Lean Classifier Models Using Boosted Decision Stumps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile devices face performance degradation and power utilization issues due to complex software and hardware interactions, which existing solutions struggle to efficiently identify and address, especially in resource-constrained environments where computationally intensive processes are not feasible.
Innovation Solution
The method involves generating lean classifier models in mobile devices using boosted decision stumps, which are derived from a full classifier model created by a network server, allowing for efficient classification of device behaviors without excessive resource consumption. This is achieved by converting a finite state machine into a list of boosted decision stumps, selecting a subset of test conditions, and applying them to classify behaviors as benign or malicious.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a full classifier model is used to accurately classify mobile device behaviors, then classification accuracy is improved, but processing power, memory usage, and energy consumption increase excessively
Solution Approach 1:
The full classifier model is segmented into multiple decision stumps, each evaluating a single feature. This segmentation allows the system to selectively apply only necessary decision stumps based on the specific behavior being classified, reducing overall computational load and energy consumption while maintaining classification accuracy.
Solution Approach 2:
The patent extracts and retains only the most relevant decision stumps from the full classifier model to form a lean classifier model. By removing unnecessary components and keeping only essential feature evaluations, the system achieves accurate behavior classification with significantly reduced processing requirements and energy usage.
2Reliability
If computationally intensive classification processes are applied to identify malicious behaviors, then detection accuracy is improved, but device performance and power utilization deteriorate
Solution Approach 1:
The system applies partial action by using only the necessary subset of decision stumps required for accurate behavior classification rather than applying the complete full classifier model. This selective approach maintains detection accuracy for malicious behaviors while avoiding excessive computational processing that would degrade device performance.
3Adaptability or versatility
If a comprehensive classifier model is deployed on mobile devices, then behavior classification capability is improved, but memory usage and processing overhead increase
Solution Approach 1:
The comprehensive classifier model is divided into discrete, independent decision stumps that can be selectively applied. This segmentation reduces model complexity by allowing the system to load and execute only the specific decision stumps needed for particular classification tasks, rather than maintaining the entire complex model in memory.
Solution Approach 2:
The patent changes the parameter of model representation from a monolithic full classifier to a collection of lightweight decision stumps. This parameter change enables the system to achieve comprehensive behavior classification capability while keeping individual model components simple and memory-efficient.
Data Source
AI summary
Methods and systems for classifying mobile device behavior include configuring a server use a large corpus of mobile device behaviors to generate a full classifier model that includes a finite state machine suitable for conversion into boosted decision stumps and/or which describes all or many of the features relevant to determining whether a mobile device behavior is benign or contributing to the mobile device's degradation over time. A mobile device may receive the full classifier model and use the model to generate a full set of boosted decision stumps from which a more focused or lean classifier model is generated by culling the full set to a subset suitable for efficiently determining whether mobile device behavior are benign. Boosted decision stumps may be culled by selecting all boosted decision stumps that depend upon a limited set of test conditions.