ProtoMind: Modeling Driven NAS and SIP Message Sequence Modeling for Smart Regression Detection
Year: 2025
Type: conference-paper
Abstract: Large language models (LLMs) face significant challenges in effectively detecting regressions in software systems, particularly in the context of dynamic environments. The need for an efficient framework that can adaptively learn and model architectural and sequential data is evident. We present ProtoMind, a unique approach that combines Modeling Driven Network Attached Storage (NAS) with Session Initiation Protocol (SIP) message sequence modeling to enhance regression detection. The NAS component autonomously discovers the most effective neural architectures specifically tailored to regression tasks, leading to notable improvements in performance. Concurrently, the SIP message sequence modeling captures complex temporal patterns and interactions, facilitating a more accurate identification of regression faults. Experiments validate that ProtoMind surpasses conventional methods in regression detection, exhibiting enhanced accuracy and expedited processing times.
Author: Tongwei Tu
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4414693127