary MatrixNet machine learning technology, which powers the company’s internet search engine. CERN, the European Organization for Nuclear Research, plans to harness MatrixNet to analyze and search through the colossal data outputs generated by the use of the Large Hadron Collider (LHC).
Initially, CERN will put MatrixNet to work combing through data sets from the B-meson decay analysis, provided by the LHCb experiment (the b stands for beauty). Yandex has provided a team of its engineers to assist the researchers in using the MatrixNet software, but the CERN Openlab collaborators are developing a user-friendly front-end interface for the CERN researchers to use in submitting queries on their own.
The LHCb experiments generate colossal amounts of raw data — as much as 35GBcovering 1 million events per second — which needs to be organized and filtered to find the extremely rare events that the scientists are looking. CERN has systems in place that can pare down the relevant data, but a good chunk of it still requires relatively extensive processing and analysis to determine relevance and usefulness. Needless to say, it is a daunting task.
More of the story, click image