The computation and memory needed for Convolutional Neural Network (CNN) inference can be reduced by pruning weights from the trained network. Pruning is guided by a pruning saliency, which heuristically approximates the change in the loss function associated with the removal of specific weights. Many pruning signals have been proposed, but the performance of each heuristic depends on the particular trained network. This leaves the data scientist with a difficult choice. When using any one saliency metric for the entire pruning process, we run the risk of the metric assumptions being invalidated, leading to poor decisions being made by the metric. Ideally we could combine the best aspects of different saliency metrics. However, despite an extensive literature review, we are unable to find any prior work on composing different saliency metrics. The chief difficulty lies in combining the numerical output of different saliency metrics, which are not directly comparable.
We propose a method to compose several primitive pruning saliencies, to exploit the cases where each saliency measure does well. Our experiments show that the composition of saliencies avoids many poor pruning choices identified by individual saliencies. In most cases our method finds better selections than even the best individual pruning saliency.
Convolutional Neural Networks (CNNs) compression is crucial to deploying these models in edge devices with limited resources. Existing channel pruning algorithms for CNNs have achieved plenty of success on complex models. They approach the pruning problem from various perspectives and use different metrics to guide the pruning process. However, these metrics mainly focus on the model's `outputs' or `weights' and neglect its `interpretations' information. To fill in this gap, we propose to address the channel pruning problem from a novel perspective by leveraging the interpretations of a model …
Structural pruning can simplify network architecture and improve inference speed. We propose Hardware-Aware Latency Pruning (HALP) that formulates structural pruning as a global resource allocation optimization problem, aiming at maximizing the accuracy while constraining latency under a predefined budget on targeting device. For filter importance ranking, HALP leverages latency lookup table to track latency reduction potential and global saliency score to gauge accuracy drop. Both metrics can be evaluated very efficiently during pruning, allowing us to reformulate global structural pruning un…
Predicting human fixations from images has recently seen large improvements by leveraging deep representations which were pretrained for object recognition. However, as we show in this paper, these networks are highly overparameterized for the task of fixation prediction. We first present a simple yet principled greedy pruning method which we call Fisher pruning. Through a combination of knowledge distillation and Fisher pruning, we obtain much more runtime-efficient architectures for saliency prediction, achieving a 10x speedup for the same AUC performance as a state of the art network on the…
SCOPE: Saliency-Coverage Oriented Token Pruning for Efficient Multimodel LLMs
University of Electronic Science and Technology of China（电子科技大学）；Shenzhen Institute for Advanced Study（深圳先进研究 institute）；CFAR, Agency for Science, Technology and Research (A*STAR)（科技研究局（A*STAR）认知与人工智能研究中心）；IHPC, Agency for Science, Technology and Research (A*STAR)（科技研究局（A*STAR）人工智能中心）
NVIDIA（英伟达）；University of California, Berkeley（加州大学伯克利分校）；Duke University（杜克大学）
Transformers yield state-of-the-art results across many tasks. However, their heuristically designed architecture impose huge computational costs during inference. This work aims on challenging the common design philosophy of the Vision Transformer (ViT) model with uniform dimension across all the stacked blocks in a model stage, where we redistribute the parameters both across transformer blocks and between different structures within the block via the first systematic attempt on global structural pruning. Dealing with diverse ViT structural components, we derive a novel Hessian-based structu…
Shanghai Jiao Tong University（上海交通大学）；Mila - Québec AI Institute（Mila-魁北克人工智能研究所）
Deep convolutional neural networks have been proved successful on a wide range of tasks, yet they are still hindered by their large computation cost in many industrial scenarios. In this paper, we propose to reduce such cost for CNNs through a self-adaptive network pruning method (SANP). Our method introduces a general Saliency-and-Pruning Module (SPM) for each convolutional layer, which learns to predict saliency scores and applies pruning for each channel. Given a total computation budget, SANP adaptively determines the pruning strategy with respect to each layer and each sample, such that t…
University of Oxford（牛津大学）；Australian National University（澳大利亚国立大学）
Network pruning is a promising avenue for compressing deep neural networks. A typical approach to pruning starts by training a model and then removing redundant parameters while minimizing the impact on what is learned. Alternatively, a recent approach shows that pruning can be done at initialization prior to training, based on a saliency criterion called connection sensitivity. However, it remains unclear exactly why pruning an untrained, randomly initialized neural network is effective. In this work, by noting connection sensitivity as a form of gradient, we formally characterize initializat…
