System

A system comprising:

• circuitry configured to:
• receive, via a communication interface coupled to a packet-switched network, text data from a terminal device, apply a sentence importance scoring algorithm to extract a plurality of information units, assign quantified importance values and relevance values to each information unit, and classify the information units into section categories to generate structured data;
• reconstruct the structured data into a section structure conforming to a predefined document format, apply a machine learning model to at least one of the text data and interaction data to estimate an affective state value, generate a first prompt sentence encoding the section structure and a tone parameter derived from the affective state value, and supply the first prompt sentence to a generative neural network model to obtain draft body text;
• generate a second prompt sentence encoding the draft body text and layout constraint parameters, supply the second prompt sentence to the generative neural network model to obtain layout data, and apply a layout template to the layout data to generate a composite document structure;
• apply a convolutional neural network to candidate image data to extract visual feature vectors and generate image attribute data, compute image relevance scores by comparing scene category labels in the image attribute data with section category labels in the structured data, generate a third prompt sentence encoding filtered candidate images and document context data, supply the third prompt sentence to the generative neural network model to obtain image selection data, and insert selected images into the composite document structure;
• and
• generate output document data from the composite document structure and transmit the output document data to the terminal device via the communication interface.

A method comprising:

• receiving, via a communication interface coupled to a packet-switched network, text data from a terminal device, and applying a computational analysis algorithm to the text data to extract a plurality of information units and assign quantified importance values and relevance values to each information unit;
• classifying the information units based on the quantified importance values and relevance values to generate structured data, and algorithmically reconstructing the structured data into a section structure and heading structure conforming to a predefined document format;
• generating a first prompt sentence encoding the structured data and supplying the first prompt sentence to a generative neural network model to obtain draft body text, and generating a second prompt sentence encoding the draft body text and layout constraint parameters to obtain layout data specifying a visual arrangement of textual regions and image regions;
• applying a layout template to the layout data to generate a composite document structure, applying an image recognition algorithm to candidate image data to extract image attribute data, generating a third prompt sentence encoding the image attribute data and document context data, supplying the third prompt sentence to the generative neural network model to obtain image selection data, and inserting selected images into the composite document structure at positions specified by the layout data;
• and
• generating output document data from the composite document structure and transmitting the output document data to the terminal device via the communication interface.