Daniel Soudry

发表

Characterizing Implicit Bias in Terms of Optimization Geometry pdf

Nathan Srebro, Daniel Soudry, Suriya Gunasekar, 2018, ICML.

No bad local minima: Data independent training error guarantees for multilayer neural networks pdf

Daniel Soudry, Yair Carmon, Daniel Soudry, 2016, ArXiv.

Binarized Neural Networks

Ran El-Yaniv, Yoshua Bengio, Daniel Soudry, 2016, ArXiv.

Train longer, generalize better: closing the generalization gap in large batch training of neural networks

Elad Hoffer, Daniel Soudry, Itay Hubara, 2017, NIPS.

Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations pdf

Ran El-Yaniv, Yoshua Bengio, Daniel Soudry, 2016, J. Mach. Learn. Res..

Binarized Neural Networks

Ran El-Yaniv, Yoshua Bengio, Daniel Soudry, 2016, NIPS.

Expectation Backpropagation: Parameter-Free Training of Multilayer Neural Networks with Continuous or Discrete Weights

Ron Meir, Daniel Soudry, Itay Hubara, 2014, NIPS.

Exponentially vanishing sub-optimal local minima in multilayer neural networks

Elad Hoffer, Daniel Soudry, Daniel Soudry, 2017, ICLR.

Efficient "Shotgun" Inference of Neural Connectivity from Highly Sub-sampled Activity Data

Daniel Soudry, Min-hwan Oh, Liam Paninski, 2015, PLoS Comput. Biol..

A structured matrix factorization framework for large scale calcium imaging data analysis

Kira E. Poskanzer, L. Paninski, R. Yuste, 2014, 1409.2903.

Memristor-Based Multilayer Neural Networks With Online Gradient Descent Training

Avinoam Kolodny, Daniel Soudry, Dotan Di Castro, 2015, IEEE Transactions on Neural Networks and Learning Systems.

Lexicographic and Depth-Sensitive Margins in Homogeneous and Non-Homogeneous Deep Models

Nathan Srebro, Daniel Soudry, Jason D. Lee, 2019, ICML.

Simultaneous Denoising, Deconvolution, and Demixing of Calcium Imaging Data

David Pfau, Daniel Soudry, Liam Paninski, 2016, Neuron.

Fast Constrained Non-negative Matrix Factorization for Whole-Brain Calcium Imaging Data

L. Paninski, Daniel Soudry, Jeremy Freeman, 2016 .

Post training 4-bit quantization of convolutional networks for rapid-deployment

Daniel Soudry, Ron Banner, Yury Nahshan, 2018, NeurIPS.

Task Agnostic Continual Learning Using Online Variational Bayes pdf

Elad Hoffer, Daniel Soudry, Chen Zeno, 2018, 1803.10123.

Why Cold Posteriors? On the Suboptimal Generalization of Optimal Bayes Estimates

Ari Pakman, Daniel Soudry, Chen Zeno, 2021 .

Extracting grid cell characteristics from place cell inputs using non-negative principal component analysis

Ron Meir, Daniel Soudry, Yedidyah Dordek, 2016, eLife.

Improving Post Training Neural Quantization: Layer-wise Calibration and Integer Programming pdf

Ron Banner, Daniel Soudry, Itay Hubara, 2020, ArXiv.

Study on Binary Multilayer Neural Networks with EBP Algorithm on Image Classification

Daniel Soudry, Zhiyong Cheng, Zexi Mao, 2015 .

Quantized Back-Propagation: Training Binarized Neural Networks with Quantized Gradients

Elad Hoffer, Daniel Soudry, Itay Hubara, 2018 .

Author response: Extracting grid cell characteristics from place cell inputs using non-negative principal component analysis

Daniel Soudry, Yedidyah Dordek, Dori Derdikman, 2015 .

XPONENTIALLY VANISHING SUB-OPTIMAL LOCAL MINIMA IN MULTILAYER NEURAL NETWORKS

Daniel Soudry, Elad Hoffer, 2018 .

NeuroResource Simultaneous Denoising , D econvolution , and Demixing of Calcium Imaging Data Highlights

Timothy A. Machado, J. Merel, L. Paninski, 2016 .

