Automated script generation and audio-visual presentations
Automatic generation of intelligent content is created using a system of computers including a user device and a cloud-based component that processes the user information. The system performs a process that includes receiving an input document and parsing the input document to generate inputs for a natural language generation model using a text analysis model. The natural language generation model generates one or more candidate presentation scripts based on the inputs. A presentation script is selected from the candidate presentation scripts and displayed. A text-to-speech model may be used to generate a synthesized audio presentation of the presentation script. A final presentation may be generated that includes a visual display of the input document and the corresponding audio presentation in sync with the visual display.
1. A computer-implemented method for automatically generating a presentation script, the method comprising:
receiving a presentation slide deck comprising a visual presentation;
determining a topic of the presentation slide deck;
generating a natural language generation model prompt for a first natural language generation model of a plurality of natural language generation models using an input design model, the generating comprising:
providing the presentation slide deck, the topic, and user information as first inputs to the input design model trained to:
select the first natural language generation model from the plurality of natural language generation models based on the first inputs, and
generate the natural language generation model prompt for the first natural language generation model based on the first inputs, wherein the natural language generation model prompt is designed to request a candidate presentation script for a speech on the topic, and
receiving the natural language generation model prompt as output from the input design model in response to the first inputs;
providing the natural language generation model prompt as an input prompt to the first natural language generation model trained to generate a natural language response to the input prompt;
receiving, in response to the input prompt, the candidate presentation script for the speech on the topic from the first natural language generation model;
generating an audio presentation by inputting the candidate presentation script into a text-to-speech model;
generating a final presentation comprising a visual display of the presentation slide deck synchronized with the audio presentation; and
outputting the final presentation to an audience.
2. The computer-implemented method of claim 1 , further comprising:
displaying one or more candidate presentation scripts comprising the candidate presentation script; and
receiving a selection of the candidate presentation script from the displayed one or more candidate presentation scripts.
3. The computer-implemented method of claim 1 , further comprising:
parsing the presentation slide deck to determine the topic.
4. The computer-implemented method of claim 1 , further comprising:
generating one or more candidate presentation scripts comprising the candidate presentation script;
ranking the one or more candidate presentation scripts with a ranking model; and
displaying the one or more candidate presentation scripts in ranked order.
5. The computer-implemented method of claim 4 , wherein each of the plurality of natural language generation models generates at least one of the one or more candidate presentation scripts.
6. The computer-implemented method of claim 1 , further comprising:
inputting a user voice into the text-to-speech model, wherein the audio presentation is generated using the user voice.
7. The computer-implemented method of claim 1 , further comprising:
receiving a request to modify an output language of the audio presentation in the final presentation to a requested language; and
translating the output language to the requested language in the final presentation.
8. The computer-implemented method of claim 1 , further comprising:
receiving feedback from the audience after outputting the final presentation to the audience; and
adjusting parameters of the input design model, the natural language generation model, or a combination of both based on the feedback.
9. A system comprising:
one or more processors; and
a memory having stored thereon instructions that, upon execution by the one or more processors, cause the one or more processors to:
receive a presentation slide deck comprising a visual presentation;
determine a topic of the presentation slide deck;
generate a natural language generation model prompt for a first natural language generation model of a plurality of natural language generation models using an input design model, the generating comprising:
provide the presentation slide deck, the topic, and user information as first inputs to the input design model trained to:
select the first natural language generation model from the plurality of natural language generation models based on the first inputs, and
generate the natural language generation model prompt for the first natural language generation model based on the first inputs,
wherein the natural language generation model prompt is designed to request a candidate presentation script for a speech on the topic, and
receive the natural language generation model prompt as output from the input design model in response to the first inputs;
provide the natural language generation model prompt as an input prompt to the first natural language generation model trained to generate a natural language response to the input prompt;
receive, in response to the input prompt, the candidate presentation script for the speech on the topic from the first natural language generation model;
generate an audio presentation by inputting the candidate presentation script into a text-to-speech model;
generate a final presentation comprising a visual display of the presentation slide deck synchronized with the audio presentation; and
output the final presentation to an audience.
10. The system of claim 9 , wherein the instructions comprise further instructions that, upon execution by the one or more processors, cause the one or more processors to:
display one or more candidate presentation scripts comprising the candidate presentation script; and
receive a selection of the candidate presentation script from the displayed one or more candidate presentation scripts.
11. The system of claim 9 , wherein the instructions comprise further instructions that, upon execution by the one or more processors, cause the one or more processors to:
parse the presentation slide deck to determine the topic.
12. The system of claim 9 , the instructions comprise further instructions that, upon execution by the one or more processors, cause the one or more processors to:
generate one or more candidate presentation scripts comprising the candidate presentation script;
rank the one or more candidate presentation scripts with a ranking model; and
display the one or more candidate presentation scripts in ranked order.
13. The system of claim 12 , wherein each of the plurality of natural language generation models generates at least one of the one or more candidate presentation scripts.
14. The system of claim 9 , wherein the instructions comprise further instructions that, upon execution by the one or more processors, cause the one or more processors to:
input a user voice into the text-to-speech model, wherein the audio presentation is generated using the user voice.
15. The system of claim 9 , wherein the instructions comprise further instructions that, upon execution by the one or more processors, cause the one or more processors to:
receive a request to modify an output language of the audio presentation in the final presentation to a requested language; and
translate the output language to the requested language in the final presentation.
16. The system of claim 9 , wherein the instructions comprise further instructions that, upon execution by the one or more processors, cause the one or more processors to:
receive feedback from the audience after outputting the final presentation to the audience; and
adjust parameters of the input design model, the natural language generation model, or a combination of both based on the feedback.