Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Structural Compositional Function Networks: Interpretable Functional Compositions for Tabular Discovery

About

Despite the ubiquity of tabular data in high-stakes domains, traditional deep learning architectures often struggle to match the performance of gradient-boosted decision trees while maintaining scientific interpretability. Standard neural networks typically treat features as independent entities, failing to exploit the inherent manifold structural dependencies that define tabular distributions. We propose Structural Compositional Function Networks (StructuralCFN), a novel architecture that imposes a Relation-Aware Inductive Bias via a differentiable structural prior. StructuralCFN explicitly models each feature as a mathematical composition of its counterparts through Differentiable Adaptive Gating, which automatically discovers the optimal activation physics (e.g., attention-style filtering vs. inhibitory polarity) for each relationship. Our framework enables Structured Knowledge Integration, allowing domain-specific relational priors to be injected directly into the architecture to guide discovery. We evaluate StructuralCFN across a rigorous 10-fold cross-validation suite on 18 benchmarks, demonstrating statistically significant improvements (p < 0.05) on scientific and clinical datasets (e.g., Blood Transfusion, Ozone, WDBC). Furthermore, StructuralCFN provides Intrinsic Symbolic Interpretability: it recovers the governing "laws" of the data manifold as human-readable mathematical expressions while maintaining a compact parameter footprint (300--2,500 parameters) that is over an order of magnitude (10x--20x) smaller than standard deep baselines.

Fang Li• 2026

Related benchmarks

TaskDatasetResultRank
RegressionCA Housing--
45
ClassificationBreast Cancer (10-fold CV)
P-Value0.047
7
ClassificationHeart Disease (10-fold CV)
P-Value0.468
5
ClassificationWine Quality (10-fold CV)
P-Value0.001
5
ClassificationIonosphere (10-fold CV)
P-Value0.481
5
RegressionDiabetes (10-fold CV)
P-Value0.435
5
RegressionCA Housing (10-fold CV)
p-value0.008
5
RegressionDiabetes
MSE0.488
3
ClassificationBlood Transf. 1464 (10-fold CV)
Log-Loss0.418
2
ClassificationOzone 1487 10-fold CV
Log-Loss0.138
2
Showing 10 of 19 rows

Other info

Follow for update