Dynamic Access Control in Question Answer Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current question and answer systems lack the ability to provide differential access to users, relying on manual and time-consuming file-space permissions and static user profiles to manage access levels, which is inefficient for scenarios requiring varying levels of information access.
Innovation Solution
A computer-implemented method and system that performs content-based determination of access permissions by receiving an information access policy, classifying document access levels, and adjusting answers based on user permission levels, allowing for flexible and dynamic access control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual file-space permissions and static user profiles are used to manage access levels, then access control can be implemented, but administrative time and complexity increase significantly
Solution Approach 1:
The system enables self-service access control by allowing documents to automatically classify their own access levels through metadata analysis and machine learning models, eliminating the need for manual administrative tagging and reducing administrative time while maintaining reliable access control
Solution Approach 2:
The system performs preliminary classification of documents during the ingestion process, assigning access levels before users query the system. This advance preparation reduces the time required for access control decisions during actual use, as the classification work is already completed
2Reliability
If manual tagging of documents with access level metadata is performed, then document-level access control can be achieved, but the process becomes time-consuming and inefficient
Solution Approach 1:
Documents automatically perform self-classification by having their metadata analyzed by machine learning models that determine appropriate access levels. This self-service approach eliminates manual tagging entirely, maintaining document-level access control while dramatically improving productivity by removing the time-consuming manual process
Solution Approach 2:
The manual mechanical process of administrators reading and tagging documents is replaced with an automated electronic system using machine learning models that analyze document metadata and automatically assign access levels, thereby maintaining control reliability while improving tagging efficiency
3Adaptability or versatility
If the same corpus content is used for all users, then system simplicity is maintained, but differential access requirements cannot be met
Solution Approach 1:
The system applies local quality by allowing different portions of the corpus to have different access characteristics based on their metadata classification. Each document or section can have its own access level assigned automatically, enabling differential access capability while managing complexity through automated rather than manual processes
Solution Approach 2:
The system changes the parameter of access level assignment from a static global setting to a dynamic per-document attribute determined by machine learning analysis of metadata. This enables differential access capability while the automated nature of the process prevents system complexity from becoming unmanageable
Data Source
AI summary
Embodiments are directed to a question and answer (QA) pipeline system that adjusts answers to input questions based on a user criteria, thus implementing a content-based determination of access permissions. The QA system allows for information to be retrieved based on permission granted to a user. Documents are ingested and assigned an access level based on a defined information access policy. The QA system is implemented with the defined information access policy, the ingested documents, and the inferred access levels. For the QA system implementation, a user enters a question; primary search and answer extraction stages are performed; candidate answer extraction is performed using only content the user is allowed to access; the candidate answers are scored, ranked, and merged; ranked answers based on user permissions are filtered; and answers are provided to the user.