Accurate intelligible models with pairwise interactions
The DOI matches its Crossref record; 21 of the 22 references it cites were found in a registry or an index.
Abstract
Standard generalized additive models (GAMs) usually model the dependent variable as a sum of univariate models. Although previous studies have shown that standard GAMs can be interpreted by users, their accuracy is significantly less than more complex models that permit interactions.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“It is designed to have high accuracy compared to the state-of-the-art machine learning models while keeping the intelligibility, and explainability [13].”
From An Empirical Comparison of Interpretable Models to Post-Hoc Explanations (2023) · cites this paper as [13]
Citations
Cites
Greedy function approximation: A gradient boosting machine
Friedman, J. H. (2001) · The Annals of Statistics · doi:10.1214/aos/1013203451 · Cited by 30,731
Generalized Additive Models: An Introduction with R
Wood, S. N. (2017) · doi:10.1201/9781315370279 · Cited by 13,852
Introduction to information retrieval
(2009) · Choice Reviews Online · doi:10.5860/choice.46-2715 · Cited by 12,278
Brown, R. A., Hastie, T. J., & Tibshirani, R. J. (1991) · Biometrics · doi:10.2307/2532174 · Cited by 8,176
An introduction to variable and feature selection
GuyonIsabelle, & ElisseeffAndré (2003) · Journal of Machine Learning Research · doi:10.5555/944919.944968 · Cited by 7,715
Koyak, R. A., Hastie, T. J., & Tibshirani, R. J. (1991) · Journal of the American Statistical Association · doi:10.2307/2290538 · Cited by 7,596
Wood, S. N. (2006) · doi:10.1201/9781420010404 · Cited by 4,953
Hastie, T., & Tibshirani, R. (2014) · Wiley StatsRef: Statistics Reference Online · doi:10.1002/9781118445112.stat03141 · Cited by 3,156
An Empirical Comparison of Voting Classification Algorithms: Bagging, Boosting, and Variants
Bauer, E., & Kohavi, R. (1999) · Machine Learning · doi:10.1023/a:1007515423169 · Cited by 2,638
Wood, S. N. (2003) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/1467-9868.00374 · Cited by 2,445
Research works printed but not linked: 2
- 10.5555/1394399 · doi:10.5555/1394399
- 10.5555/1622620.1622633 · doi:10.5555/1622620.1622633
Cited by
Items 501–565 of 565
Explainable underwater target recognition models: principles, methods, and applications
Xu, T., Jia, H., & Qin, J. (2025) · Frontiers in Physics · doi:10.3389/fphy.2025.1682253 · Cited by 3
“The first category is model transparency, the opposing attribute of “black-box” systems, encompassing model-level explainability such as simulatability [4], parameter decomposability at the single-component level [5], and training algorithm transparency [6].”
Predicting blood pressure variability in hemodialysis using an explainable boosting machine model
Huang, W.-Y., Lin, C.-J., Zheng, Y.-X., Chen, H.-T., Sun, F., Wu, C., & Wu, K. (2025) · Clinical Kidney Journal · doi:10.1093/ckj/sfaf349
“In order to overcome the limitations of previous studies, we employed the Explainable Boosting Machine ( EBM) algorithm [23 ], which can capture non-linear relationships while maintaining interpretability through its additive structure.”
Localized Additive Models for Interpretable High-Dimensional Predictions
Guru, D., Subramaniam, S. R., Patel, R. S., & Prabhakaran, P. (2025) · International Conference on Information Technology and Computing (ICITCOM) · doi:10.1109/icitcom66635.2025.11265625
“Among these, Generalized Additive Models with Interactions (GA 2 Ms) [6] were pioneering, incorporating second-order components to extend the formulation as: This extension preserves interpretability since the interaction terms can be visualized using two-dimensional plots alongside the…”
XAI in Wireless Communications: A Case Study on Interpretable 5G Performance Analysis
Şahin, A. F., Güven, Y., Başaran, S. T., & Kumbasar, T. (2025) · IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC) · doi:10.1109/pimrc62392.2025.11275455 · Cited by 3
“Aiming to provide an accurate yet interpretable model, GAMs plus Interactions (GA 2 M) have been proposed in [12].”
