arXiv subjects

Proma Chowdhury

Publications and source records attributed to Proma Chowdhury.

2 recordsLinked to original sources

arXiv2026

Alteron: A Tool for Behavioral Regression Testing Across NLP Classifier Versions

Evaluating evolving Natural Language Processing (NLP) models is important for ensuring reliable behavior across updates, but standard benchmark metrics do not fully capture how model behavior changes across versions. Existing work has focused mainly on testing models in isolation rather than comparing successive versions in continuous integration workflows. We present Alteron, a tool for detecting behavioral regressions across NLP model versions with metamorphic testing. Alteron constructs a test corpus from labeled source examples and compares model versions on metamorphically transformed inputs. In an evaluation spanning 10 metamorphic relations (MRs), 4 model versions, and 3 model-update transitions, Alteron identified 16 behavioral regressions, 11 of which were release-blocking. The results show that common model updates can preserve overall task performance while still introducing undesirable behavior changes, and that behavioral checks across model versions can reveal failures that aggregate benchmark metrics alone do not capture. The tool is open-source and available at https://github.com/shazzad5709/alteron. A screencast demonstration is available at https://youtu.be/szwiWW5O4do.

cs.SE↗

arXiv2025

Towards Bridging Formal Methods and Human Interpretability

Labeled Transition Systems (LTS) are integral to model checking and design repair tools. System engineers frequently examine LTS designs during model checking or design repair to debug, identify inconsistencies, and validate system behavior. Despite LTS's significance, no prior research has examined human comprehension of these designs. To address this, we draw on traditional software engineering and graph theory, identifying 7 key metrics: cyclomatic complexity, state space size, average branching factor, maximum depth, Albin complexity, modularity, and redundancy. We created a dataset of 148 LTS designs, sampling 48 for 324 paired comparisons, and ranked them using the Bradley-Terry model. Through Kendall's Tau correlation analysis, we found that Albin complexity ($\tau = 0.444$), state space size ($\tau = 0.420$), cyclomatic complexity ($\tau = 0.366$), and redundancy ($\tau = 0.315$) most accurately reflect human comprehension of LTS designs. To showcase the metrics' utility, we applied the Albin complexity metric within the Fortis design repair tool, ranking system redesigns. This ranking reduced annotators' comprehension time by 39\%, suggesting that metrics emphasizing human factors can enhance formal design interpretability.

cs.SE↗