Emotion recognition method, emotion recognition model pre-training method and electronic equipment
Through large-scale model technology and emotional joint modeling method, the problem of low accuracy in data emotion recognition in the prior art is solved, and higher accuracy and cross-task generalization capabilities are achieved.
CN119646219APending Publication Date: 2025-03-18HANGZHOU ALICLOUD FEITIAN INFORMATION TECH CO LTD
0 Cites 0 Cited by
Patent Information
- Application Number
- CN202311125828.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-09-01
- Publication Date
- 2025-03-18
AI Technical Summary
Technical Problem
The accuracy of emotional identification of data in the prior art is low, especially when processing multiple data modes.
Method used
The large model technology is adopted to process single-modal and multimodal data through emotional joint modeling methods, and task-specific prompt methods are introduced, combining new pre-training tasks and training strategies to improve the multimodal emotional perception ability of the model.
Benefits of technology
It has achieved improved the accuracy of data emotion recognition, enhanced the model's cross-task generalization ability, and can effectively handle multiple emotion recognition tasks.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention discloses an emotion recognition method, a pre-training method of an emotion recognition model and electronic equipment, and relates to the field of large model technologies and data recognition. The method comprises the following steps: responding to query data received in a generative dialogue interface, the query data comprising data of at least one of the following modalities: a visual modality, an acoustic modality, a text modality and a context modality; emotion recognition is conducted on the query data, an initial emotion recognition set is obtained, and the initial emotion recognition set comprises emotion recognition results corresponding to a plurality of emotion recognition tasks; and based on the target emotion recognition task corresponding to the query data, determining a target emotion recognition result corresponding to the query data. According to the method and the device, the technical problem of relatively low accuracy of emotion recognition on data in related technologies is solved.
Need to check novelty before this filing date? Find Prior Art