Regionally Additive Models: Explainable-by-design models minimizing feature interactions (2309.12215v1)
Abstract: Generalized Additive Models (GAMs) are widely used explainable-by-design models in various applications. GAMs assume that the output can be represented as a sum of univariate functions, referred to as components. However, this assumption fails in ML problems where the output depends on multiple features simultaneously. In these cases, GAMs fail to capture the interaction terms of the underlying function, leading to subpar accuracy. To (partially) address this issue, we propose Regionally Additive Models (RAMs), a novel class of explainable-by-design models. RAMs identify subregions within the feature space where interactions are minimized. Within these regions, it is more accurate to express the output as a sum of univariate functions (components). Consequently, RAMs fit one component per subregion of each feature instead of one component per feature. This approach yields a more expressive model compared to GAMs while retaining interpretability. The RAM framework consists of three steps. Firstly, we train a black-box model. Secondly, using Regional Effect Plots, we identify subregions where the black-box model exhibits near-local additivity. Lastly, we fit a GAM component for each identified subregion. We validate the effectiveness of RAMs through experiments on both synthetic and real-world datasets. The results confirm that RAMs offer improved expressiveness compared to GAMs while maintaining interpretability.
- Neural additive models: Interpretable machine learning with neural nets. Advances in Neural Information Processing Systems, 34:4699–4711, 2021.
- Visualizing the effects of predictor variables in black box supervised learning models. Journal of the Royal Statistical Society Series B: Statistical Methodology, 82(4):1059–1086, 2020.
- Tabnet: Attentive interpretable tabular learning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 6679–6687, 2021.
- Node-gam: Neural generalized additive model for interpretable deep learning. arXiv preprint arXiv:2106.01613, 2021.
- Sparse interaction additive networks via feature interaction detection and sparse selection. In Alice H. Oh, Alekh Agarwal, Danielle Belgrave, and Kyunghyun Cho, editors, Advances in Neural Information Processing Systems, 2022. URL https://openreview.net/forum?id=Q6DJ12oQjrp.
- Hadi Fanaee-T. Bike Sharing Dataset. UCI Machine Learning Repository, 2013. DOI: https://doi.org/10.24432/C5W894.
- The false hope of current approaches to explainable artificial intelligence in health care. The Lancet Digital Health, 3(11):e745–e750, 2021.
- Dale: Differential accumulated local effects for efficient and accurate global explanations. In Asian Conference on Machine Learning, pages 375–390. PMLR, 2023a.
- Rhale: Robust and heterogeneity-aware accumulated local effects. In European Conferenece in AI (ECAI), 2023b.
- Generalized additive models: some applications. Journal of the American Statistical Association, 82(398):371–386, 1987.
- Repid: Regional effect plots with implicit interaction detection. In International Conference on Artificial Intelligence and Statistics, pages 10209–10233. PMLR, 2022.
- Decomposing global feature effects based on feature interactions, 2023.
- Accurate intelligible models with pairwise interactions. In Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 623–631, 2013.
- Sparse spatial autoregressions. Statistics & Probability Letters, 33(3):291–297, 1997.
- Cynthia Rudin. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nature machine intelligence, 1(5):206–215, 2019.
- Sparse neural additive model: Interpretable deep learning with feature selection via group sparsity. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, pages 343–359. Springer, 2023.