Semantic behavior conflict-driven social robot adaptive detection method, electronic equipment and storage medium

Abstract

The invention provides a semantic behavior conflict-driven social robot adaptive detection method, electronic equipment and a storage medium, and the method comprises the steps: obtaining text data, behavior time sequence data and behavior sequence derived features based on the text content of a to-be-detected account; constructing a joint detection network model comprising a text semantic expert module, a behavior time sequence expert module, a conflict calculation unit, a decision strategy module and a multipath output module, and defining the joint detection network model as a semantic behavior conflict driven social robot detection model; obtaining a time sequence representation and a time sequence detection output probability; determining a decision path based on a semantic-behavior conflict index, and outputting a decision path standard error; and outputting a social robot detection result corresponding to the to-be-detected account according to the decision path. Detection is completed only based on the text + behavior time sequence, the method does not depend on social relation graph / adjacency information, the data acquisition and deployment complexity is reduced, and the applicability under the scene that relation data cannot be obtained or is limited is improved.

Images