Natural-Language Retrieval of Siloed Data in Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data silos in manufacturing industries create barriers to information sharing and collaboration across departments due to fragmented data views and inconsistencies, making it difficult for leaders to obtain a holistic view of company data.
Innovation Solution
A computer-implemented method and system that facilitates fetching and using siloed data by receiving natural language input, extracting fetch parameters, and storing data sets from separate repositories into a data store, utilizing natural language processing and configurable parameter types to ensure accurate data retrieval and storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If data is stored in separate repositories for different departments, then data accessibility for specific teams is improved, but information sharing and collaboration across departments deteriorates
Solution Approach 1:
The patent introduces a data federation layer that acts as an intermediary between siloed repositories and users. This layer provides a unified virtual data model that mediates access to distributed data sources, enabling cross-departmental information sharing while maintaining the structural independence of original repositories. The federation layer translates diverse data formats and access patterns into consistent queries, resolving the contradiction between localized accessibility and global sharing.
Solution Approach 2:
The patent implements a universal data access interface that serves multiple functions: it provides department-specific data access, enables cross-departmental queries, maintains data consistency, and supports various user roles. This multi-functional system allows the same infrastructure to simultaneously preserve data sovereignty and facilitate information sharing, eliminating the need to choose between localized accessibility and organizational-wide collaboration.
2Reliability
If data is stored in separate repositories, then data quality for specific departments is maintained, but overall data quality and consistency deteriorates
Solution Approach 1:
The patent implements feedback mechanisms through data quality monitoring and validation layers that continuously assess data consistency across repositories. The system detects quality issues, triggers alerts, and enables automated corrections by propagating changes across the federation. This feedback loop maintains overall data consistency without requiring centralized control, allowing each department to manage local quality while the system ensures global coherence.
Solution Approach 2:
The patent changes the fundamental parameters of data access from physical storage locations to virtual data views. By transforming data from static repository files to dynamic federation objects, the system enables consistent data representation across departments while maintaining local storage independence. This parameter transformation allows data to be both locally optimized and globally consistent through virtual layering.
3Loss of information
If manual data fetching from multiple repositories is performed, then data can be retrieved, but time consumption and operational complexity increases
Solution Approach 1:
The patent segments data access into two independent layers: physical data storage in distributed repositories and virtual data presentation through the federation layer. This segmentation allows the system to maintain complex data retrieval capabilities while simplifying user interaction. Users query the unified virtual model rather than manually accessing multiple repositories, reducing time consumption while preserving comprehensive data access.
Solution Approach 2:
The patent creates virtual copies of data across the federation layer that reflect the state of underlying repositories without physically replicating the data storage. This virtual copying mechanism enables fast data retrieval through the unified interface while avoiding the time-consuming manual process of accessing multiple sources. The virtual layer caches and pre-processes data, providing instant access to information that would otherwise require sequential queries across repositories.
4Loss of information
If centralized data management is implemented, then information sharing and collaboration are improved, but system complexity and loss of autonomy deteriorates
Solution Approach 1:
The patent adds a virtual dimension to the existing physical data architecture. Instead of managing complexity within a single centralized system, the federation layer operates as an additional dimensional space that abstracts the complexity of distributed repositories. This dimensional change allows centralized information sharing capabilities to coexist with distributed data storage, reducing overall system complexity by organizing complexity into separate layers rather than consolidating it.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for fetching and using siloed data in an automation environment. To facilitate fetching and using siloed data in an automation environment, the following steps are suggested: • receiving natural language input (120A) indicative of fetching data sets (122A, 122B) stored in at least two separate repositories (124A, 124B); • extracting at least two fetch parameters (126A, 126B) according to configurable parameter types (128A, 128B) from the received input (120A); • fetching the data sets (122A, 122B) corresponding to the respective, extracted fetch parameter (126A, 126B) from repositories (124A, 124B); and • storing the respective, fetched data set (122A, 122B) in a data storage (130).