Histogram-based Parameter-efficient Tuning for Passive and Active Sonar Classification

2025/04/21 by Amirmohammad Mohammadi, Davelle Carreiro, Mohammadi, Amirmohammad +5 · 1 citation
Computer Science · #Digital Filter Design and Implementation #FOS: Computer and information sciences #Machine Learning (cs.LG) #Music Technology and Sound Studies #Neural Networks and Applications #Sound (cs.SD)

paper · pdf · doi:10.48550/arxiv.2504.15214

openalex publication_date 2025/04/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

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, Vessel Type Underwater Acoustic Data (VTUAD)) demonstrate that HPT outperforms conventional adapters. Notably, HPT achieves 91.8% vs. 89.8% accuracy on VTUAD. For active sonar imagery (Watertank, Turntable), HPT is competitive with other PETL methods. Furthermore, HPT yields feature representations closer to those of fully fine-tuned models. Overall, HPT balances parameter savings and provides a distribution-aware alternative to existing adapters and shows a promising direction for transfer learning in resource-constrained environments. The code is publicly available: https://github.com/Advanced-Vision-and-Learning-Lab/HLAST_DeepShip_ParameterEfficient.

Citations

• Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey
• Advancing Test-Time Adaptation in Wild Acoustic Test Settings
• Histogram Layer Time Delay Neural Networks for Passive Sonar Classification
• Scaling & Shifting Your Features: A New Baseline for Efficient Model Tuning
• DeepShip: An underwater acoustic benchmark dataset and a separable convolution based autoencoder for classification
• UniPELT: A Unified Framework for Parameter-Efficient Language Model Tuning
• Towards a Unified View of Parameter-Efficient Transfer Learning
• The Marine Debris Dataset for Forward-Looking Sonar Semantic Segmentation
• BitFit: Simple Parameter-efficient Fine-tuning for Transformer-based Masked Language-models
• LoRA: Low-Rank Adaptation of Large Language Models
• The Power of Scale for Parameter-Efficient Prompt Tuning
• AST: Audio Spectrogram Transformer
• Training data-efficient image transformers & distillation through attention
• Histogram Layers for Texture Analysis
• Weakly Labelled AudioSet Tagging with Attention Neural Networks
• Parameter-Efficient Transfer Learning for NLP
• A multi-device dataset for urban acoustic scene classification
• Return of Frustratingly Easy Domain Adaptation
• Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
• Prefix-Tuning: Optimizing Continuous Prompts for Generation
• Prefix-Tuning: Optimizing Continuous Prompts for Generation

Related

• Fine-grained Prompt Tuning: A Parameter and Memory Efficient Transfer Learning Method for High-resolution Medical Image Classification
• Optimizing Specific and Shared Parameters for Efficient Parameter Tuning
• Rethinking Efficient Tuning Methods from a Unified Perspective
• UniPT: Universal Parallel Tuning for Transfer Learning with Efficient Parameter and Memory
• Boosting Memory Efficiency in Transfer Learning for High-Resolution Medical Image Classification