Intuitive defect prevention with swarm learning intelligence over blockchain network
Aspects of the disclosure relate to s computing system that is configured to use heuristic and/or metaheuristic algorithms based on swarm learning (SL) intelligence frameworks and combine SL with blockchain and edge computing frameworks to provide a technologically efficient, responsive, and/or adaptable solution to detecting and preventing defects in software applications.
1 . A swarm learning (SL) computing system that categorizes application defects using a plurality of nodes communicatively coupled by a peer-to-peer communications network, the SL computing system comprising:
a swarm of boids comprising at least a first quality assurance (QA) agent and a second QA agent, wherein the swarm of boids correspond to the plurality of nodes;
a blockchain comprising a plurality of blocks corresponding to a plurality of software application failure logs and defects data, wherein the plurality of software application failure logs indicates events associated with a plurality of software applications that failed, and wherein the defects data corresponds to what caused the plurality of software applications to fail;
a particle swarm optimization (PSO) model configured to categorize the defects data as either a false positive defect or a true positive defect,
wherein the PSO model comprises a field based, gene expression programming algorithm, and wherein the gene expression programming algorithm adjusts a crossover rate, a mutation rate, or a population size based upon the field variation improving a defect prediction accuracy;
wherein the PSO model is configured to test the plurality of software applications,
wherein the PSO model is configured to receive training data to cause generation of the swarm of boids, wherein the swarm of boids correspond to the plurality of application failure logs and the defects data,
wherein the swarm of boids are configured to calculate a first grouping of software application failure logs that comprise false positive defects and a second grouping of software application failure logs that comprise true positive defects,
wherein parameters of the PSO model are adjusted by one or more metaheuristic algorithms,
wherein a defect prevention framework is generated by the PSO, and wherein the defect prevention network provides a guide to a software developer with a solution or a fix to true positive defects; and
an edge node of the plurality of nodes, wherein the edge node comprises a personal computing device corresponding to the software developer, wherein based on the true positive defects that are prioritized depending on a severity of the true positive defects, the edge node is configured to receive a feedback loop from the PSO model.
2 . The SL computing system of claim 1 , wherein the feedback loop from the PSO model to the edge node is configured to prevent a future defect.
3 . The SL computing system of claim 1 , wherein the training data comprises data at least extracted from the blockchain, and wherein the blockchain comprises a distributed ledger system.
4 . The SL computing system of claim 1 , wherein the metaheuristic algorithms enhance accuracy and relevance of data-driven insights.
5 . The SL computing system of claim 1 , wherein the PSO model further comprises an ant colony optimization (ACO).
6 . The SL computing system of claim 1 , wherein the receiving into the PSO model occurs asynchronously and non-sequentially for the plurality of software application logs and the defects data.
7 . The SL computing system of claim 1 , wherein the blockchain stores immutable records of at least defect lifecycle events comprising: detection, classification, assignment, and resolution.
8 . The SL computing system of claim 1 , wherein the swarm of boids uses pattern identification by learning from historical data from the blockchain.
9 . The SL computing system of claim 1 , wherein the first QA agent is configured to autonomously adhere to code scanning rules and interactions with other agents in the swarm of boids without a centralized server dictating which parts of the plurality of software applications to analyze by each QA agent.
10 . The SL computing system of claim 1 , wherein the feedback loop from the PSO model comprises embeddings for use with one or more machine learning models.
11 . The SL computing system of claim 1 , wherein the swarm of boids are configured to self-organize into groups for optimal division of labor and adaptation.
12 . A method for categorizing application defects using a plurality of nodes communicatively coupled by a peer-to-peer communications network, wherein the plurality of nodes corresponds to a swarm of boids, the method comprising:
(a) accessing a blockchain, which comprises a plurality of blocks corresponding to a plurality of software application failure logs indicating events associated with software application failures and defects data that caused the software application failures, to extract, as training data, the one or more application failure logs and the defects data from the plurality of blocks;
(b) categorizing, by a particle swarm optimization (PSO) model, the defects data as either a false positive defect or a true positive defect, wherein the PSO model comprises a field based, gene expression programming algorithm, and wherein the gene expression programming algorithm adjusts a crossover rate, a mutation rate, or a population size based upon the field variation improving a defect prediction accuracy;
(c) testing, by the PSO model, the plurality of software applications;
(d) receiving training data to cause generation of the swarm of boids, wherein the swarm of boids correspond to the plurality of application failure logs and the defects data,
(e) calculating a first grouping of software application failure logs that comprise false positive defects and a second grouping of software application failure logs that comprise true positive defects;
(f) generating a defect prevention framework, wherein the defect prevention network provides a guide to a software developer with a solution or a fix to true positive defects;
(g) adjusting, by one or more metaheuristic algorithms, one or more parameters of the PSO model; and
(h) receiving, by an edge node of the plurality of nodes, a feedback loop from the PSO model.
13 . The method of claim 12 , wherein the feedback loop from the PSO model to the edge node prevents a future defect in the plurality of software applications.
14 . The method of claim 12 , wherein the PSO model further comprises an ant colony optimization (ACO).
15 . The method of claim 12 , wherein the swarm of boids are configured to self-organize into groups for optimal division of labor and adaptation.
16 . A non-transitory computer-readable medium storing instruction that, when executed by one or more computer processors, cause the one or more computer processors to perform a method comprising steps of:
(a) accessing a blockchain, which comprises a plurality of blocks corresponding to a plurality of software application failure logs indicating events associated with software application failures and defects data that caused the software application failures, to extract, as training data, the one or more application failure logs and the defects data from the plurality of blocks;
(b) categorizing, by a particle swarm optimization (PSO) model, the defects data as either a false positive defect or a true positive defect, wherein the PSO model comprises a field based, gene expression programming algorithm, and wherein the gene expression programming algorithm adjusts a crossover rate, a mutation rate, or a population size based upon the field variation improving a defect prediction accuracy;
(c) testing, by the PSO model, the plurality of software applications;
(d) receiving training data to cause generation of a swarm of boids, wherein the swarm of boids correspond to the plurality of application failure logs and the defects data;
(e) calculating a first grouping of software application failure logs that comprise false positive defects and a second grouping of software application failure logs that comprise true positive defects;
(f) generating a defect prevention framework, wherein the defect prevention network provides a guide to a software developer with a solution or a fix to true positive defects;
(g) adjusting, by one or more metaheuristic algorithms, one or more parameters of the PSO model; and
(h) receiving, by an edge node of the swarm of boids, a feedback loop from the PSO model.
17 . The non-transitory computer-readable medium of claim 16 , wherein the feedback loop from the PSO model comprises embeddings for use with one or more machine learning models.