Purpose of this paper is to propose approach and technical infrastructure for improvement of inter-organizational networks' response in product information acquisition and management. Different approaches (industrial categorization schemes, functional decomposition and semantic web) for management of product information are analyzed in context of inter-organizational networks. Process for semantic alignment of product information is defined, resulting with generalized, two-dimensional model, consisting of design and functional perspective. The process is expected to decrease human intervention in product data exchange in networked environments, as well as to create added value, through possible recognition of design intent, automated referencing to related manufacturing competences and reuse potential. Current prototype of system comprises of product ontologies and interfaces for topological model submission and refinement by using lexical term and predicate matching and property transfer. Impact of using formalized functional perspective is only theoretically justified and it still needs to be verified.
A Deep Semantic Alignment Network for Cross-Modal Image-Text Retrieval in Remote Sensing
