Prototype-rule supports improved neural tensor learning on scarce data
Prototype-Rule Neurosymbolic Regularization for Rank-Constrained Tensor Neural Networks under Label Scarcity
Machine LearningComputer Vision and Pattern Recognition
Summary
Machine learning models that handle complex data often struggle when there isn't much labeled data to learn from. The authors studied a method that combines rank-constrained tensor neural networks with a special rule-based system called a prototype-rule to better guide learning when labels are scarce. This approach helps the model organize data classes more clearly during training, improving accuracy especially when data points are few and spatially separated. The method was tested on four different real-world hyperspectral image datasets, showing consistent gains in performance except for one case.
What this means in practice
- •For remote sensing analysts: Improve classification accuracy of hyperspectral images when only limited labeled training samples are available by using prototype-rule regularization during model training.
- •For environmental monitoring teams: Better identify land cover types from satellite data with fewer labeled examples by applying rank-constrained tensor learning combined with prototype-based rules.
Authors
Eftychios Protopapadakis, Konstantinos Makantasis, Konstantinos M. Giannoutakis
Abstract
Rank-constrained tensor neural networks reduce the parameterization of high-order inputs, but they do not explicitly constrain class geometry in the learned representation. This study investigates whether a differentiable prototype-rule can provide a complementary inductive bias for Rank-R tensor learning under limited supervision. The proposed framework augments the Rank-R objective with prototype-based regularization and optionally fuses prototype evidence with neural logits at inference. Four hyperspectral benchmarks are evaluated with four Rank-R configurations under both seven-fold stratification and spatially separated folds that mitigate leakage; a separate spatial study varies the class support budget from 2 to 20 samples. Under spatial evaluation, full neurosymbolic inference changes Macro-F1 score by +8.82 percentage points on Botswana, +5.49 on Indian Pines, +1.59 on Pavia University, and -0.62 on Salinas. Most of the benefit arises from training-time regularization, whereas inference fusion is small and dataset dependent.