초록 ▼

Content generation leverages an unsupervised, generative pre-trained language model (e.g., a generative-AI). In this approach, a model derived by applying to given content relevant competitive content and one or more optimization targets is received. Based on optimization criteria encoded as embeddi

Content generation leverages an unsupervised, generative pre-trained language model (e.g., a generative-AI). In this approach, a model derived by applying to given content relevant competitive content and one or more optimization targets is received. Based on optimization criteria encoded as embedding signals in the model, a determination is made regarding whether a template suitable for use as an input to the generative-AI exists in a set of templates. If so, the model embedding signals are merged into the template, or the template itself is transformed using the embedding signals, in either case creating a modified template. If, however, no template suitable as the input exists, the model and other information are input to a natural language processor to generate a generative-AI input. Either the modified template or the generative-AI input, as the case may be, is then applied through the generative-AI to generate an output competitively-optimized with respect to the optimization targets.

대표청구항 ▼

1. A computer program product comprising a non-transitory computer-readable medium, the computer program product comprising program code executable in a hardware processor for content generation, the program code comprising computer program instructions configured to: receive a target content derive

1. A computer program product comprising a non-transitory computer-readable medium, the computer program product comprising program code executable in a hardware processor for content generation, the program code comprising computer program instructions configured to: receive a target content derived model, the target content derived model having been derived by applying to given content (i) search engine-indexed content, and (ii) one or more optimization targets, the search engine-indexed content including content portions in which the given content is expressed;based on one or more optimization criteria encoded as embedding signals in the target content derived model, identify a template suitable for use as an input to a generative-AI exists in the set;create a modified template by merging the target content derived model embedding signals into the template or transforming the template using the embedding signals; andapply the modified template through the generative-AI to generate an output that is competitively-optimized with respect to the one or more optimization targets, wherein the output comprises generative-AI-generated text. 2. The computer program product as described in claim 1 wherein the given content is one of: a user prompt, and a user prompt together with a knowledge graph of general facts. 3. The computer program product as described in claim 1 wherein the one or more optimization targets are: search engine optimization (SEO), semantic relevance, accuracy, engagement, emotion, conversion, style, tone and voice. 4. The computer program product as described in claim 1 wherein the generative-AI is an unsupervised, generative pre-trained language model. 5. The computer program product as described in claim 4 wherein the language model is GPT-3. 6. The computer program product as described in claim 1 wherein the target content derived model includes topically-relevant content as identified from a user prompt, or from a user prompt together with a knowledge graph of general facts, and wherein a user prompt is one of: a key phrase, and a URL. 7. The computer program product as described in claim 1 wherein the program code further includes computer program instructions configured to transform the target content derived model into a knowledge graph, and to process the knowledge graph to generate one or more statements that are applied through the generative-AI. 8. The computer program product as described in claim 1 wherein the program code further includes computer program instructions configured to post-process the output. 9. The computer program product as described in claim 8 wherein post-processing includes fact-checking. 10. The computer program product as described in claim 1 wherein the target content derived model is a knowledge graph. 11. The computer program product as described in claim 1 wherein n number of optimization targets are used and the target content derived model is an n-dimensional target content derived model. 12. The computer program product as described in claim 1 wherein the computer program instructions configured to apply the given content further include computer program instructions configured to process a content model derived from the given content to a set of content models derived from the relevant competitive content to generate a target optimization score. 13. The computer program product as described in claim 12 wherein the output is competitively-optimized by a given competitive performance metric. 14. A computer program product comprising a non-transitory computer-readable medium, the computer program product comprising program code executable in a hardware processor for content generation, the program code comprising computer program instructions configured to: receive a target content derived model, the target content derived model having been derived by applying to given content (i) search engine-indexed content, and (ii) one or more optimization targets, the search engine-indexed content including content portions in which the given content is expressed;based on one or more optimization criteria encoded as embedding signals in the target content derived model, identify a template suitable for use as an input to a generative-AI exists in the set;apply the target content derived model and other information as input to a natural language processor (NLP) algorithm to generate an generative-AI input; andapply the generative-AI input through the generative-AI to generate an output that is competitively-optimized with respect to the one or more optimization targets. 15. The computer program product as described in claim 14 wherein the computer program instructions are further configured to score the output with respect to the target content derived model. 16. The computer program product as described in claim 15 wherein the computer program instructions are further configured to train the NLP algorithm using a result of the scoring.