Image de couverture de livre

Fundamental Limits of Neural Network Sparsification: Evidence from Catastrophic Interpretability Collapse

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Roy, Dip, Misra, Rajiv, Singh, Sanjay Kumar
Format:	Preprint
Publié:	2026
Sujets:	Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

A Multi-Level Causal Intervention Framework for Mechanistic Interpretability in Variational Autoencoders
par: Roy, Dip, et autres
Publié: (2025)

Posterior-Calibrated Causal Circuits in Variational Autoencoders: Why Image-Domain Interpretability Fails on Tabular Data
par: Roy, Dip, et autres
Publié: (2026)

Detection Without Correction: A Robust Asymmetry in Activation-Based Hallucination Probing
par: Roy, Dip, et autres
Publié: (2026)

ARDDQN: Attention Recurrent Double Deep Q-Network for UAV Coverage Path Planning and Data Harvesting
par: Kumar, Praveen, et autres
Publié: (2024)

MemGuard-Alpha: Detecting and Filtering Memorization-Contaminated Signals in LLM-Based Financial Forecasting via Membership Inference and Cross-Model Disagreement
par: Roy, Anisha, et autres
Publié: (2026)

Bayesian Autoencoder for Medical Anomaly Detection: Uncertainty-Aware Approach for Brain 2 MRI Analysis
par: Roy, Dip
Publié: (2025)

Structure-Aware Spectral Sparsification via Uniform Edge Sampling
par: He, Kaiwen, et autres
Publié: (2025)

Efficient Distributed Training through Gradient Compression with Sparsification and Quantization Techniques
par: Singh, Shruti, et autres
Publié: (2024)

Spectral Neural Graph Sparsification
par: Liguori, Angelica, et autres
Publié: (2025)

Graph Sparsification for Enhanced Conformal Prediction in Graph Neural Networks
par: He, Yuntian, et autres
Publié: (2024)

Mitigating Over-Squashing in Graph Neural Networks by Spectrum-Preserving Sparsification
par: Liang, Langzhang, et autres
Publié: (2025)

Some Fundamental Aspects about Lipschitz Continuity of Neural Networks
par: Khromov, Grigory, et autres
Publié: (2023)

Imitation Learning of MPC with Neural Networks: Error Guarantees and Sparsification
par: Alsmeier, Hendrik, et autres
Publié: (2025)

A Methodology-Oriented Study of Catastrophic Forgetting in Incremental Deep Neural Networks
par: Kumar, Ashutosh, et autres
Publié: (2024)

SGS-GNN: A Supervised Graph Sparsification method for Graph Neural Networks
par: Das, Siddhartha Shankar, et autres
Publié: (2025)

Less Is More -- On the Importance of Sparsification for Transformers and Graph Neural Networks for TSP
par: Lischka, Attila, et autres
Publié: (2024)

Fundamental Limits of Perfect Concept Erasure
par: Chowdhury, Somnath Basu Roy, et autres
Publié: (2025)

Shaving Weights with Occam's Razor: Bayesian Sparsification for Neural Networks Using the Marginal Likelihood
par: Dhahri, Rayen, et autres
Publié: (2024)

Large-Scale Spectral Graph Neural Networks via Laplacian Sparsification: Technical Report
par: Ding, Haipeng, et autres
Publié: (2025)

Automatic and Structure-Aware Sparsification of Hybrid Neural ODEs
par: Zou, Bob Junyi, et autres
Publié: (2025)

Consequentialist Objectives and Catastrophe
par: Marklund, Henrik, et autres
Publié: (2026)

Theoretical Learning Performance of Graph Neural Networks: The Impact of Jumping Connections and Layer-wise Sparsification
par: Sun, Jiawei, et autres
Publié: (2025)

On the Robustness of Neural Collapse and the Neural Collapse of Robustness
par: Su, Jingtong, et autres
Publié: (2023)

Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning
par: Zhang, Ruiqi, et autres
Publié: (2024)

Efficient and Interpretable Information Retrieval for Product Question Answering with Heterogeneous Data
par: Biswas, Biplob, et autres
Publié: (2024)

The Fundamental Limits of Fraud Detection in Card Payment Networks
par: Dhama, Gaurav
Publié: (2026)

Fundamental Limits of Crystalline Equivariant Graph Neural Networks: A Circuit Complexity Perspective
par: Cao, Yang, et autres
Publié: (2025)

Hybrid Quantum-Classical Ridgelet Neural Networks for Portfolio Optimization
par: Yadav, Bahadur, et autres
Publié: (2026)

Spectral Graph Sparsification Preserves Representation Geometry in Graph Neural Networks
par: Krishnagopal, Sanjukta
Publié: (2026)

Towards Lightweight Graph Neural Network Search with Curriculum Graph Sparsification
par: Xie, Beini, et autres
Publié: (2024)

Visual Prompting Upgrades Neural Network Sparsification: A Data-Model Perspective
par: Jin, Can, et autres
Publié: (2023)

I Can't Believe It's Not Robust: Catastrophic Collapse of Safety Classifiers under Embedding Drift
par: Sahoo, Subramanyam, et autres
Publié: (2026)

Optimization Proxies using Limited Labeled Data and Training Time -- A Semi-Supervised Bayesian Neural Network Approach
par: Pareek, Parikshit, et autres
Publié: (2024)

How Interpretable Are Interpretable Graph Neural Networks?
par: Chen, Yongqiang, et autres
Publié: (2024)

A Study on Quantum Neural Networks in Healthcare 5.0
par: Chakraborty, Sanjay
Publié: (2024)

Leveraging Intermediate Neural Collapse with Simplex ETFs for Efficient Deep Neural Networks
par: Liu, Emily
Publié: (2024)

Beyond Unconstrained Features: Neural Collapse for Shallow Neural Networks with General Data
par: Hong, Wanli, et autres
Publié: (2024)

Mask in the Mirror: Implicit Sparsification
par: Jacobs, Tom, et autres
Publié: (2024)

Neural CRNs: A Natural Implementation of Learning in Chemical Reaction Networks
par: Nagipogu, Rajiv Teja, et autres
Publié: (2024)

Natural Mitigation of Catastrophic Interference: Continual Learning in Power-Law Learning Environments
par: Gandhi, Atith, et autres
Publié: (2024)