Cover Image

Do Sparse Subnetworks Exhibit Cognitively Aligned Attention? Effects of Pruning on Saliency Map Fidelity, Sparsity, and Concept Coherence

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Suwal, Sanish, Bhusal, Dipkamal, Clifford, Michael, Rastogi, Nidhi
Format:	Preprint
Published:	2025
Subjects:	Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Smaller is Better: Enhancing Transparency in Vehicle AI Systems via Pruning
by: Suwal, Sanish, et al.
Published: (2025)

Training for Trustworthy Saliency Maps: Adversarial Training Meets Feature-Map Smoothing
by: Bhusal, Dipkamal, et al.
Published: (2026)

FACE: Faithful Automatic Concept Extraction
by: Bhusal, Dipkamal, et al.
Published: (2025)

Learning to Look: Cognitive Attention Alignment with Vision-Language Models
by: Yang, Ryan L., et al.
Published: (2025)

H-Sets: Hessian-Guided Discovery of Set-Level Feature Interactions in Image Classifiers
by: Mehrotra, Ayushi, et al.
Published: (2026)

Concept-Based Masking: A Patch-Agnostic Defense Against Adversarial Patch Attacks
by: Mehrotra, Ayushi, et al.
Published: (2025)

Adversarial Patterns: Building Robust Android Malware Classifiers
by: Bhusal, Dipkamal, et al.
Published: (2022)

R+R: Revisiting Static Feature-Based Android Malware Detection using Machine Learning
by: Alam, Md Tanvirul, et al.
Published: (2024)

Multi-Label Classification of Thoracic Diseases using Dense Convolutional Network on Chest Radiographs
by: Bhusal, Dipkamal, et al.
Published: (2022)

CTIBench: A Benchmark for Evaluating LLMs in Cyber Threat Intelligence
by: Alam, Md Tanvirul, et al.
Published: (2024)

PASA: Attack Agnostic Unsupervised Adversarial Detection using Prediction & Attribution Sensitivity Analysis
by: Bhusal, Dipkamal, et al.
Published: (2024)

Rethinking Saliency Maps: A Cognitive Human Aligned Taxonomy and Evaluation Framework for Explanations
by: Elisha, Yehonatan, et al.
Published: (2025)

AthenaBench: A Dynamic Benchmark for Evaluating LLMs in Cyber Threat Intelligence
by: Alam, Md Tanvirul, et al.
Published: (2025)

SparseMM: Head Sparsity Emerges from Visual Concept Responses in MLLMs
by: Wang, Jiahui, et al.
Published: (2025)

Gaussian Smoothing in Saliency Maps: The Stability-Fidelity Trade-Off in Neural Network Interpretability
by: Ye, Zhuorui, et al.
Published: (2024)

Attention Sparsity is Input-Stable: Training-Free Sparse Attention for Video Generation via Offline Sparsity Profiling and Online QK Co-Clustering
by: Luo, Jiayi, et al.
Published: (2026)

One-Cycle Structured Pruning via Stability-Driven Subnetwork Search
by: Ghimire, Deepak, et al.
Published: (2025)

Sparse Prior Is Not All You Need: When Differential Directionality Meets Saliency Coherence for Infrared Small Target Detection
by: Zhou, Fei, et al.
Published: (2024)

A General Framework for Group Sparsity in Hyperspectral Unmixing Using Endmember Bundles
by: Bhusal, Gokul, et al.
Published: (2025)

OrthCaps: An Orthogonal CapsNet with Sparse Attention Routing and Pruning
by: Geng, Xinyu, et al.
Published: (2024)

SCOPE: Saliency-Coverage Oriented Token Pruning for Efficient Multimodel LLMs
by: Deng, Jinhong, et al.
Published: (2025)

CASL: Concept-Aligned Sparse Latents for Interpreting Diffusion Models
by: He, Zhenghao, et al.
Published: (2026)

Pushing the Limits of Sparsity: A Bag of Tricks for Extreme Pruning
by: Li, Andy, et al.
Published: (2024)

The Limits of Perception: Analyzing Inconsistencies in Saliency Maps in XAI
by: Stubbin, Anna, et al.
Published: (2024)

A Psychophysically Oriented Saliency Map Prediction Model
by: Li, Qiang
Published: (2020)

SUM: Saliency Unification through Mamba for Visual Attention Modeling
by: Hosseini, Alireza, et al.
Published: (2024)

Visual-TCAV: Concept-based Attribution and Saliency Maps for Post-hoc Explainability in Image Classification
by: De Santis, Antonio, et al.
Published: (2024)

AlignVid: Training-Free Attention Scaling for Semantic Fidelity in Text-Guided Image-to-Video Generation
by: Liu, Yexin, et al.
Published: (2025)

SPARC: Concept-Aligned Sparse Autoencoders for Cross-Model and Cross-Modal Interpretability
by: Nasiri-Sarvi, Ali, et al.
Published: (2025)

Dynamic Sparse Training with Structured Sparsity
by: Lasby, Mike, et al.
Published: (2023)

Learning to Incorporate Texture Saliency Adaptive Attention to Image Cartoonization
by: Gao, Xiang, et al.
Published: (2022)

Gun Detection Using Combined Human Pose and Weapon Appearance
by: Maligireddy, Amulya Reddy, et al.
Published: (2025)

SalM$^{2}$: An Extremely Lightweight Saliency Mamba Model for Real-Time Cognitive Awareness of Driver Attention
by: Zhao, Chunyu, et al.
Published: (2025)

2D or 3D: Who Governs Salience in VLA Models? -- Tri-Stage Token Pruning Framework with Modality Salience Awareness
by: Zheng, Zihao, et al.
Published: (2026)

SLA: Beyond Sparsity in Diffusion Transformers via Fine-Tunable Sparse-Linear Attention
by: Zhang, Jintao, et al.
Published: (2025)

SparseVILA: Decoupling Visual Sparsity for Efficient VLM Inference
by: Khaki, Samir, et al.
Published: (2025)

DiffSparse: Accelerating Diffusion Transformers with Learned Token Sparsity
by: Zhu, Haowei, et al.
Published: (2026)

Learning Compatible Multi-Prize Subnetworks for Asymmetric Retrieval
by: Sun, Yushuai, et al.
Published: (2025)

High-Fidelity Mobile Avatars with Pruned Local Blendshapes
by: Zhan, Youyi, et al.
Published: (2026)

Pruning for Robust Concept Erasing in Diffusion Models
by: Yang, Tianyun, et al.
Published: (2024)