Systems and methods for continual updating of response generation by an artificial intelligence chatbot
Inventors
Ma, Weicheng • Cao, Kai • Pan, Bei • Chen, Lin • Li, Xiang
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11823061-B2
Publication Date
2023-11-21
Expiration Date
Abstract
Methods and systems are provided for a natural language processing system comprising a chatbot adapted for dialog generation. In one example, the system may include a combination of a variational autoencoder (VAE) and a generative adversarial network (GAN) for generating natural responses to input queries. The VAE may convert queries into vector embeddings that may then be used by the GAN to continuously update and improve responses provided by the chatbot.
Core Innovation
The invention relates to a chatbot that performs dialog generation using a Variational-sequence-to-sequence dialog generator (VSDG) that includes a variational autoencoder (VAE) combined with a generative adversarial network (GAN). User queries are preprocessed to form a processed query and sent to the VSDG, and the chatbot outputs a response at a response dialog box in a user interface displayed at a display device.
In response to a generated response not satisfying the user query, the chatbot outputs additional responses configured to focus a subsequent response towards a final response that satisfies the user query. The chatbot continues to output additional responses until a final response satisfying the user query is generated, using the VSDG in the dialog flow.
For a healthcare website implementation, the invention provides a chatbot with a user interface device having at least one of a query text box and a response dialog box, and a processor communicatively coupled to the user interface device. The processor is configured with the VSDG stored in non-transitory memory and instructions stored in the non-transitory memory that cause the processor to receive a user query, send it to the VSDG, output a response, and continue outputting additional responses until a final response satisfying the user query is generated.
Claims Coverage
The partial content provides two independent claims, with inventive features centered on a VSDG (VAE + GAN) for dialog generation and a continuing additional-response mechanism to reach a final response satisfying the user query.
Variational-sequence-to-sequence dialog generator with VAE and GAN
A processed query is sent to a Variational-sequence-to-sequence dialog generator (VSDG) that includes a variational autoencoder (VAE) and a generative adversarial network (GAN), and a response is output from the VSDG.
Additional responses steered to a final satisfying response
Responsive to the response not satisfying the user query, the chatbot outputs additional responses configured to focus a subsequent response towards a final response that satisfies the user query, including continuing to output additional responses until a final response satisfying the user query is generated.
Healthcare website chatbot user interface and processor using stored VSDG
A chatbot for a healthcare website includes a user interface device with at least one of a query text box and a response dialog box, and a processor configured with a VSDG that includes a VAE combined with a GAN stored in non-transitory memory and instructions causing the processor to receive a user query, send it to the VSDG, output a response, and continue outputting additional responses until a final response satisfying the user query is generated.
Across the independent claims, the coverage centers on dialog generation using a VSDG combining a VAE and a GAN, together with a loop-like additional-response mechanism that is responsive to dissatisfaction and continues until a final response satisfies the user query. One independent claim grounds this architecture specifically in a healthcare website chatbot with a query text box/response dialog box interface and a processor configured with the VSDG stored in non-transitory memory.
Stated Advantages
Not explicitly described in patent.
Documented Applications
A healthcare website chatbot example use case.
Interested in licensing this patent?