The Knowledge Within: Methods for Data-Free Model Compression

Daniel Soudry, Itay Hubara, Elad Hoffer, 2019, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Correction: Efficient "Shotgun" Inference of Neural Connectivity from Highly Sub-sampled Activity Data

L. Paninski, Daniel Soudry, G. Iyengar, 2015, PLoS computational biology.

An exact reduction of the master equation to a strictly stable system with an explicit expression for the stationary distribution

Ron Meir, Daniel Soudry, R. Meir, 2012, 1207.4436.

How catastrophic can catastrophic forgetting be in linear regression?

Daniel Soudry, Rachel A. Ward, Itay Evron, 2022, COLT.

Training of quantized deep neural networks using a magnetic tunnel junction-based synapse

Daniel Soudry, Itay Hubara, S. Kvatinsky, 2019, Semiconductor Science and Technology.

At Stability's Edge: How to Adjust Hyperparameters to Preserve Minima Selection in Asynchronous Training of Neural Networks?

Mor Shpigel Nacson, Daniel Soudry, Elad Hoffer, 2019, ICLR.

The Implicit Bias of Minima Stability: A View from Function Space

Daniel Soudry, T. Michaeli, Rotem Mulayoff, 2021, NeurIPS.

Logarithmic Unbiased Quantization: Simple 4-bit Training in Deep Learning

Hilla Ben Yaacov, Daniel Soudry, Elad Hoffer, 2021, 2112.10769.

Implicit Bias of the Step Size in Linear Diagonal Neural Networks

Mor Shpigel Nacson, Nathan Srebro, Daniel Soudry, 2022, ICML.

A Function Space View of Bounded Norm Infinite Width ReLU Nets: The Multivariate Case pdf

Nathan Srebro, Daniel Soudry, Rebecca Willett, 2019, ICLR.

How do infinite width bounded norm networks look in function space? pdf

Nathan Srebro, Daniel Soudry, Pedro H. P. Savarese, 2019, COLT.

The Global Optimization Geometry of Shallow Linear Neural Networks

Yonina C. Eldar, Zhihui Zhu, Michael B. Wakin, 2018, Journal of Mathematical Imaging and Vision.

Conductance-Based Neuron Models and the Slow Dynamics of Excitability

Ron Meir, Daniel Soudry, R. Meir, 2012, Front. Comput. Neurosci..

Simple, Fast and Accurate Implementation of the Diffusion Approximation Algorithm for Stochastic Ion Channels with Multiple States

Daniel Soudry, Patricio Orio, Daniel Soudry, 2011, PloS one.

Multi-scale approaches for high-speed imaging and analysis of large neural populations

Rafael Yuste, Daniel Soudry, Liam Paninski, 2016, bioRxiv.

Stochastic Gradient Descent on Separable Data: Exact Convergence with a Fixed Learning Rate

Nathan Srebro, Daniel Soudry, Mor Shpigel Nacson, 2018, AISTATS.

The Implicit Bias of Gradient Descent on Separable Data pdf

Nathan Srebro, Elad Hoffer, Daniel Soudry, 2017, J. Mach. Learn. Res..

On the Implicit Bias of Initialization Shape: Beyond Infinitesimal Mirror Descent

Nathan Srebro, Daniel Soudry, Amir Globerson, 2021, ICML.

Kernel and Rich Regimes in Overparametrized Models

Nathan Srebro, Daniel Soudry, Jason D. Lee, 2019, COLT.

Geochemistry of cadmium in the Negev phosphorites

Daniel Soudry, Y. Nathan, D. Shitrit, 1997 .

Scalable Methods for 8-bit Training of Neural Networks

Elad Hoffer, Ron Banner, Daniel Soudry, 2018, NeurIPS.

Hebbian Learning Rules with Memristors

Dotan Di Castro, Daniel Soudry, A. Kolodny, 2013 .

History-Dependent Dynamics in a Generic Model of Ion Channels – An Analytic Study

Ron Meir, Daniel Soudry, R. Meir, 2009, Front. Comput. Neurosci..