Explainable Gradient-boosting for Explainability in Energy Consumption Management
Ramos, D., Faria, P., & Vale, Z. (2025) · IEEE PES Innovative Smart Grid Technologies Europe · doi:10.1109/isgteurope64741.2025.11305666
“The use of pairwise interactions has been considered in [14] to rank all possible pairs of features as candidates for inclusion into standard GAMs (generalized additive model).”
Synthetic Disease Trajectories for Predicting Chronic Kidney Disease Post-diabetes Onset
de la Oliva, V., Kreil, D. P., Dopazo, J., Ortuño, F. M., & Loucera, C. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-10661-2_47
Zhan, M. (2025) · International Conference on Smart Healthcare and Wearable Intelligent Devices (SHWID) · doi:10.1145/3788112.3788141
“Intrinsically interpretable models such as Generalized Additive Models (GAM) [25] and its variant GA 2 M [26] are noted for their transparent structures that do not necessitate additional interpretation tools.”
Explanations Go Linear: Post-Hoc Explainability for Tabular Data with Interpretable Meta-Encoding
Piaggesi, S., Guidotti, R., Giannotti, F., & Pedreschi, D. (2025) · Proceedings · doi:10.1109/icdm65498.2025.00074 · Cited by 4
“NAMs define the mapping as 663 2025 IEEE International Conference on Data Mining (ICDM) a sum of non-linear functions NN j ( x j ) , for each feature j , possibly including feature interactions NN jj ′ ( x j , x j ′ ) [25].”
The most important features in generalized additive models might be groups of features
Bosschieter, T. M., França, L., Wolk, J., Wu, Y., Mehta, B. Y., Dehoney, J., Kiss, O., Baker, F. C., Zhao, Q., Caruana, R., et al. (2026) · Scientific Reports · doi:10.1038/s41598-026-43928-4
“Explainable Boosting Machines (EBMs) 2 are Generalized Additive Models (GAMs) 3 with interaction terms, i.e., The shape functions f i ( · ) and f i , j ( · , · ) are trained using cyclic gradient boosting with shallow learners 2 .”
InfluentialSelf-citationPredicting self‐reported injury status among runners training for the New York City Marathon
Fontana, M. A., Egbert, J. S., & Toresdahl, B. G. (2026) · PM&R · doi:10.1002/pmrj.70127
Human-Supervised AI Control Architectures for Accountable and Transparent Enterprise Data Governance
Yamsani, N. (2024) · International Journal of Scientific Research in Science Engineering and Technology · doi:10.32628/ijsrset24118051
Interpretable machine learning for CLAS12 data analysis
Cherrier, N. (2021) · theses.fr (ABES) · doi:10.70675/171a54daz808cz478az83bcz0276fd7612c3 · Cited by 5
Lam, Q. T., Fan, F.-Y., Wang, Y., Wu, C.-Y., Sun, Y.-S., Vo, T. T. T., Kuo, H., Kha, Q. H., Minh, L. H. N., Vu, G. T., et al. (2026) · medRxiv · doi:10.64898/2026.04.03.26350106
A Neural-Symbolic learning framework to produce interpretable predictions for image classification
Bennetot, A. (2022) · theses.fr (ABES) · doi:10.70675/2b59f598z2028z49afz8a20z29b21f19f394 · Cited by 1
Trustworthy machine learning : explainability and distribution-free uncertainty quantification
Amoukou, S. I. (2023) · theses.fr (ABES) · doi:10.70675/a022ddcbz8047z488bzb33bz68f6b99639b2 · Cited by 6
Interpretable Music Recommender Systems
Afchar, D. (2023) · theses.fr (ABES) · doi:10.70675/acfb6676z8613z4e08zad14z98d4a9c2ee74 · Cited by 3
Cuthbertson, R., Hancock, O. E., & Vuksanović, V. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.04.24.715834
“In addition, an Explainable Boosting Machine (EBM) – an interpretable additive model that represents predictions as a sum of feature-specific effects – was used on the test dataset to quantify feature importance in terms of their age-dependent contribution profiles (25).”
Interpretable AI and Classical Statistics: Revisiting the Surrogate Model Paradigm
Abdelhay, O., Altamimi, T., Shatnawi, A., & Abdelhay, H. (2026) · International Conference on Advances in Artificial Intelligence and Machine Learning (AAIML) · doi:10.1109/aaiml67890.2026.11498135
“For instance, utilizing Generalized Additive Models with pairwise interactions (GA2M), which are inherently interpretable (they generate risk contributions for each feature) and are often as accurate as black-box models [37, 38].”
