VP
Vardan Papyan
cs.LGstat.MLcs.AIcs.CVcs.CLcs.ITmath.ITcs.CRSurgeonsadversarial robustness
On Valency
published · living versionsW_724c9fq5·v1 · currentpublished
Prevalence of Neural Collapse during the terminal phase of deep learning training
with X. Y. Han, David L. Donoho
1 version
Preprints & journals
31 papers in the corpus · 2015–2026Gradient Smoothing: Coupling Layer-wise Updates for Improved Optimization2606.30813v1 · Haoming Meng, Anton Sugolov, Vardan Papyan2026 · 0 citationsarXiv
Laplacian Heads Improve Transformers by Smoothing Token Representations2602.09297v3 · Yuchong Zhang, Vardan Papyan2026 · 0 citationsarXiv
On the Importance of Gaussianizing Representations2505.00685v2 · Daniel Eftekhari, Vardan Papyan2025 · 0 citationsarXiv
Attention Sinks: A 'Catch, Tag, Release' Mechanism for Embeddings2502.00919v2 · Stephen Zhang, Mustafa Khan, Vardan Papyan2025 · 0 citationsarXiv
OATS: Outlier-Aware Pruning Through Sparse and Low Rank Decomposition2409.13652v3 · Stephen Zhang, Vardan Papyan2024 · 0 citationsarXiv
Transformer Block Coupling and its Correlation with Generalization in LLMs2407.07810v5 · Murdock Aubry, Haoming Meng, Anton Sugolov et al.2024 · 0 citationsarXiv
Sparsest Models Elude Pruning: An Exposé of Pruning's Current Capabilities2407.04075v1 · Stephen Zhang, Vardan Papyan2024 · 0 citationsarXiv
Breach By A Thousand Leaks: Unsafe Information Leakage in `Safe' AI Responses2407.02551v2 · David Glukhov, Ziwen Han, Ilia Shumailov et al.2024 · 0 citationsarXiv
Linguistic Collapse: Neural Collapse in (Large) Language Models2405.17767v3 · Robert Wu, Vardan Papyan2024 · 1 citationarXiv
Pushing Boundaries: Mixup's Influence on Neural Collapse2402.06171v1 · Quinn Fisher, Haoming Meng, Vardan Papyan2024 · 0 citationsarXiv
Residual Alignment: Uncovering the Mechanisms of Residual Networks2401.09018v1 · Jianing Li, Vardan Papyan2024 · 8 citationsarXiv
Out of the Ordinary: Spectrally Adapting Regression for Covariate Shift2312.17463v1 · Benjamin Eyre, Elliot Creager, David Madras et al.2023 · 0 citationsarXiv
LLM Censorship: A Machine Learning Challenge or a Computer Security Problem?2307.10719v1 · David Glukhov, Ilia Shumailov, Yarin Gal et al.2023 · 19 citationsarXiv
Expert surgeons and deep learning models can predict the outcome of surgical hemorrhage from 1 min of video.35581213 · Pangal, Dhiraj J, Kugener, Guillaume, Zhu, Yichao et al.2022 · 26 citationsScientific reports. 2022;12(1):8137
Utility of the Simulated Outcomes Following Carotid Artery Laceration Video Data Set for Machine Learning Applications.35311962 · Kugener, Guillaume, Pangal, Dhiraj J, Cardinal, Tyler et al.2022 · 14 citationsJAMA network open. 2022;5(3):e223177
Neural Collapse Under MSE Loss: Proximity to and Dynamics on the Central Path2106.02073v4 · X.Y. Han, Vardan Papyan, David L. Donoho2021 · 28 citationsarXiv
Prevalence of neural collapse during the terminal phase of deep learning training.32958680 · Papyan, Vardan, Han, X Y, Donoho, David L2020 · 400 citationsProceedings of the National Academy of Sciences of the United States of America. 2020;117(40):24652-24663
Traces of Class/Cross-Class Structure Pervade Deep Learning Spectra2008.11865v1 · Vardan Papyan2020 · 14 citationsarXiv
Degrees of Freedom Analysis of Unrolled Neural Networks1906.03742v1 · Morteza Mardani, Qingyun Sun, Vardan Papyan et al.2019 · 5 citationsarXiv
Measurements of Three-Level Hierarchical Structure in the Outliers in the Spectrum of Deepnet Hessians1901.08244v1 · Vardan Papyan2019 · 35 citationsarXiv
The Full Spectrum of Deepnet Hessians at Scale: Dynamics with SGD Training and Sample Size1811.07062v2 · Vardan Papyan2018 · 39 citationsarXiv
Neural Proximal Gradient Descent for Compressive Imaging1806.03963v1 · Morteza Mardani, Qingyun Sun, Shreyas Vasawanala et al.2018 · 144 citationsarXiv
Recurrent Generative Adversarial Networks for Proximal Learning and Automated Compressive Image Recovery1711.10046v1 · Morteza Mardani, Hatef Monajemi, Vardan Papyan et al.2017 · 41 citationsarXiv
Multi-Layer Convolutional Sparse Modeling: Pursuit and Dictionary Learning1708.08705v2 · Jeremias Sulam, Vardan Papyan, Yaniv Romano et al.2017 · 120 citationsIEEE Transactions on Signal Processing, vol. 66, no. 15, pp. 4090-4104, Aug.1, 1 2018
Working Locally Thinking Globally: Theoretical Guarantees for Convolutional Sparse Coding1707.06066v1 · Vardan Papyan, Jeremias Sulam, Michael Elad2017 · 126 citationsarXiv
Convolutional Dictionary Learning via Local Processing1705.03239v1 · Vardan Papyan, Yaniv Romano, Jeremias Sulam et al.2017 · 139 citationsarXiv
Multimodal Latent Variable Analysis1611.08472v1 · Vardan Papyan, Ronen Talmon2016 · 0 citationsarXiv
Convolutional Neural Networks Analyzed via Convolutional Sparse Coding1607.08194v4 · Vardan Papyan, Yaniv Romano, Michael Elad2016 · 156 citationsarXiv
Working Locally Thinking Globally - Part I: Theoretical Guarantees for Convolutional Sparse Coding1607.02005v2 · Vardan Papyan, Jeremias Sulam, Michael Elad2016 · 4 citationsarXiv
Working Locally Thinking Globally - Part II: Stability and Algorithms for Convolutional Sparse Coding1607.02009v3 · Vardan Papyan, Jeremias Sulam, Michael Elad2016 · 11 citationsarXiv
Multi-Scale Patch-Based Image Restoration.26571527 · Papyan, Vardan, Elad, Michael2016 · 262 citationsIEEE transactions on image processing : a publication of the IEEE Signal Processing Society. 2016;25(1):249-61
Career total: 44 works. 31 are in this corpus.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.