A Study of the Learnability of Relational Properties: Model Counting Meets Machine Learning (MCML)
This work addresses the challenge of accurately quantifying ML model performance for relational properties in software engineering, though it is incremental as it applies existing model counting techniques to a new evaluation context.
The paper tackles the problem of evaluating the learnability of relational properties in software design by introducing the MCML approach, which uses model counting to assess ML model performance across entire bounded input spaces rather than just datasets. The results show that while simple models achieve high accuracy on standard datasets, their performance degrades significantly when evaluated against the full input space, revealing the complexity of precisely learning these properties.
This paper introduces the MCML approach for empirically studying the learnability of relational properties that can be expressed in the well-known software design language Alloy. A key novelty of MCML is quantification of the performance of and semantic differences among trained machine learning (ML) models, specifically decision trees, with respect to entire (bounded) input spaces, and not just for given training and test datasets (as is the common practice). MCML reduces the quantification problems to the classic complexity theory problem of model counting, and employs state-of-the-art model counters. The results show that relatively simple ML models can achieve surprisingly high performance (accuracy and F1-score) when evaluated in the common setting of using training and test datasets - even when the training dataset is much smaller than the test dataset - indicating the seeming simplicity of learning relational properties. However, MCML metrics based on model counting show that the performance can degrade substantially when tested against the entire (bounded) input space, indicating the high complexity of precisely learning these properties, and the usefulness of model counting in quantifying the true performance.