Net-Trim: A Layer-wise Convex Pruning of Deep Neural Networks
نویسندگان
چکیده
and quantum settings Model reduction is a highly desirable process for deep neural networks. While large networks are theoretically capable of learning arbitrarily complex models, overfitting and model redundancy negatively affects the prediction accuracy and model variance. Net-Trim is a layer-wise convex framework to prune (sparsify) deep neural networks. The method is applicable to neural networks operating with the rectified linear unit (ReLU) as the nonlinear activation. The basic idea is to retrain the network layer by layer keeping the layer inputs and outputs close to the originally trained model, while seeking a sparse transform matrix. We present both the parallel and cascade versions of the algorithm. While the former enjoys computational distributability, the latter is capable of achieving simpler models. In both cases, we mathematically show a consistency between the retrained model and the initial trained network. We also derive the general sufficient conditions for the recovery of a sparse transform matrix. In the case of standard Gaussian training samples of dimension N being fed to a layer, and s being the maximum number of nonzero terms across all columns of the transform matrix, we show that O(s.logN) samples are enough to accurately learn the layer model.
منابع مشابه
Net-Trim: Convex Pruning of Deep Neural Networks with Performance Guarantee
Model reduction is a highly desirable process for deep neural networks. While large networks are theoretically capable of learning arbitrarily complex models, overfitting and model redundancy negatively affects the prediction accuracy and model variance. NetTrim is a layer-wise convex framework to prune (sparsify) deep neural networks. The method is applicable to neural networks operating with ...
متن کاملLearning to Prune Deep Neural Networks via Layer-wise Optimal Brain Surgeon
How to develop slim and accurate deep neural networks has become crucial for realworld applications, especially for those employed in embedded systems. Though previous work along this research line has shown some promising results, most existing methods either fail to significantly compress a well-trained deep network or require a heavy retraining process for the pruned deep network to re-boost...
متن کاملLayer-wise Relevance Propagation for Deep Neural Network Architectures
We present the application of layer-wise relevance propagation to several deep neural networks such as the BVLC reference neural net and googlenet trained on ImageNet and MIT Places datasets. Layerwise relevance propagation is a method to compute scores for image pixels and image regions denoting the impact of the particular image region on the prediction of the classifier for one particular te...
متن کاملNeuron Pruning for Compressing Deep Networks Using Maxout Architectures
This paper presents an efficient and robust approach for reducing the size of deep neural networks by pruning entire neurons. It exploits maxout units for combining neurons into more complex convex functions and it makes use of a local relevance measurement that ranks neurons according to their activation on the training set for pruning them. Additionally, a parameter reduction comparison betwe...
متن کاملمعرفی شبکه های عصبی پیمانه ای عمیق با ساختار فضایی-زمانی دوگانه جهت بهبود بازشناسی گفتار پیوسته فارسی
In this article, growable deep modular neural networks for continuous speech recognition are introduced. These networks can be grown to implement the spatio-temporal information of the frame sequences at their input layer as well as their labels at the output layer at the same time. The trained neural network with such double spatio-temporal association structure can learn the phonetic sequence...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- CoRR
دوره abs/1611.05162 شماره
صفحات -
تاریخ انتشار 2016