aiqbench analysis
This study tests whether gains from graph neural networks in software architecture recovery come from the model or from simpler signals. It uses a Heterogeneous Graph Transformer on ten Java, C, and C++ systems and finds that dependency-aware representations and folder information help, while a plain directory rule reaches 89-100% median accuracy. For ML readers, the takeaway is a useful caution: graph architecture gains need baselines that match what representations already encode.
What Carries Architectural Signal? A Code-to-Architecture Mapping Study
arXiv:2610.41628v1 Announce Type: cross Abstract: Code-to-architecture mapping assigns source files to modules of a reference architecture. Representations of implementation artifacts can capture code text, dependencies, and directory organization, but predictive gains alone do not establish the contribution of additional graph processing. Dependencies may already be encoded in these representations, while directories may closely match the reference decomposition. We investigate how artifact representations and graph context affect mapping on ten Java, C, and C++ systems with a limited labeling budget. We compare lexical, structural, and contextual file representations and assess each predictor with and without dependency neighbors. Using a Heterogeneous Graph Transformer (HGT), we compare file-level graphs extended with member nodes summarizing each file's program elements, folder nodes, or both, alongside a variant encoding folder paths directly in file representations. We compare folder-aware predictions with a directory rule using the same labeled files. Representations encoding dependency information outperform lexical representations on most systems. Additional neighbor aggregation can improve mapping, depending on the predictor and aggregation direction. Folder-aware variants yield the largest and most consistent gains across the evaluated HGT configurations. However, the directory rule achieves per-system median accuracy of 89-100%, and adding folder information does not consistently improve predictions for files the rule misclassifies. Member nodes provide no consistent improvement over file-level graphs. These findings support assessing what file representations and directory organization already provide before attributing improvements to graph neural networks.
Title:What Carries Architectural Signal? A Code-to-Architecture Mapping Study
Abstract:Code-to-architecture mapping assigns source files to modules of a reference architecture. Representations of implementation artifacts can capture code text, dependencies, and directory organization, but predictive gains alone do not establish the contribution of additional graph processing. Dependencies may already be encoded in these representations, while directories may closely match the reference decomposition. We investigate how artifact representations and graph context affect mapping on ten Java, C, and C++ systems with a limited labeling budget. We compare lexical, structural, and contextual file representations and assess each predictor with and without dependency neighbors. Using a Heterogeneous Graph Transformer (HGT), we compare file-level graphs extended with member nodes summarizing each file's program elements, folder nodes, or both, alongside a variant encoding folder paths directly in file representations. We compare folder-aware predictions with a directory rule using the same labeled files. Representations encoding dependency information outperform lexical representations on most systems. Additional neighbor aggregation can improve mapping, depending on the predictor and aggregation direction. Folder-aware variants yield the largest and most consistent gains across the evaluated HGT configurations. However, the directory rule achieves per-system median accuracy of 89-100%, and adding folder information does not consistently improve predictions for files the rule misclassifies. Member nodes provide no consistent improvement over file-level graphs. These findings support assessing what file representations and directory organization already provide before attributing improvements to graph neural networks.
Software Engineering (cs.SE); Machine Learning (cs.LG)
arXiv:2610.41628 [cs.SE]
(or arXiv:2610.41628v1 [cs.SE] for this version)
https://doi.org/10.48550/arXiv.2610.41628 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 331 words