Seizure pathways: A model-based investigation

Levin Kuhlmann, Daniel Soudry, David B Grayden, 2018, PLoS Comput. Biol..

Bifurcation analysis of two coupled Jansen-Rit neural mass models

Daniel Soudry, David B Grayden, Dragan Nešić, 2018, PloS one.

ACIQ: Analytical Clipping for Integer Quantization of neural networks pdf

Elad Hoffer, Ron Banner, Daniel Soudry, 2018, ArXiv.

Convergence of Gradient Descent on Separable Data

Nathan Srebro, Daniel Soudry, Jason D. Lee, 2018, AISTATS.

Accurate Post Training Quantization With Small Calibration Sets

Ron Banner, Daniel Soudry, Itay Hubara, 2021, ICML.

Diffusion approximation-based simulation of stochastic ion channels: which method to use?

Daniel Soudry, Patricio Orio, Danilo Pezo, 2014, Front. Comput. Neurosci..

The neuronal response at extended timescales: a linearized spiking input–output relation

Ron Meir, Daniel Soudry, R. Meir, 2014, Front. Comput. Neurosci..

The neuronal response at extended timescales: long-term correlations without long-term memory

Ron Meir, Daniel Soudry, R. Meir, 2014, Front. Comput. Neurosci..

Augment Your Batch: Improving Generalization Through Instance Repetition

Torsten Hoefler, Elad Hoffer, Daniel Soudry, 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Implicit Bias in Deep Linear Classification: Initialization Scale vs Training Accuracy pdf

Nathan Srebro, Daniel Soudry, Jason D. Lee, 2020, NeurIPS.

Implicit Bias of Gradient Descent on Linear Convolutional Networks

Nathan Srebro, Daniel Soudry, Suriya Gunasekar, 2018, NeurIPS.

Norm matters: efficient and accurate normalization schemes in deep networks

Elad Hoffer, Ron Banner, Daniel Soudry, 2018, NeurIPS.

Mix & Match: training convnets with mixed image sizes for improved accuracy, speed and scale resiliency pdf

Torsten Hoefler, Elad Hoffer, Daniel Soudry, 2019, ArXiv.

Kernel and Deep Regimes in Overparametrized Models pdf

Nathan Srebro, Daniel Soudry, Jason D. Lee, 2019, ArXiv.

Accelerated Sparse Neural Training: A Provable and Efficient Method to Find N: M Transposable Masks pdf

Brian Chmiel, Ron Banner, Daniel Soudry, 2021, NeurIPS.

"Neuronal spike generation mechanism as an oversampling, noise-shaping A-to-D converter"

Daniel Soudry, Dmitri B. Chklovskii, Daniel Soudry, 2012, NIPS.

Alias-Free Convnets: Fractional Shift Invariance via Polynomial Activations

Daniel Soudry, T. Michaeli, Hagay Michaeli, 2023, 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

Neural gradients are near-lognormal: improved quantized and sparse training

Brian Chmiel, Liad Ben-Uri, Elad Hoffer, 2020, ICLR.

Logarithmic Unbiased Quantization: Practical 4-bit Training in Deep Learning pdf

Brian Chmiel, Elad Hoffer, Ron Banner, 2021, ArXiv.

Augment your batch: better training with larger batches pdf

Torsten Hoefler, Elad Hoffer, Daniel Soudry, 2019, ArXiv.

A Shotgun Sampling Solution for the Common Input Problem in Neural Connectivity Inference

Daniel Soudry, Min-hwan Oh, Liam Paninski, 2022 .

Explore to Generalize in Zero-Shot RL

Aviv Tamar, Daniel Soudry, E. Zisselman, 2023, ArXiv.

Regularization Guarantees Generalization in Bayesian Reinforcement Learning through Algorithmic Stability pdf

Daniel Soudry, Aviv Tamar, Ev Zisselman, 2021, AAAI.

Physics-Aware Downsampling with Deep Learning for Scalable Flood Modeling pdf

Yossi Matias, Daniel Soudry, Zvika Ben-Haim, 2021, NeurIPS.

A Statistical Framework for Efficient Out of Distribution Detection in Deep Neural Networks

Daniel Soudry, R. Heller, T. Frostig, 2021, ICLR.

