Histogram-based Parameter-efficient Tuning for Passive Sonar Classification
This work addresses distributional shifts in feature embeddings for passive sonar classification, offering a more efficient alternative to existing adapters in resource-constrained environments, though it is incremental as it builds on prior PETL methods.
The paper tackles the problem of parameter-efficient transfer learning for passive sonar classification by proposing a histogram-based tuning technique that captures target domain statistics, achieving 91.8% accuracy on the VTUAD dataset compared to 89.8% with conventional adapters.
Parameter-efficient transfer learning (PETL) methods adapt large artificial neural networks to downstream tasks without fine-tuning the entire model. However, existing additive methods, such as adapters, sometimes struggle to capture distributional shifts in intermediate feature embeddings. We propose a novel histogram-based parameter-efficient tuning (HPT) technique that captures the statistics of the target domain and modulates the embeddings. Experimental results on three downstream passive sonar datasets (ShipsEar, DeepShip, VTUAD) demonstrate that HPT outperforms conventional adapters. Notably, HPT achieves 91.8% vs. 89.8% accuracy on VTUAD. Furthermore, HPT trains faster and yields feature representations closer to those of fully fine-tuned models. Overall, HPT balances parameter savings and performance, providing a distribution-aware alternative to existing adapters and shows a promising direction for scalable transfer learning in resource-constrained environments. The code is publicly available: https://github.com/Advanced-Vision-and-Learning-Lab/HLAST_DeepShip_ParameterEfficient.