发明名称 |
Information processing system, network structure learning device, link strength prediction device, link strength prediction method and program |
摘要 |
An information processing system separately generates sample sequences from a posterior distribution of each random variable in a probability model representing the structure of a template network that serves as a template for a plurality of networks whose network structures are to be learned, and from a posterior distribution of each random variable in a probability model representing the structures of the plurality of networks, using learning data and hyperparameters relating to the plurality of networks. Next, the information processing system derives a predictive value of the strength of a link specified by an external variable based on the external variable and on the sample sequences. |
申请公布号 |
US9424527(B2) |
申请公布日期 |
2016.08.23 |
申请号 |
US201214128691 |
申请日期 |
2012.06.22 |
申请人 |
NEC Corporation |
发明人 |
Aoki Kenji |
分类号 |
G06N99/00;G06Q10/10;G06Q30/02 |
主分类号 |
G06N99/00 |
代理机构 |
Finnegan, Henderson, Farabow, Garrett & Dunner LLP |
代理人 |
Finnegan, Henderson, Farabow, Garrett & Dunner LLP |
主权项 |
1. An information processing system, comprising:
a memory that stores a set of instructions; and one or more processors configured to execute the set of instructions to:
receive learning data and hyperparameters related to a plurality of networks;generate a first sample sequence based on a posterior distribution of each random variable in a probability model representing a structure of a template network that serves as a template for a plurality of networks whose network structures are to be learned, and second sample sequences based on a posterior distribution of each random variable in probability models representing structures of the plurality of networks, using the learning data and the hyperparameters;output the first and second sample sequences;receive an external variable for a pair of network nodes between which a link strength is to be predicted, and the first and second sample sequences;derive a predictive value of the link strength based on the external variable and on the first and second sample sequences; andoutput the predictive value. |
地址 |
Tokyo JP |