Don't stop me now: Rethinking Validation Criteria for Model Parameter Selection

📅 2026-02-25
📈 Citations: 0
Influential: 0
📄 PDF

career value

188K/year
🤖 AI Summary
This study investigates the impact of model selection criteria—such as accuracy versus loss—on test performance in neural classifier training, particularly under early stopping with patience. Through systematic empirical evaluation using k-fold cross-validation on standard benchmarks, the work compares multiple validation metrics, including accuracy, cross-entropy, C-Loss, and PolyLoss, under both early stopping and post-hoc full-trajectory selection strategies. The findings reveal that validation loss–based criteria consistently outperform validation accuracy, which exhibits not only inferior performance but also lower stability. More critically, regardless of the selection criterion employed, the chosen models are typically substantially worse than the best test performance observed during training, thereby exposing a fundamental limitation in current model selection paradigms.

Technology Category

Application Category

📝 Abstract
Despite the extensive literature on training loss functions, the evaluation of generalization on the validation set remains underexplored. In this work, we conduct a systematic empirical and statistical study of how the validation criterion used for model selection affects test performance in neural classifiers, with attention to early stopping. Using fully connected networks on standard benchmarks under $k$-fold evaluation, we compare: (i) early stopping with patience and (ii) post-hoc selection over all epochs (i.e. no early stopping). Models are trained with cross-entropy, C-Loss, or PolyLoss; the model parameter selection on the validation set is made using accuracy or one of the three loss functions, each considered independently. Three main findings emerge. (1) Early stopping based on validation accuracy performs worst, consistently selecting checkpoints with lower test accuracy than both loss-based early stopping and post-hoc selection. (2) Loss-based validation criteria yield comparable and more stable test accuracy. (3) Across datasets and folds, any single validation rule often underperforms the test-optimal checkpoint. Overall, the selected model typically achieves test-set performance statistically lower than the best performance across all epochs, regardless of the validation criterion. Our results suggest avoiding validation accuracy (in particular with early stopping) for parameter selection, favoring loss-based validation criteria.
Problem

Research questions and friction points this paper is trying to address.

validation criteria
model selection
early stopping
generalization
neural classifiers
Innovation

Methods, ideas, or system contributions that make the work stand out.

validation criteria
early stopping
model selection
loss functions
generalization
🔎 Similar Papers
No similar papers found.
A
Andrea Apicella
Department of Information Engineering, Electrical Engineering, and Applied Mathematics (DIEM), University of Salerno, Via Giovanni Paolo II, 132, Fisciano (Salerno), 84084, Italy.
F
Francesco Isgrò
Department of Electrical Engineering and Information Technology, University of Naples Federico II, Via Claudio 21, Naples, 80125, Italy.
A
Andrea Pollastro
Department of Electrical Engineering and Information Technology, University of Naples Federico II, Via Claudio 21, Naples, 80125, Italy.
R
Roberto Prevete
Department of Electrical Engineering and Information Technology, University of Naples Federico II, Via Claudio 21, Naples, 80125, Italy.