Distributed computing system and method and storage medium

Abstract

The invention discloses a distributed computing system and method and a storage medium. The distributed computing system includes at least two computing nodes and at least two parameter service nodes.A computing node initializes vectors of corresponding users in a user matrix according to the users, which are included by a subset of training data, to obtain a user sub-matrix composed of the initialized vectors; the computing node carries out iterative computing on the user sub-matrix and an item sub-matrix according to the subset of the training data and the item sub-matrix obtained from a parameter service node, and transmits an item sub-matrix, which is obtained after each time of iterative computing, to the corresponding parameter service node; the parameter service node initializes vectors, which correspond to partial items, to obtain the item sub-matrix composed of the initialized vectors; and the parameter service node updates an item sub-matrix, which is stored by the parameterservice node, according to the item sub-matrix transmitted by the computing node. By implementing the system, computing tasks can be completed in an intensive manner.

Description

technical field [0001] The present invention relates to computer technology, in particular to a distributed computing system, method and storage medium. Background technique [0002] Artificial intelligence has developed rapidly and is widely used in various industries. Taking the application scenario of product recommendation as an example, according to the user's behavior data, the machine learning method is used to train a model that predicts the user's ratings for different products, so that the ranking of the user's ratings for different products can be calculated, and the products with high ratings can be selected. Recommending to users can help users quickly locate the products they are interested in, and realize accurate and efficient product marketing. [0003] For example, the current commodity recommendation relies on big data processing technology, and it is necessary to analyze and process the massive collected behavioral data in order to train a model with sco...

AI Technical Summary

Problems solved by technology

[0004] However, the resources of a single node in the computing system provided by related technologies are limited, and the upgrade of the computing system often lags behind. The contradiction between the current situation of limited resources of a single node and the high resource overhead required for model training calculations has become difficult to solve. technical problem

Images