LLM Field Object Generation for Accurate Fillable Documents

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing digital form filling systems suffer from inaccuracies due to generalized machine learning models that hallucinate and require excessive user interactions, leading to inefficiencies and resource wastage.

Innovation Solution

A field object generation system utilizing a large language model to determine and combine relevant data from source content items with fillable digital documents, reducing hallucinations by using a summarization and aggregation validation approach and minimizing user interactions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If generalized machine learning models are used to generate content for fillable fields, then broad coverage of output generation is achieved, but accuracy deteriorates due to hallucinations

Engineering Contradiction:

Improvebroad coverage of output generationVSAvoidaccuracy of generated content

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent introduces an intermediary validation system that acts as a mediator between the generalized machine learning model and the final output. This validation system checks generated content against the source document structure and predefined field requirements, filtering out hallucinated content while preserving accurate generated content, thus maintaining both broad coverage and high accuracy

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent implements a feedback mechanism where the system validates generated content by comparing it with the source document structure and field definitions. When inaccuracies are detected, the system adjusts its generation process accordingly, creating a closed-loop system that continuously improves accuracy while maintaining versatility

Inventive Principle:

Principle #23Feedback

2Adaptability or versatility

If multiple interfaces and applications are used to input digital content into fillable documents, then comprehensive data input capability is achieved, but navigational efficiency deteriorates

Engineering Contradiction:

Improvedata input capabilityVSAvoidnavigational efficiency

Core Design Contradiction:

Adaptability or versatilityVSEase of operation

Solution Approach 1:

The patent merges multiple interfaces and applications into a single integrated interface that provides comprehensive data input capability. This unified interface combines source document viewing, field identification, content generation, and validation functions, eliminating the need to navigate between multiple applications while maintaining full functionality

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent creates a universal interface that performs multiple functions: viewing source documents, identifying fillable fields, generating content, validating accuracy, and populating fields. This multi-functional interface replaces the need for separate specialized applications, improving navigational efficiency without sacrificing capability

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Productivity

If multiple applications run simultaneously during data input stages, then comprehensive processing capability is achieved, but resource consumption increases

Engineering Contradiction:

Improveprocessing capabilityVSAvoidcomputer resource expenditure

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent combines multiple processing functions that previously required separate applications into a single integrated system. By consolidating source document processing, field identification, content generation, and validation into one application, the system maintains comprehensive processing capability while significantly reducing the number of simultaneously running processes and associated resource consumption

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS20260044671A1Generating field objects for auto-populating fillable documents utilizing a large language model

Publication Date: 2026.02.12 DROPBOX INC

AI summary

The present disclosure relates to systems, non-transitory computer-readable media, and methods for modifying a fillable digital document. In particular, the disclosed systems can receive a user interaction requesting to populate one or more aggregated data fields in a fillable digital document. In response to the request, the field object generation system can determine the data relevant to one or more aggregated data fields in the fillable digital document by utilizing a large language model to process one or more source content items for a user account. Further the systems and generate a field object from the data relevant to one or more aggregated data field and modify the fillable digital document by including the field object in the fillable digital document.