Continual Learning in Linear Classification on Separable Data

Daniel Soudry, Itay Evron, N. Srebro, 2023, ICML.

Are deep networks broken, and should we fix them?

Daniel Soudry, Y. Carmon, T. Michaeli, 2023 .

On the Blindspots of Convolutional Networks pdf

Elad Hoffer, Daniel Soudry, Shai Fine, 2018, ArXiv.

Infer2Train: leveraging inference for better training of deep networks

Elad Hoffer, Daniel Soudry, Berry Weinstein, 2018 .

Bayesian Gradient Descent: Online Variational Bayes Learning with Increased Robustness to Catastrophic Forgetting and Weight Pruning pdf

Elad Hoffer, Daniel Soudry, Chen Zeno, 2018, ArXiv.

Task-Agnostic Continual Learning Using Online Variational Bayes With Fixed-Point Updates

Daniel Soudry, Elad Hoffer, Chen Zeno, 2020, Neural Computation.

Fix your classifier: the marginal value of training the last weight layer

Daniel Soudry, Itay Hubara, Elad Hoffer, 2018, ICLR.

Beyond Signal Propagation: Is Feature Diversity Necessary in Deep Neural Network Initialization?

Daniel Soudry, Dar Gilboa, Yaniv Blumenfeld, 2020, ICML.

A Mean Field Theory of Quantized Deep Networks: The Quantization-Depth Trade-Off

Daniel Soudry, Yaniv Blumenfeld, D. Gilboa, 2019, Neural Information Processing Systems.

Exponentially vanishing sub-optimal local minima in multilayer neural networks

Daniel Soudry, Elad Hoffer, 2017, ICLR.

Augment your batch: better training with larger batches

Daniel Soudry, Itay Hubara, Elad Hoffer, 2019, ArXiv.

Gradient Descent Monotonically Decreases the Sharpness of Gradient Flow Solutions in Scalar Networks and Beyond

Daniel Soudry, Y. Carmon, M. S. Nacson, 2023, ArXiv.

A fully analog memristor-based neural network with online gradient training

Daniel Soudry, Shahar Kvatinsky, Eyal Rosenthal, 2016, 2016 IEEE International Symposium on Circuits and Systems (ISCAS).

Statistical Testing for Efficient Out of Distribution Detection in Deep Neural Networks pdf

Daniel Soudry, Ruth Heller, Matan Haroush, 2021, ArXiv.

Optimal Fine-Grained N: M sparsity for Activations and Neural Gradients

Daniel Soudry, Itay Hubara, Ron Banner, 2022, ArXiv.

Neural gradients are lognormally distributed: understanding sparse and quantized training pdf

Brian Chmiel, Liad Ben-Uri, Moran Shkolnik, 2020, ArXiv.

A shotgun sampling solution for the common input problem in neural connectivity inference

L. Paninski, Daniel Soudry, G. Iyengar, 2013, 1309.3724.

Simultaneous Denoising, Deconvolution, and Demixing of Calcium Imaging Data

Timothy A. Machado, J. Merel, L. Paninski, 2016, Neuron.

Conductance-Based Neuron Models and the Slow Dynamics of Excitability

R. Meir, Daniel Soudry, 2012, Front. Comput. Neurosci..

Spiking input-output relation for general biophysical neuron models

Ron Meir, Daniel Soudry, R. Meir, 2012, 1207.6251.

The Implicit Bias of Minima Stability in Multivariate Shallow ReLU Networks

Daniel Soudry, T. Michaeli, Greg Ongie, 2023, ICLR.

Accurate Neural Training with 4-bit Matrix Multiplications at Standard Formats

Daniel Soudry, Elad Hoffer, Ron Banner, 2023, ICLR.

Training Binary Multilayer Neural Networks for Image Classification using Expectation Backpropagation pdf

Daniel Soudry, Zhiyong Cheng, Zhen-Zhong Lan, 2015, ArXiv.

Augment Your Batch: Improving Generalization Through Instance Repetition

Daniel Soudry, Itay Hubara, Elad Hoffer, 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).