Alphanumeric Character Vectors for Reliable LLM Action Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing software development systems lack intuitive and reliable methods for selecting appropriate large language models (LLMs) and validating their outputs, leading to inefficiencies, security risks, and inconsistencies in compliance with evolving regulatory requirements.

Innovation Solution

A data generation platform that dynamically evaluates machine learning prompts, selects models based on resource availability, and validates outputs in an isolated environment using generative AI models to ensure secure, reliable, and modular LLM usage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If organizations manually process regulatory guidelines to generate actionable items, then compliance accuracy can be maintained, but the process becomes time-consuming and cannot keep pace with evolving regulations

Engineering Contradiction:

Improvespeed of generating actionable compliance itemsVSAvoidaccuracy of compliance interpretation

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system creates a virtual copy of the regulatory guideline in the form of a structured data model that captures all actionable items. This digital replica can be processed, searched, and analyzed without affecting the original document, enabling rapid generation of compliance actions while maintaining interpretive accuracy through the structured representation.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces manual mechanical processing of regulatory text with an automated computer system that uses natural language processing and data modeling techniques. This substitution eliminates human processing time constraints while maintaining through algorithmic analysis the accuracy needed for reliable compliance interpretation.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If organizations implement comprehensive compliance programs, then regulatory adherence improves, but operational complexity and resource requirements increase

Engineering Contradiction:

Improvecompliance adherenceVSAvoidcomplexity of compliance management system

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system segments the comprehensive compliance program into discrete, manageable components by extracting individual actionable items from the regulatory guideline. Each actionable item becomes a separate, identifiable unit that can be independently tracked and managed, reducing the perceived complexity while maintaining comprehensive coverage.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent creates a universal data model structure that can handle multiple types of regulatory guidelines and compliance scenarios within a single system framework. This multi-functional approach allows the same system to manage diverse compliance requirements without proportionally increasing complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Productivity

If organizations use existing LLMs for compliance processing, then productivity increases, but security risks and output validation challenges arise

Engineering Contradiction:

Improveefficiency of compliance processingVSAvoidsecurity risks from LLM outputs

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The patent introduces a structured data model as an intermediary layer between the LLM and the compliance processing system. This intermediary validates and structures the LLM's output before it is used for compliance actions, filtering out potential security risks while maintaining the productivity benefits of LLM processing.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements feedback mechanisms that continuously monitor and validate LLM outputs against the structured data model. This feedback loop detects and corrects potential security issues or inaccuracies in real-time, enabling safe use of LLMs for compliance processing while maintaining high productivity.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20250322153A1Identifying and analyzing actions from vector representations of alphanumeric characters using a large language model

Publication Date: 2025.10.16 CITIBANK N A

AI summary

The systems and methods disclosed herein receive an output generation request from that includes input for generating an output using a language model. The input includes a set of alphanumeric characters associated with operative standards for a first set of actions. The system divides the set of alphanumeric characters into text subsets. For each text subset, a vector representation is determined. Prompts are created for each vector representation including the set of alphanumeric characters, query contexts, keywords, and/or the text subset. Each vector representation's prompt is input into the language model, which generates a second set of actions of related actions, where subsequently generated actions are based on prior generated actions. The system aggregates the second set of actions into a third set of actions and displays a graphical layout. The graphical layout displays a representation of the set of alphanumeric characters and the corresponding actions.