METHOD AND DEVICE FOR REACTION TYPE INFORMATION PROCESSING
PROBLEM TO BE SOLVED: To realize flexible learning and adaptation of high level and to facilitate constituting a system which has a high information operability and is compounded by autonomously changing information of a node while maintaining the operability of information. SOLUTION: An image react...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Patent |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | PROBLEM TO BE SOLVED: To realize flexible learning and adaptation of high level and to facilitate constituting a system which has a high information operability and is compounded by autonomously changing information of a node while maintaining the operability of information. SOLUTION: An image reactor 2 constituting an image storage network consists of a feature extraction module 2a and a reaction module 2b. Image are stored as unit information in nodes (a) and (b). Each node has the node state representing the state, and nodes (a) and (b) are connected by a link having weights, and the change of states of nodes (a) and (b) is propagated through a stimulus signal transmitted through the link. Recognition and learning are executed through the change of weights of nodes (a) and (b) as well as the change of information in nodes (a) and (b). Information in nodes (a) and (b) autonomously change by each other of nodes (a) and (b), and new information, nodes (a) and (b), and link are generated by mutual operation of information in plural nodes (a) and (b). New internal information is generated by the independent autonomous change of nodes (a) and (b). |
---|