Entity Recognition for Document Productivity

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Users of productivity applications face challenges in enhancing document content due to limited contextual awareness, often requiring external resources to supplement information, and existing data grids suffer from accuracy and quality issues due to human error and equipment limitations.

Innovation Solution

The system identifies entities within documents and surfaces additional relevant information from external data sources, allowing users to enhance entries by selecting and incorporating this information, and uses predictive probability distributions to suggest values and identify errors in data grids.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If users manually consult external resources to supplement document information, then document completeness improves, but time consumption increases

Engineering Contradiction:
Improvedocument completenessVSAvoidtime consumption
Core Design Contradiction:
Loss of informationVSLoss of time

Solution Approach 1:

The system performs preliminary actions by proactively identifying entities in the document and automatically retrieving related information from external data sources before the user needs it. The information is pre-processed and organized, ready for immediate presentation to the user, eliminating the need for manual resource consultation and significantly reducing time consumption while maintaining document completeness.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system enables self-service by automatically detecting when external information is needed and retrieving it without user intervention. The entity recognition and information retrieval processes operate autonomously, with the system serving itself by identifying knowledge gaps and filling them automatically, thus improving document completeness without increasing user time consumption.

Inventive Principle:
Principle #25Self-service

2Reliability

If automated spell-check and formula checking are implemented, then document accuracy improves, but system complexity increases

Engineering Contradiction:
Improvedocument accuracyVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system applies universality by using a single entity recognition and information retrieval module that serves multiple functions: identifying entities, retrieving related information, validating data accuracy, and supplementing document content. This multi-functional approach improves document accuracy through comprehensive checking while avoiding the complexity of separate specialized systems for each checking function.

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

3Manufacturing precision

If predictive probability distributions are used to suggest values for data grids, then data quality improves, but computational complexity increases

Engineering Contradiction:
Improvedata qualityVSAvoidcomputational complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The system applies partial action by using predictive probability distributions selectively only for cells that are blank or contain outliers, rather than processing every cell in the data grid. The inference engine identifies specific cells needing prediction and applies computational resources only where necessary, improving data quality for critical cells while limiting computational complexity to essential operations only.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3170096B1Entity recognition for enhanced document productivity
Publication Date: 2022.03.02 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3170096B1 patent drawingFigure 1
  • EP3170096B1 patent drawingFigure 2~3
  • EP3170096B1 patent drawingFigure 4

AI summary

Disclosed herein are systems, methods, and software for enhancing document productivity. In one implementation, various entries in a document are examined to identify at least an entry recognizable as an entity that is potentially related to at least one of various additional entities external to the document. At least a subset of the additional entities may be identified for surfacing in a user interface for potential inclusion in the document. In response to a selection of at least one of the subset of the additional entities, at least the one additional entity of the subset of the additional entities is included in the document in association with the entry.