Explainability-Guided Credit Risk Prediction Using Machine Learning Model
Bhandari, R., Ramya, S., & Ebenezar, U. S. (2026) · International Conference on Innovative Trends in Information Technology (ICITIIT) · doi:10.1109/icitiit68860.2026.11499474 · Cited by 1
ERP-RiskBench: Leakage-Safe Ensemble Learning for Financial Risk
Mishra*, S. (2026) · Conference of the Open Innovations Association · doi:10.23919/fruct70069.2026.11506535 · Cited by 1
“Explainable Boosting Machine (EBM) provides a glass-box alternative with pairwise interaction detection [29], [30].”
Patil, S. P. (2025) · Journal of Artificial Intelligence Machine Learning and Neural Network · doi:10.55529/jaimlnn.51.122.136
Dapaah, E. C., & Grabowski, J. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21480-5_14
Kang, J., & Kim, K. (2026) · Applied Sciences · doi:10.3390/app16125809 · Cited by 1
When Efficiency Shapes Fairness: Cost-Optimised Human-in-the-Loop Routing
Breene, V., O'Hagan, A., & Costello, F. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812278
“These concerns have driven renewed interest in inherently interpretable model classes—such as generalized additive models with interactions (GA2Ms) and their extensions—which offer stronger guarantees of transparency and conceptual soundness [2, 31, 40].”
An AI-Reinforced Traffic Digital Twin for Testing Emergency Vehicle Interventions
Namdarpour, F., Zuo, F., Sha, D., Bonestroo, H., Gao, J., Chow, J. Y. J., & Özbay, K. M. A. (2026) · Data Science for Transportation · doi:10.1007/s42421-026-00166-4
“The explainable boosting machine (EBM) is a tree-based, cyclic gradient boosting generalized additive model, developed to be as accurate as state-of-the-art black box models while remaining interpretable (Lou et al. 2013).”
Bustos-Márquez, L. F., & Hegedus, S. (2026) · Algorithms · doi:10.3390/a19060496
Camastra, C., Novellino, F., Pelagi, A., Pullano, G. M., Quattrone, A., & Sarica, A. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-9575-4_24
An Explainable Multimodal Framework for Cyclist Safety Perception in Mixed Traffic Environments
Chiang, C.-Y., Wang, M., Fathy, Y., Jaber, M., & Abdelmoniem, A. M. (2026) · Applied Sciences · doi:10.3390/app16136690
Using Explainability Techniques for Alzheimer Disease and Cancer Detection
Moreira, A., Ferreira, A. J., & Leite, N. (2026) · Studies in computational intelligence · doi:10.1007/978-3-032-22633-4_8
Ivančić, H., Šagovac, M., Novalić, A., Blažević, F., Vuran, E., Potoč, D., & Jagušt, T. (2026) · International Convention on Information and Communication Technology, Electronics and Microelectronics · doi:10.1109/mipro70003.2026.11592048
“Comparative studies demonstrate that EBMs often achieve performance comparable to more complex models while maintaining human-readable feature contributions [15].”
Shrestha, K., Duddukunta, K. R., Lal, P., Hasan, M. M., Rahman, M. M., & Pradhan, N. S. (2026) · Natural Hazards · doi:10.1007/s11069-026-08310-7 · Cited by 1
“…variable using many shallow decision trees created through gradient boosting, with the critical distinction that tree generation and gradient updates are performed using a single predictor variable at a time in a round-robin fashion with a low learning rate (Lou et al. 2013; Nori et al. 2019).”
Tamm, A., SHINE, B., James, T. J., Withers, J., Salih, H., East, J. E., Oke, J., Davies, J., Morris, E. J., & Nicholson, B. D. (2026) · medRxiv · doi:10.64898/2026.07.16.26358105
Hybrid Interpretability and Performance Healthcare Model
Mcheick, H., Faye, M., & Msheik, B. (2026) · International Conference on Smart Applications, Communications and Networking (SmartNets) · doi:10.1109/smartnets69662.2026.11604935
“Nevertheless, the performance gap with black-box models may persist on certain highly nonlinear problems, limiting its adoption as a direct replacement in the most complex scenarios [16].”
