Interpretability-by-Design with Accurate Locally Additive Models and Conditional Feature Effects

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gkolemis, Vasilis, Kavouras, Loukas, Kyriakopoulos, Dimitrios, Tsopelas, Konstantinos, Rontogiannis, Dimitrios, Casalicchio, Giuseppe, Dalamagas, Theodore, Diou, Christos
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908839823015936
author Gkolemis, Vasilis
Kavouras, Loukas
Kyriakopoulos, Dimitrios
Tsopelas, Konstantinos
Rontogiannis, Dimitrios
Casalicchio, Giuseppe
Dalamagas, Theodore
Diou, Christos
author_facet Gkolemis, Vasilis
Kavouras, Loukas
Kyriakopoulos, Dimitrios
Tsopelas, Konstantinos
Rontogiannis, Dimitrios
Casalicchio, Giuseppe
Dalamagas, Theodore
Diou, Christos
contents Generalized additive models (GAMs) offer interpretability through independent univariate feature effects but underfit when interactions are present in data. GA$^2$Ms add selected pairwise interactions which improves accuracy, but sacrifices interpretability and limits model auditing. We propose \emph{Conditionally Additive Local Models} (CALMs), a new model class, that balances the interpretability of GAMs with the accuracy of GA$^2$Ms. CALMs allow multiple univariate shape functions per feature, each active in different regions of the input space. These regions are defined independently for each feature as simple logical conditions (thresholds) on the features it interacts with. As a result, effects remain locally additive while varying across subregions to capture interactions. We further propose a principled distillation-based training pipeline that identifies homogeneous regions with limited interactions and fits interpretable shape functions via region-aware backfitting. Experiments on diverse classification and regression tasks show that CALMs consistently outperform GAMs and achieve accuracy comparable with GA$^2$Ms. Overall, CALMs offer a compelling trade-off between predictive accuracy and interpretability.
format Preprint
id arxiv_https___arxiv_org_abs_2602_16503
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Interpretability-by-Design with Accurate Locally Additive Models and Conditional Feature Effects
Gkolemis, Vasilis
Kavouras, Loukas
Kyriakopoulos, Dimitrios
Tsopelas, Konstantinos
Rontogiannis, Dimitrios
Casalicchio, Giuseppe
Dalamagas, Theodore
Diou, Christos
Machine Learning
Artificial Intelligence
Generalized additive models (GAMs) offer interpretability through independent univariate feature effects but underfit when interactions are present in data. GA$^2$Ms add selected pairwise interactions which improves accuracy, but sacrifices interpretability and limits model auditing. We propose \emph{Conditionally Additive Local Models} (CALMs), a new model class, that balances the interpretability of GAMs with the accuracy of GA$^2$Ms. CALMs allow multiple univariate shape functions per feature, each active in different regions of the input space. These regions are defined independently for each feature as simple logical conditions (thresholds) on the features it interacts with. As a result, effects remain locally additive while varying across subregions to capture interactions. We further propose a principled distillation-based training pipeline that identifies homogeneous regions with limited interactions and fits interpretable shape functions via region-aware backfitting. Experiments on diverse classification and regression tasks show that CALMs consistently outperform GAMs and achieve accuracy comparable with GA$^2$Ms. Overall, CALMs offer a compelling trade-off between predictive accuracy and interpretability.
title Interpretability-by-Design with Accurate Locally Additive Models and Conditional Feature Effects
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2602.16503