Explainable Text Summarization with Segment Contribution Tracking

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current text summarization systems lack transparency, as users have no insight into which portions of the original document were used to generate the summary or whether the summary covers the entire document, due to their reliance on deep learning models with hidden decision-making processes.

Innovation Solution

A device receives relevancy parameters from a user interface to segment a text document into relevant portions, generates a summary using these segments, and provides the summary along with an indication of each segment's contribution to the user interface.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If deep learning models with hidden layers are used for text summarization, then summarization capability is improved, but transparency and user insight into the decision-making process deteriorates

Engineering Contradiction:

Improvesummarization capabilityVSAvoidtransparency of decision-making process

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

The patent introduces an intermediary component that tracks and records the decision-making process of the deep learning model. This intermediary captures which portions of the input text were selected for inclusion in the summary and why, making the hidden decision-making process visible to users without interfering with the model's summarization capability.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements feedback by providing users with information about which text portions were used to generate the summary and the rationale behind the model's decisions. This feedback loop allows users to understand and verify the summarization process while maintaining the efficiency of deep learning models.

Inventive Principle:

Principle #23Feedback

2Productivity

If deep learning models are used for text summarization, then summarization efficiency is improved, but user control and insight into document coverage deteriorates

Engineering Contradiction:

Improvesummarization efficiencyVSAvoiduser control over summarization

Core Design Contradiction:

ProductivityVSEase of operation

Solution Approach 1:

The system provides feedback to users by displaying which portions of the document were selected for the summary and allowing users to review and adjust these selections. This maintains summarization efficiency while giving users control over the final output.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system allows users to self-adjust the summary by providing them with the ability to review selected portions and make modifications based on their own judgment of what should be included or excluded, combining automated efficiency with user control.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20250077562A1Creating explainable summaries of text documents with controllable levels of detail

Publication Date: 2025.03.06 CISCO TECHNOLOGY INC

AI summary

In one implementation, a device receives relevancy parameters from a user interface indicative of how relevant different portions of a text document are to a user. The device segments the text document into segments based on the relevancy parameters. The device generates a summary of the text document using the segments. The device provides, to the user interface, the summary of the text document and an indication of how much each of the segments contributed to the summary.