Cooperative Coevolution versus Monolithic Evolutionary Search for Semi-Supervised Tabular Classification

TL;DR

This study introduces a cooperative coevolutionary approach (CC-SSL) for semi-supervised tabular classification with very limited labels, outperforming lightweight baselines and comparing favorably to monolithic methods.

Contribution

The paper proposes a novel cooperative coevolutionary method (CC-SSL) that evolves feature views and pseudo-labeling policies, demonstrating improved performance in low-label regimes.

Findings

CC-SSL and EA-SSL outperform lightweight baselines in MacroF1 and accuracy.

EA-SSL shows higher diversity and best-so-far fitness during search.

No significant difference in pseudo-label volume and validation metrics between CC-SSL and EA-SSL.

Abstract

This paper studies semi-supervised tabular classification in the extreme low-label regime using lightweight base learners. The paper proposes a cooperative coevolutionary method (CC-SSL) that evolves (i) two feature-subset views and (ii) a pseudo-labeling policy, and compares it to a matched monolithic evolutionary baseline (EA-SSL) and three lightweight SSL baselines. Experiments on 25 OpenML datasets with labeled fractions {1%,5%,10%} evaluate test MacroF1 and accuracy, together with evolutionary and pseudo-label diagnostics. CC-SSL and EA-SSL achieve higher median test MacroF1 than the lightweight baselines, with the largest separations at 1% labeled data. Most CC-SSL vs. EA-SSL comparisons are statistical draws on final test performance. EA-SSL shows higher best-so-far fitness and higher diversity during search, while time-to-target is comparable and generations-to-target favors EA-SSL in several multiclass settings. Pseudo-label volume, ProbeDrop, and validation optimism show no significant differences between CC-SSL and EA-SSL under the shared protocol.