Is this project an undergraduate, graduate, or faculty project?
Undergraduate
Project Type
group
Campus
Daytona Beach
Authors' Class Standing
Caleb Stone, Senior
Lead Presenter's Name
Caleb Stone
Lead Presenter's College
DB College of Arts and Sciences
Faculty Mentor Name
Dr. Vidhyashree Nagaraju
Abstract
Ensuring the reliability of software intensive and safety critical systems is a persistent challenge across aerospace, defense, transportation, and other mis- sion focused domains. Traditional software relia- bility growth models (SRGM) provide useful quanti- tative insight into defect discovery trends, but they rely mostly only on numerical failure data and do not use the rich contextual information contained in test logs, anomaly reports, and engineering notes. This paper presents a hybrid framework that com- bines semantic features extracted by a large lan- guage model (LLM) with a non-homogeneous Pois- son process (NHPP) based software reliability growth model. The LLM analyzes unstructured engineering text to identify meaningful attributes such as fault severity, subsystem impact, fault class, recurrence indicators, test phase references, integration points (changepoints), fix complexity, and any other indi- cators of defect difficulty or operational relevance. These attributes are then incorporated as covariates into the NHPP SRGM, which produces reliability pre- dictions that are more sensitive to system context and that improve interpretability for engineering de- cision support. Using synthetic but representative defect data, the study demonstrates improvements in predictive accuracy, earlier identification of risk prone fault classes, and clearer visibility into changes across testing phases. The results show that integrat- ing LLM enabled semantic analysis with classical reli- ability modeling can strengthen verification and val- idation workflows and can enhance reliability assess- ment for safety critical software systems.
Did this research project receive funding support (Spark, SURF, Research Abroad, Student Internal Grants, Collaborative, Climbing, or Ignite Grants) from the Office of Undergraduate Research?
No
Included in
Artificial Intelligence and Robotics Commons, Risk Analysis Commons, Software Engineering Commons
A Hybrid LLM-SRGM Framework for AI-Enabled Reliability Assessment in Safety-Critical Software Systems
Ensuring the reliability of software intensive and safety critical systems is a persistent challenge across aerospace, defense, transportation, and other mis- sion focused domains. Traditional software relia- bility growth models (SRGM) provide useful quanti- tative insight into defect discovery trends, but they rely mostly only on numerical failure data and do not use the rich contextual information contained in test logs, anomaly reports, and engineering notes. This paper presents a hybrid framework that com- bines semantic features extracted by a large lan- guage model (LLM) with a non-homogeneous Pois- son process (NHPP) based software reliability growth model. The LLM analyzes unstructured engineering text to identify meaningful attributes such as fault severity, subsystem impact, fault class, recurrence indicators, test phase references, integration points (changepoints), fix complexity, and any other indi- cators of defect difficulty or operational relevance. These attributes are then incorporated as covariates into the NHPP SRGM, which produces reliability pre- dictions that are more sensitive to system context and that improve interpretability for engineering de- cision support. Using synthetic but representative defect data, the study demonstrates improvements in predictive accuracy, earlier identification of risk prone fault classes, and clearer visibility into changes across testing phases. The results show that integrat- ing LLM enabled semantic analysis with classical reli- ability modeling can strengthen verification and val- idation workflows and can enhance reliability assess- ment for safety critical software systems.