Discrete Choice Model with Generalized Additive Utility Network for Interpretable Policy Evaluation
Nishi, T., & Hara, Y. (2026) · Transportation Research Record Journal of the Transportation Research Board · doi:10.1177/03611981261461509
Acoustic Features as Potential Vocal Biomarkers for Multiple Sclerosis
Giugliano, S., Basile, I., Dubbioso, R., Iodice, R., Aruta, L., & Sannino, G. (2026) · IEEE International Symposium on Cluster, Cloud and Internet Computing Workshops (CCGridW) · doi:10.1109/ccgridw69005.2026.00055
“To identify vocal biomarkers capable of differentiating between patients with MS and healthy controls, a set of interpretable ML models was examined: LR [34], Linear Discriminant Analysis (LDA) [35], Decision Tree (DT) [36] and Explainable Boosting Machine (EBM) [37].”
Contextualizing Generalized Additive Models
Loor, M., & Tré, G. D. (2026) · IEEE International Conference on Fuzzy Systems (FUZZ-IEEE) · doi:10.1109/fuzz69877.2026.11626460
“Six independent variables x 1 , · · · , x 6 uniformly distributed in [1 , 10] have been considered for modeling the impact on QLI of treatments A and B .”
Luo, Z., Zhu, Z., Wang, X., Liu, Y., & Cai, J. (2026) · Fuel · doi:10.1016/j.fuel.2026.140910
Explainability Assistant: a Conversational XAI Interface for Interpreting Energy Consumption Models
Krjutškov, R., Barbu, E., Sakkas, N., & Yfanti, S. (2026) · International Conference on Electrical, Computer and Energy Technologies (ICECET) · doi:10.1109/icecet65726.2026.11632877
“Early work focused on inherently interpretable models [6], [7], while post-hoc meth-ods emerged for black-box models.”
Lam, Q.-T., Fan, F., Wang, Y.-L., Feng, S.-W., Vo, T., Le, M. H. N., Vu, G., Le, N., & Lee, I.-T. (2026) · Diagnostics · doi:10.3390/diagnostics16162569
“Predictive modeling used a survey-weighted EBM restricted to main effects [34,35].”
Werner, A., Rosner, A., Żywiec, J., Mrozek, D., & Bach, M. (2026) · Applied Sciences · doi:10.3390/app16168154
Rajendiran, I., & Pillay, S. C. H. (2026) · International Journal of Computational Intelligence and Applications · doi:10.1142/s1469026826500392
Nie, Y., Fang, X., Wang, G., & Zhao, Y. (2026) · IEEE Access · doi:10.1109/access.2026.3728903
“EBM belongs to the family of generalized additive models (GAMs) [56], [57].”
Lam, Q. T., Fan, F.-Y., Peng, T., Feng, S., Huang, C. C., Chow, T.-H., Minh, L. H. N., Vu, G. T., Wang, Y.-L., Le, N. Q. K., et al. (2026) · Bioengineering · doi:10.3390/bioengineering13091020
“The additive EBM architecture also enabled coefficient-level localization of secondary contrasts, including the BMI contribution to the income-axis signal [17].”
Del-Valle-Soto, C., Cruz-Diaz, C.-S., Cardona, M., Ponce-Espinosa, H., Valdivia, L. J., & Visconti, P. (2026) · Algorithms · doi:10.3390/a19090741
Wei, S., Yang, H. C., Gao, Y., & Wang, Y. (2026) · Transport Policy · doi:10.1016/j.tranpol.2026.104407
Wang, L., Liu, X., Tian, T., Chen, J., Liu, L., Wang, Y., Li, Y., & Li, J. (2026) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2026.3729263
“…inherently inter-pretable models (e.g., linear models, Neural Additive Models (NAMs) [7], Generalized Additive Models with interactions (GA2Ms) [8], Explainable Boosting Machine (EBM) [9]), post-hoc explanations (e.g., Partial Dependence Plots (PDP) [10], SHAP [11], LIME [12]), example-based…”
Mohammad, S. A. A., Ye, X., Qian, C., & AL-Bakhrani, A. A. (2026) · IEEE Access · doi:10.1109/access.2026.3736824
“…regression, naive Bayes, kNN, RBF-SVM, decision tree, random forest, extra trees, histogram gradient boosting, XGBoost [20], CatBoost [21], LightGBM [28], MLP, and an Explainable Boosting Machine [29], [30], a generalized additive model included as a directly interpretable glass-box reference).”
