Neural Feature Governance: Extending Atom Prevalence

A new Bayesian framework called Neural Atom Prevalence (NAP) achieves state-of-the-art structural sparsity in feedforward neural networks, reducing active nodes to as few as 8% of the original dense architecture on MNIST while maintaining near-nominal predictive interval coverage (93.4% observed against a 95% target). The method, introduced in arXiv:2607.21671v1, also provides well-calibrated uncertainty quantification with model ignorance accounting for only 3-4% of total predictive variance.

arXiv:2607.21671v1 Announce Type: new Abstract: Neural network compression and interpretability remain open challenges in modern deep learn- ing, where billion-parameter architectures deliver impressive accuracy at the cost of trans- parency, computational efficiency, and reliable uncertainty quantification. This paper introduces Neural Atom Prevalence NAP , a principled Bayesian framework for structured node-level model selection in feedforward neural networks. NAP introduces the neural atom activation unit and functions as a hybrid method operating through a four-phase pipeline: Bayesian Lottery Ticket BLT identification via Iterative Magnitude Pruning IMP , soft variational training of the Spike and Slab Independent Gaussian SS-IG model, Poisson-Binomial PB optimal layer-size selection, and Bayesian fine-tuning to produce a sparse, stable, interpretable, and accurate model. Extensive empirical validation across simulated nonlinear regression, two UCI benchmark datasets Concrete, YearPredictionMSD , and the MNIST image classification task demonstrates that NAP achieves state-of-the-art structural sparsity, reducing active nodes to as few as 8% of the original dense architecture on MNIST, while well-calibrated probabilisti- cally: the aleatoric-epistemic uncertainty decomposition reveals that model ignorance accounts for only 3 to 4% of total predictive variance across all experiments, and regression reliability diagrams confirm a near-nominal predictive interval coverage 93.4% observed against a 95% target . These results establish NAP as a reliable, theoretically grounded, and computation- ally tractable solution to the simultaneous pursuit of sparsity, accuracy, interpretability, and uncertainty quantification in Bayesian neural networks.