17897887. DETECTING OUT-OF-DOMAIN TEXT DATA IN DIALOG SYSTEMS USING ARTIFICIAL INTELLIGENCE simplified abstract (International Business Machines Corporation)
DETECTING OUT-OF-DOMAIN TEXT DATA IN DIALOG SYSTEMS USING ARTIFICIAL INTELLIGENCE
Organization Name
International Business Machines Corporation
Inventor(s)
Saloni Potdar of Cambridge MA (US)
Ladislav Kunc of Cambridge MA (US)
DETECTING OUT-OF-DOMAIN TEXT DATA IN DIALOG SYSTEMS USING ARTIFICIAL INTELLIGENCE - A simplified explanation of the abstract
This abstract first appeared for US patent application 17897887 titled 'DETECTING OUT-OF-DOMAIN TEXT DATA IN DIALOG SYSTEMS USING ARTIFICIAL INTELLIGENCE
Simplified Explanation
The patent application describes methods, systems, and computer program products for detecting out-of-domain text data in dialog systems using artificial intelligence techniques.
- Updating artificial intelligence techniques for out-of-domain text data detection based on encoding training data and generating regularized representations.
- Encoding input text data.
- Computing out-of-domain scores for the encoded input text data using updated artificial intelligence techniques.
- Performing automated actions based on the computed out-of-domain scores.
Potential Applications
- Improving the accuracy and efficiency of dialog systems.
- Enhancing user experience in chatbots and virtual assistants.
- Streamlining customer service interactions.
Problems Solved
- Identifying and handling out-of-domain text data in dialog systems.
- Ensuring that responses are relevant and appropriate.
- Minimizing errors and misunderstandings in communication.
Benefits
- Increased accuracy in detecting out-of-domain text data.
- Enhanced performance of dialog systems.
- Improved user satisfaction and engagement.
Original Abstract Submitted
Methods, systems, and computer program products for detecting out-of-domain text data in dialog systems using artificial intelligence techniques are provided herein. A computer-implemented method includes updating artificial intelligence techniques related to out-of-domain text data detection, the updating based on encoding training data and generating regularized representations of at least a portion of the encoded training data by combining the at least a portion of the encoded training data and at least one intent centroid associated with the updated artificial intelligence techniques; encoding input text data; computing out-of-domain scores, in connection with the at least one dialog system, for at least a portion of the encoded input text data by processing the at least a portion of encoded input data using at least a portion of the one or more updated artificial intelligence techniques; and performing one or more automated actions based on the computed out-of-domain scores.
