训练未知意图检测模型的方法、未知意图检测方法及装置
By training an unknown intent detection model, constructing unknown intent samples using generative adversarial networks and dropout mechanisms, and combining them with historical context for detection, the accuracy problem of unknown intent detection in intelligent dialogue systems is solved, thus improving the user experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ALIBABA (CHINA) CO LTD
- Filing Date
- 2022-12-28
- Publication Date
- 2026-07-17
AI Technical Summary
Existing intelligent dialogue systems cannot effectively detect unknown intents, resulting in poor user experience and an inability to provide services that meet user needs. Current technologies mainly rely on the current statement for detection.
The unknown intent detection model is trained by acquiring training data, including multiple samples labeled with known intent and unknown intent, using a feature extraction network and a classification network for training, combining generative adversarial networks and dropout mechanism to construct unknown intent samples, and taking into account the historical context of the input text for detection.
It improves the accuracy of unknown intent detection, reduces manual costs, enhances the robustness of the model and its ability to detect unknown intent, and enables a better understanding of dialogue semantics.
Smart Images
Figure CN116186255B_ABST