Statistical Inspection of Information Objects for Component Decomposition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current reverse engineering and malware detection methods are inefficient and inaccurate, as they rely on detailed information about individual objects and heuristic decisions, and require frequent updates to detect evolving malware, especially in large populations of software applications and genomes.

Innovation Solution

A system and method for inspecting a population of information objects to identify and analyze symbols, constructing directed graphs to determine dependencies and hierarchical relationships between components, allowing for rapid decomposition of objects into logical components and prediction of object behavior.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If reverse engineering mechanisms examine individual objects in isolation using detailed information and heuristic decisions, then component decomposition can be performed, but the process becomes slow and inaccurate

Engineering Contradiction:
Improvecomponent decomposition accuracyVSAvoidanalysis speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent combines multiple objects into populations for batch analysis, grouping objects by shared symbols and analyzing them collectively rather than individually. This merging approach enables simultaneous processing of multiple objects, dramatically improving analysis speed while maintaining accuracy through statistical methods.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The system implements a universal analysis framework that handles diverse object types (software applications, genomes, etc.) using the same statistical population-based approach. By creating component models that can represent multiple object types, the system achieves both high accuracy and efficiency across different domains without requiring domain-specific heuristic rules for each object type.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Measurement precision

If malware detection relies on large databases with signatures of previously identified malware, then detection accuracy improves, but frequent database updates are required to keep abreast of evolving malware

Engineering Contradiction:
Improvemalware detection accuracyVSAvoidtime for database updates
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary analysis by examining populations of objects to identify shared symbols and construct component models in advance. By pre-processing and grouping objects by their symbolic content, the system creates a structured framework that enables rapid malware detection without requiring frequent updates of raw signature databases.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

Instead of storing and updating large databases of complete malware signatures, the system creates compressed component models that capture the essential structural characteristics of object populations. These component models serve as efficient representations that can be updated less frequently while still enabling accurate detection of new malware variants.

Inventive Principle:
Principle #26Copying

3Quantity of substance

If genome analysis identifies individual genes in sequences of DNA, then genetic components can be recognized, but the function of individual genes or their interactions remains unclear without significant research

Engineering Contradiction:
Improvenumber of identified genesVSAvoidfunctional information about genes
Core Design Contradiction:
Quantity of substanceVSLoss of information

Solution Approach 1:

The patent merges information from multiple genomes into populations and analyzes shared symbols to construct component models. By combining data across multiple genomes and examining which symbols (genes) appear together in populations, the system infers functional relationships and interactions among genes, recovering functional information that would be lost in individual gene analysis.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The system introduces component models as intermediary representations that capture the functional relationships between genes. These component models serve as mediators that encode not just individual gene identities but also their interactions and functional contexts, preserving functional information that would otherwise require extensive individual gene research.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS9069963B2Statistical inspection systems and methods for components and component relationships
Publication Date: 2015.06.30 RTX BBN TECH INC
  • US9069963B2 patent drawing
  • US9069963B2 patent drawing
  • US9069963B2 patent drawing

AI summary

Embodiments of an inspection system and method for a collection of information objects, for example, a collection of executable software applications may be inspected for computer viruses, or a collection of genomes may be inspected for common or unique gene sequences. Information objects may contain identified sequences of instructions, each of which may be labeled with a symbol. In the software context, programming languages may include symbols that indicate functionality. In some embodiments, an inspection of the statistical properties of the information objects and their included symbols may allow for the symbols (and thus instruction sequences) to be grouped into logical components. In some embodiments, objects that include individual logical components may be grouped together. These groupings and their dependencies may be used to determine the structure of each object by detailing its constituent components, how they relate or depend on one another, and how the information object may function.