Information processing method and system

Abstract

The invention discloses an information processing method and an information processing system. The method comprises the steps as follows: a server is used for decomposing a training sample into i datablocks to acquire a first piece of training data to an i-th piece of training data, wherein i is a positive integer more than 1, distributing the first piece of training data to the i-th piece of training data to at least one information processing node which communicates with at least one server to perform computation of model parameters; and the information processing node is used for separately performing the computation of the model parameters according to the first piece of training data to the i-th piece of training data, working out the first model parameter to the i-th model parameter, updating the first model parameter to the i-th model parameter into the at least one server, accessing the at least one server, reading overall data formed by the first model parameter to the i-th model parameter, and acquiring a sample which is closest to respective training data in the information processing node according to the overall data to serve as an updated model parameter.

Description

technical field [0001] The present invention relates to information processing technology, in particular to an information processing method and system. Background technique [0002] With the development of Internet technology, communication technology, smart terminals and other fields, users' lives are becoming more and more intelligent. Intelligence relies on the accuracy of big data analysis and data mining. For example, video services, social advertising services, user portrait mining and other services that require accurate recommendations are inseparable from big data analysis and the accuracy of data mining. The purpose is to find out the information that users care about from the vast amount of data. In order to better achieve this goal, it is possible to design a machine learning platform and deploy large-scale machine learning tasks to create a system environment with a machine learning ecosystem that can support a larger order of magnitude of data and better perf...

AI Technical Summary

Problems solved by technology

However, with the existing clustering solutions, the random probability is too high, resulting in unsatisfactory clustering results and not meeting expectations

There is currently no effective solution to this problem

Images