Machine content generation
OwnerBao Traninventor
Life of this patent
Estimating the term…
Abstract
Computerized systems and methods are disclosed to generate a document from one or more first and second text prompts, generating one or more context-sensitive text suggestions using a transformer with an encoder on the text prompts and a decoder that produces a text expansion to provide the context-sensitive text suggestions based on the one or more first and second text prompts by applying generative artificial intelligence with token biased weights for zero-shot, one-shot or some-shot generation of the artificial intelligence context-sensitive text suggestions from the one or more first and second text prompts.
Claims31
A method for generating content using artificial intelligence (AI), comprising:
- providing an interface to receive one or more multimodal prompts including text, structured data, file, image, audio, sound, or video;
- applying a transformer model with an encoder to encode the one or more prompts;
- using a decoder to produce a multimodal expansion based on the encoded one or more prompts;
- generating one or more content suggestions using the multimodal expansion;
- applying generative artificial intelligence with token biased weights to the multimodal suggestions;
- and
- outputting artificial intelligence generated content based on the one or more prompts.
A method, comprising:
- generating one or more images or videos with a multimedia structure having one or more seed landmark images therein, the landmark image including a milestone overview and a plurality of components;
- from the milestone overview, generating one or more computer-generated image suggestions to supplement the landmark image;
- combining the milestone overview with each component and generating one or more computer-generated component image suggestions;
- and
- creating the content by combining the landmark image and the one or more computer-generated image suggestions.
A method to respond to a verbal query with a chatbot, comprising:
- providing the chatbot trained with context sensitive data to generate a realistic human voice response by applying a transformer model with an encoder to encode the prompt, using a decoder to produce a multimodal expansion based on the encoded prompt, generating a suggestion using the multimodal expansion;
- and applying generative artificial intelligence with token biased weights to the multimodal suggestions;
- and
- using the chatbot in place of a human agent to respond to the query, wherein the chat bot detects an emotion and upon detecting a predetermined emotion, adjusting the human voice response based on the detected emotion or referring the query to one or more AI agents trained on responding to the predetermined emotion.
A method to summarize a multi-modal document, comprising:
- applying a transformer model with an encoder to encode the multi-modal document;
- using a decoder to produce a multimodal expansion based on the encoded multi-modal document;
- and
- applying generative artificial intelligence with token biased weights to the multimodal document;
- and generating a summary of the multi-modal document.
A method for generating code using artificial intelligence (AI), comprising:
- providing an interface to receive one or more prompts including text, structured data, file, image, audio, sound, or video;
- applying a transformer model with an encoder to encode the one or more prompts;
- using a decoder to produce a code expansion based on the encoded one or more prompts;
- generating one or more code suggestions;
- applying generative artificial intelligence with token biased weights to the code suggestions;
- and
- outputting artificial intelligence generated code.
Dependencies are inferred from the claim text.