Semantic Data Representation for Explainable Automated Reasoning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computer systems struggle to deeply understand and process the meaning of natural language due to the complexity and heterogeneity of data, leading to inefficiencies in applications that require broad and unstructured data, such as health management, nutrition tracking, and accounting, where conventional methods fail to represent semantic differences and require enormous schemas, making them impractical to build and code.
Innovation Solution
A machine-readable language, referred to as UL, is used to represent data in a structured and expressive format that allows for semantic nodes and links, enabling automated systems to process and understand a broad range of information, including user speech and text inputs, and facilitate faster response times through simplified processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional structured data schemas are used to represent broad heterogeneous data, then data coverage can be expanded, but system complexity and difficulty of maintenance grow more than linearly
Solution Approach 1:
The patent introduces an intermediary layer (semantic network/graph structure) between the raw heterogeneous data and the application logic. This intermediary uses standardized nodes and edges to represent concepts and relationships, allowing broad data coverage without proportional increases in schema complexity. The intermediary translates diverse data sources into a unified representation that applications can consume efficiently.
Solution Approach 2:
The patent creates a universal data representation framework that can handle multiple types of heterogeneous data (structured, unstructured, semi-structured) through a single flexible schema. The graph-based model uses universal node and edge concepts that can represent any entity and relationship type, eliminating the need for separate schemas for different data categories and reducing overall system complexity.
2Reliability
If hand-crafted schemas are used to cover all possible data types, then complete data representation is achieved, but building and coding becomes impractical
Solution Approach 1:
The patent enables the system to automatically generate and maintain data schemas through self-service mechanisms. The framework can autonomously discover data patterns, infer relationships, and create appropriate graph structures without requiring manual schema design for each new data type. This automation maintains complete data representation while making system implementation practical and scalable.
Solution Approach 2:
The patent implements dynamic schema evolution where the data representation structure adapts automatically to new data types and relationships. Rather than requiring static, pre-defined schemas that cover all possibilities, the system dynamically creates and modifies graph structures based on incoming data, maintaining completeness while keeping implementation feasible through on-demand schema generation.
3Ease of operation
If virtual ledgers with natural language names are used, then data storage is simplified, but semantic differences between transactions are lost
Solution Approach 1:
The patent segments the data representation into distinct components: nodes for entities, edges for relationships, and attributes for properties. This segmentation allows the system to maintain simple storage operations while preserving rich semantic information. Each transaction can be represented as a structured graph with explicit semantic relationships, avoiding the information loss that occurs with flat natural language labels.
Solution Approach 2:
The patent uses a composite data structure that combines simple storage-friendly formats with rich semantic representations. The graph model uses standardized node and edge types (analogous to composite materials) that provide both storage efficiency and semantic richness, allowing the system to achieve simplicity in operation without sacrificing semantic meaning.
Data Source
AI summary
There is provided a computer implemented method for the automated analysis or use of data, comprising the steps of: (a) storing in a non-transitory storage medium a structured, machine-readable representation of data that conforms to a machine-readable processable language, in which the structured, machine-readable representation of data includes reasoning passages, wherein the reasoning passages are represented in the processable language to represent semantics of reasoning steps; (b) automatically processing the structured, machine-readable representation of data, including processing at least some of the reasoning passages represented in the processable language to represent semantics of reasoning steps, to reason and to generate an explanation of the reasoning, and (c) storing a result of the reasoning and the explanation of the reasoning.