Neural Network Conversation Segment Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic document review tools face challenges in accurately reconstructing conversation threads from electronic communication documents due to variations in software clients and formats, leading to information loss and inefficiencies in the discovery process.
Innovation Solution
A computer-implemented method using an artificial neural network to identify conversation segments and their portions within electronic communication documents, which processes text-based content to generate position indicators for segment identification and determine ordered relationships between documents, enabling more accurate thread reconstruction across different software clients and formats.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional threading techniques are used to organize electronic communication documents, then documents can be arranged into conversation threads, but accuracy of thread reconstruction deteriorates due to variations in software clients and formats
Solution Approach 1:
The patent implements a universal threading system that can process electronic communication documents from multiple software clients and formats (Outlook, Lotus Notes, Gmail, etc.) using a single machine learning model. The system extracts features from diverse document formats and applies consistent threading logic, enabling one system to handle many different input formats without requiring client-specific processing logic.
Solution Approach 2:
The patent transforms the threading problem from exact matching of header fields to probabilistic pattern recognition using machine learning. By changing from deterministic parameter matching to statistical parameter analysis, the system can accommodate variations in software clients and formats while maintaining threading accuracy. The neural network learns to identify conversation segments despite differences in field names, positions, and formats across different email clients.
2Reliability
If strict segment identification is required for thread reconstruction, then threading accuracy may improve, but information loss increases when segment sections or fields cannot be fully identified
Solution Approach 1:
The patent applies partial action by allowing thread reconstruction to proceed with incomplete segment information. Instead of requiring all segment sections and fields to be perfectly identified, the system uses available features to probabilistically determine thread membership. Documents can be added to threads even if certain segment portions cannot be fully identified, reducing information loss while maintaining reasonable threading accuracy through the machine learning model's ability to work with partial data.
Solution Approach 2:
The patent introduces machine learning features as an intermediary between raw document segments and thread reconstruction. Rather than directly matching exact segment fields, the system extracts multiple features from segments (including partial matches) and uses these intermediate features to determine thread membership. This intermediary layer allows the system to bridge gaps in incomplete segment identification while still achieving accurate threading.
3Loss of information
If manual review of each electronic communication document is performed, then complete information extraction is achieved, but time consumption increases significantly
Solution Approach 1:
The patent implements self-service by enabling the system to automatically perform thread reconstruction and information extraction without manual review. The machine learning model autonomously processes electronic communication documents, identifies conversation segments, extracts relevant features, and organizes documents into threads. This automation eliminates the need for manual document analysis while maintaining high information extraction completeness, resolving the contradiction between thoroughness and time efficiency.
Solution Approach 2:
The patent replaces the mechanical process of manual document review with an automated machine learning system. Instead of human reviewers manually analyzing each document, the system uses neural networks to automatically extract information and reconstruct threads. This substitution of mechanical human labor with automated computational processes achieves both complete information extraction and time efficiency, as the machine learning model can process documents rapidly without sacrificing accuracy.
Data Source
AI summary
In a computer-implemented method, an artificial neural network is trained to identify portions of conversation segments within electronic communication documents, wherein an input layer of the artificial neural network includes a plurality of input parameters each corresponding to a different characteristic of text-based content. The method also includes receiving a first electronic communication document that includes first text-based content, and processing the first text-based content using the trained artificial neural network. Processing the first text-based content includes generating one or more position indicators for the first electronic communication document, and the one or more position indicators include one or more segment portion indicators denoting positions of one or more portions of one or more conversation segments within the first electronic communication document. The method also includes determining an ordered relationship between the first electronic communication document and one or more other electronic communication documents using the position indicator(s).