Abstract: This application relates to a graph data processing method performed by a distributed computer node cluster including a plurality of computer devices, each computer device distributed on a respective computing node of the distributed computer node cluster, the method including: obtaining subgraph data divided from to-be-processed graph data; performing a computation task on the subgraph data to obtain corresponding global data and local data; writing the global data to a blockchain network, the global data of the blockchain network being updated by the distributed computing node cluster; obtaining latest global data from the blockchain network; and iteratively performing, according to the obtained latest global data and the local data, the computation task on the subgraph data without obtaining a computation result until an iteration stopping condition is met.

Abstract: A method and an apparatus to build a machine learning based network model are described. For example, processing circuitry of an information processing apparatus obtains a data processing procedure of a first network model and a reference dataset that is generated by the first network model in the data processing procedure. The data processing procedure includes a first data processing step. Further, the processing circuitry builds a first sub-network in a second network model of a neural network type. The second network model is the machine learning based network model to be built. The first sub-network performs the first data processing step. Then, the processing circuitry performs optimization training on the first sub-network by using the reference dataset.

Abstract: This application relates to a graph data processing method performed by a distributed computer node cluster including a plurality of computer devices, each computer device distributed on a respective computing node of the distributed computer node cluster, the method including: obtaining subgraph data divided from to-be-processed graph data; performing a computation task on the subgraph data to obtain corresponding global data and local data; writing the global data to a blockchain network, the global data of the blockchain network being updated by the distributed computing node cluster; obtaining latest global data from the blockchain network; and iteratively performing, according to the obtained latest global data and the local data, the computation task on the subgraph data without obtaining a computation result until an iteration stopping condition is met.

Abstract: A method for publishing multimedia data is disclosed. A user identifier of a user using a terminal and a target media position of the terminal are obtained. A plurality of multimedia data to be published is acquired. An outcome, associated with each of the plurality of multimedia data when the respective multimedia data is published at the target media position, is determined based on a set of characteristics. The set of characteristics includes at least one of a first credit characteristic, a second credit characteristic and a third credit characteristic, a user characteristic of the user identifier, a content characteristic of the respective multimedia data, and a position characteristic of the target media position. Based on the outcomes associated with the plurality of multimedia data, the target multimedia data is selected from the plurality of multimedia data, and sent to the terminal to be published at the target media position.

Abstract: A method and an apparatus to build a machine learning based network model are described. For example, processing circuitry of an information processing apparatus obtains a data processing procedure of a first network model and a reference dataset that is generated by the first network model in the data processing procedure. The data processing procedure includes a first data processing step. Further, the processing circuitry builds a first sub-network in a second network model of a neural network type. The second network model is the machine learning based network model to be built. The first sub-network performs the first data processing step. Then, the processing circuitry performs optimization training on the first sub-network by using the reference dataset.