Smaller is Better: Enhancing Transparency in Vehicle AI Systems via Pruning
2025/09/24 by Suwal, Sanish, Garg, Shaurya, Bhusal, Dipkamal +2
#Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences
paper · doi:10.48550/arxiv.2509.20148
Abstract
Connected and autonomous vehicles continue to heavily rely on AI systems, where transparency and security are critical for trust and operational safety. Post-hoc explanations provide transparency to these black-box like AI models but the quality and reliability of these explanations is often questioned due to inconsistencies and lack of faithfulness in representing model decisions. This paper systematically examines the impact of three widely used training approaches, namely natural training, adversarial training, and pruning, affect the quality of post-hoc explanations for traffic sign classifiers. Through extensive empirical evaluation, we demonstrate that pruning significantly enhances the comprehensibility and faithfulness of explanations (using saliency maps). Our findings reveal that pruning not only improves model efficiency but also enforces sparsity in learned representation, leading to more interpretable and reliable decisions. Additionally, these insights suggest that pruning is a promising strategy for developing transparent deep learning models, especially in resource-constrained vehicular AI systems.
Citations
- Do Sparse Subnetworks Exhibit Cognitively Aligned Attention? Effects of Pruning on Saliency Map Fidelity, Sparsity, and Concept Coherence
- Structured Gradient-based Interpretations via Norm-Regularized Adversarial Training
- A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recommendations
- A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis, and Recommendations
- Less is More: The Influence of Pruning on the Explainability of CNNs
- A Consistent and Efficient Evaluation Strategy for Attribution Methods
- NoiseGrad: Enhancing Explanations by Introducing Stochasticity to Model Weights
- What is the State of Neural Network Pruning?
- A Survey of Deep Learning Applications to Autonomous Vehicle Control
- Improving Feature Attribution through Input-specific Network Pruning
- Fooling LIME and SHAP: Adversarial Attacks on Post hoc Explanation\n Methods
- ML-LOO: Detecting Adversarial Examples with Feature Attribution
- On the Connection Between Adversarial Robustness and Saliency Map Interpretability
- Sanity Checks for Saliency Maps
- Soft Filter Pruning for Accelerating Deep Convolutional Neural Networks
- RISE: Randomized Input Sampling for Explanation of Black-box Models
- The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
- Interpreting Convolutional Neural Networks Through Compression
- The (Un)reliability of saliency methods
- Towards Deep Learning Models Resistant to Adversarial Attacks
- SmoothGrad: removing noise by adding noise
- A Unified Approach to Interpreting Model Predictions
- More is Less: A More Complicated Network with Less Inference Complexity
- Axiomatic Attribution for Deep Networks
- Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization
- "Why Should I Trust You?": Explaining the Predictions of Any Classifier
- Deep Residual Learning for Image Recognition
- Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding
- Learning both Weights and Connections for Efficient Neural Networks
- Striving for Simplicity: The All Convolutional Net
- Explaining and Harnessing Adversarial Examples
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
Cited by
Related