Tymoshchuk, D., Lyashuk, О., Maruschak, P. O., Mironov, D., & Tymoshchuk, V. (2026) · Vehicles · doi:10.3390/vehicles8100225
Interpretable Model-Based Anomaly Detection for LoRaWAN Link Behavior Across Deployments
Fonseca, M. S. P., de Bittencourt, J. G. P., de Oliveira, F. G., Pedroso, C. M., & Pacheco, E. J. (2026) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2026.3738684
“…additive models fit one shape function per feature and remain readable while capturing nonlinear e ff ects [22], and extending them with a small number of selected pairwise interactions achieves almost the same performance as the best full-complexity models on a number of real datasets [42].”
State of Health Estimation for Lithium-Ion Batteries Using Charge-Segment Timing Indicators
Rehman, F. U., Manoharan, A., Andersen, H., Cecati, C., & Buccella, C. (2026) · IEEE International Conference on Compatibility, Power Electronics and Power Engineering · doi:10.1109/cpe-powereng67301.2026.11704737
“In addition to reporting predictive performance, Explainable Boosting Machine (EBM) [20], [21] is used to examine how each charging feature contributes to the estimated SoH.”
Zhang, J., Li, S., Mawalim, C. O., & Okada, S. (2026) · INTERNATIONAL CONFERENCE ON MULTIMODAL INTERACTION · doi:10.1145/3776574.3831183
Gao, T., Shen, L., & Lu, Z. (2026) · Journal of Marine Science and Engineering · doi:10.3390/jmse14191832
“In addition, this study incorporates the GA 2 M algorithm, along with its industrial implementation, the Explainable Boosting Machine (EBM) [39,40].”
Baydili, K. N. (2026) · Mathematics · doi:10.3390/math14193588
Predicting ESG adoption among Malaysian SMEs using explainable machine learning
Ad-Din, S., & Seera, M. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-02136-9
“Explainable Boosting Machine (EBM) is a glass-box generalised additive model with optional pairwise interactions (GA2M), trained by round-robin gradient boosting on shallow trees with a low learning rate [30, 35].”
Wang, G., Bennamoun, H., Kwok, W. H., Raza, R., Sohel, F. A., Kelly, B., Ratajczak, T., Reid, C. L., Marriott, R., Walker, R., et al. (2026) · PLOS Digital Health · doi:10.1371/journal.pdig.0001709
Automated Interpretable Neuron Screening for Calcium Imaging
Pospelov, N., Rogozhnikova, O., Plyusnin, V., Ivanova, A., Toropova, K. A., Ivashkina, O. I., Evgenev, A., & Anokhin, K. V. (2026) · Scientific School Dynamics of Complex Networks and their Applications (DCNA) · doi:10.1109/dcna72004.2026.11733380
“Candidates are classified by an Explainable Boosting Machine (EBM) [9], [10], a generalized additive model with logit link g , with intercept f 0 , per-feature shape functions f j , and |I| = 20 automatically selected pairwise interactions.”
Beshti, A., Adegboye, O. R., & Alzubi, A. B. (2026) · Applied Sciences · doi:10.3390/app162110532
Xie, Y., Liu, Y., & Ning, H. (2026) · International Journal of Financial Studies · doi:10.3390/ijfs14110292 · Cited by 1
An Interpretable Interaction–Based Kolmogorov–Arnold Network
Winkler, A., Marcial‐Romero, J. R., & Hernández, J. A. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-43214-8_7
Metadata
- Authors
- Lou, Y., Caruana, R., Gehrke, J., & Hooker, G.
- Affiliations
- Cornell University (US)
- Microsoft (United States) (US)
- Venue
- Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining
- Publisher
- ACM
- Published
- 2013-08-11
- Pages
- 623-631
- OpenAlex
- W2046945713
- Cited by
- 564
- References
- 22
- Citations by year
- 2026: 66 · 2025: 78 · 2024: 88 · 2023: 87 · 2022: 65 · 2021: 57 · 2020: 45 · 2019: 29 · 2018: 24 · 2017: 11 · 2016: 8 · 2015: 4 · 2014: 2
- FWCI
- 3.01
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Applications
- Physical Sciences › Computer Science › Computer Graphics and Computer-Aided Design › Data Visualization and Analytics
- Keywords
- generalized additive models, pairwise interactions, model interpretability, model accuracy
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.