Associative Memory Update via Dual-Table Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current associative memory updating methods require recreating the entire memory system, which is time-consuming and resource-intensive, especially as the memory size increases, and do not effectively track changes in underlying data sources.
Innovation Solution
A system comprising an associative memory, a first table, a second table, a comparator, and an updater, where the associative memory ingests only updated data from the first table while avoiding the second table's fields, using the comparator to determine whether to ingest or forget data based on the second table's status indicators.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the associative memory is updated by recreating the entire memory using new versions of data sources, then the memory can be fully updated with current data, but the time and processing resources required increase significantly as memory size increases
Solution Approach 1:
The patent divides the update process into two independent tables: a first table containing data to be ingested and a second table containing data to be forgotten. This segmentation allows the system to process updates selectively rather than recreating the entire memory, thereby reducing update time while maintaining data freshness.
Solution Approach 2:
The patent extracts only the necessary data for updating from the complete data sources. By separating the update process into identifying which data to ingest (first table) and which data to forget (second table), the system processes only relevant portions of the data rather than the entire memory content.
2Productivity
If the associative memory is updated by treating updates as new data and ingesting them, then the memory can be updated efficiently, but inconsistencies arise when updates modify existing data rather than adding new data
Solution Approach 1:
The patent uses feedback mechanisms through the second table that tracks which data should be forgotten. The comparator references this feedback information to determine whether to ingest or forget specific data, ensuring consistent updates that reflect actual data changes rather than treating all updates as new data.
Solution Approach 2:
Instead of only adding new data to the memory, the patent inverts the approach by also removing outdated data through the second table. This inversion ensures that updates accurately represent changes in the underlying data sources, maintaining data consistency while preserving update efficiency.
3Reliability
If the system tracks all data changes in underlying sources, then the associative memory can be accurately updated, but the complexity of tracking and managing changes increases
Solution Approach 1:
The patent segments the tracking of data changes into two separate tables: the first table for data to be ingested and the second table for data to be forgotten. This segmentation simplifies the tracking process by organizing changes into distinct categories rather than managing all changes uniformly, thereby reducing complexity while maintaining update accuracy.
Solution Approach 2:
The patent introduces intermediary tables (first and second tables) that mediate between the underlying data sources and the associative memory. These intermediary structures simplify the tracking of changes by providing a structured way to manage which data should be ingested and which should be forgotten, reducing the complexity of direct tracking.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
A system (100) includes an associative memory (102), a first table (134), a second table (136), a comparator (164), and an updater (166). The associative memory (102) may include data and associations among data and may be built from the first table (134). The first table (134) may include a record (144) with a first and second field (150, 152). The associative memory may be configured to ingest the first field (150) and avoid ingesting the second field (152). The second table (136) may include a record (160) with a third field (162) storing information indicating whether the first field (150) has been ingested by the associative memory (102) or has been forgotten by the associative memory (102). The comparator (164) may be configured to compare the first and second table (134, 136) to identify one of whether the first field (150) should be forgotten or ingested by the associative memory (102). The updater (166) may be configured to update the associative memory (102) by performing one of ingesting or forgetting the first field (150).