Neural Query Approximation With Exact Fallback for Large Data Sets

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing database optimization solutions, such as indexing and caching, fall short in providing timely insights from large and unknown data sets, as they are not programmed to process all types of data effectively, leading to prolonged query execution times.

Innovation Solution

A method and system utilizing a neural network to generate approximations of query results by sending queries to the network for predicted results, with real results provided when predictions are insufficiently accurate, allowing for faster response times and improved accuracy through multiple neural networks and local storage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If traditional database optimization methods (indexing, caching) are used, then query processing is accelerated for known data sets, but query execution time becomes excessively long for large and unknown data sets

Engineering Contradiction:

Improvequery execution speedVSAvoidquery completion time

Core Design Contradiction:

SpeedVSLoss of time

Solution Approach 1:

The patent introduces an intermediary system (approximation server with neural network) that sits between the query and the database. This intermediary provides approximate query results without requiring full database execution, thereby resolving the contradiction between speed and accuracy for unknown data sets.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system creates a neural network model that learns to replicate database query behavior. This copy (neural network approximation) can provide results instantly without executing actual database queries, solving the time loss problem while maintaining reasonable accuracy.

Inventive Principle:

Principle #26Copying

2Loss of time

If approximate results from neural networks are provided, then query response time is reduced, but result accuracy may be insufficient for certain queries

Engineering Contradiction:

Improvequery response timeVSAvoidquery result accuracy

Core Design Contradiction:

Loss of timeVSMeasurement precision

Solution Approach 1:

The system performs partial action by providing approximate results when sufficient and exact results when necessary. The approximation server determines whether to return neural network predictions or trigger full database execution based on query characteristics and accuracy requirements.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system implements feedback mechanisms where the accuracy of approximate results is evaluated, and when accuracy thresholds are not met, the system falls back to exact database execution. This feedback loop ensures accuracy requirements are satisfied while maximizing the use of fast approximation methods.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS11321320B2System and method for approximating query results using neural networks

Publication Date: 2022.05.03 SISENSE LTD

AI summary

A system and method for generating approximations of query results. The method includes sending a received query to a neural network, wherein the received query is executable on a target data set; receiving from the neural network a predicted result to the received query; providing the predicted result as a first output to a device having initiated the received query; determining a real result of the query from a data set stored in the database when the predicted result is insufficiently accurate; and providing the real result as a second output to a device having initiated the received query.