Structural pruning can simplify network architecture and improve inference speed. We propose Hardware-Aware Latency Pruning (HALP) that formulates structural pruning as a global resource allocation optimization problem, aiming at maximizing the accuracy while constraining latency under a predefined budget. For filter importance ranking, HALP leverages latency lookup table to track latency reduction potential and global saliency score to gauge accuracy drop. Both metrics can be evaluated very efficiently during pruning, allowing us to reformulate global structural pruning under a reward maximiz…
Pruning large neural networks while maintaining their performance is often desirable due to the reduced space and time complexity. In existing methods, pruning is done within an iterative optimization procedure with either heuristically designed pruning schedules or additional hyperparameters, undermining their utility. In this work, we present a new approach that prunes a given network once at initialization prior to training. To achieve this, we introduce a saliency criterion based on connection sensitivity that identifies structurally important connections in the network for the given task.…
In this paper, we propose an approach for filter-level pruning with hierarchical knowledge distillation based on the teacher, teaching-assistant, and student framework. Our method makes use of teaching assistants at intermediate pruning levels that share the same architecture and weights as the target student. We propose to prune each model independently using the gradient information from its corresponding teacher. By considering the relative sizes of each student-teacher pair, this formulation provides a natural trade-off between the capacity gap for knowledge distillation and the bias of th…
One-Cycle Structured Pruning via Stability-Driven Subnetwork Search
IT Application Research Center, Korea Electronics Technology Institute（韩国电子技术研究所应用信息技术中心）；Department of Intelligent Semiconductors, Soongsil University（顺慈大学智能半导体系）；Department of Computer Science & Engineering, Seoul National University（首尔国立大学计算机科学与工程系）
Channel pruning is used to reduce the number of weights in a Convolutional Neural Network (CNN). Channel pruning removes slices of the weight tensor so that the convolution layer remains dense. The removal of these weight slices from a single layer causes mismatching number of feature maps between layers of the network. A simple solution is to force the number of feature map between layers to match through the removal of weight slices from subsequent layers. This additional constraint becomes more apparent in DNNs with branches where multiple channels need to be pruned together to keep the net…
Fudan University（复旦大学）；Shanghai AI Laboratory（上海人工智能实验室）
Global channel pruning (GCP) aims to remove a subset of channels (filters) across different layers from a deep model without hurting the performance. Previous works focus on either single task model pruning or simply adapting it to multitask scenario, and still face the following problems when handling multitask pruning: 1) Due to the task mismatch, a well-pruned backbone for classification task focuses on preserving filters that can extract category-sensitive information, causing filters that may be useful for other tasks to be pruned during the backbone pruning stage; 2) For multitask predic…
Deep neural network (DNN) pruning has become a de facto component for deploying on resource-constrained devices since it can reduce memory requirements and computation costs during inference. In particular, channel pruning gained more popularity due to its structured nature and direct savings on general hardware. However, most existing pruning approaches utilize importance measures that are not directly related to the task utility. Moreover, few in the literature focus on visual detection models. To fill these gaps, we propose a novel gradient-based saliency measure for visual detection and us…
Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）；Harbin Institute of Technology（哈尔滨工业大学）；Tsinghua University（清华大学）；The University of Hong Kong（香港大学）
City University of Hong Kong（香港城市大学）；Baidu USA LLC（百度美国有限责任公司）
Model pruning aims to reduce the deep neural network (DNN) model size or computational overhead. Traditional model pruning methods such as l-1 pruning that evaluates the channel significance for DNN pay too much attention to the local analysis of each channel and make use of the magnitude of the entire feature while ignoring its relevance to the batch normalization (BN) and ReLU layer after each convolutional operation. To overcome these problems, we propose a new model pruning method from a new perspective of gradient flow in this paper. Specifically, we first theoretically analyze the channe…
CAS Key Laboratory of Technology in Geo-spatial Information Processing and Application System, University of Science and Technology of China（中国科学技术大学 中国科学院空间信息处理与应用系统技术重点实验室）；KDDI Research（KDDI研究所）
Accelerating the inference speed of CNNs is critical to their deployment in real-world applications. Among all the pruning approaches, those implementing a sparsity learning framework have shown to be effective as they learn and prune the models in an end-to-end data-driven manner. However, these works impose the same sparsity regularization on all filters indiscriminately, which can hardly result in an optimal structure-sparse network. In this paper, we propose a Saliency-Adaptive Sparsity Learning (SASL) approach for further optimization. A novel and effective estimation of each filter, i.e.…
Adversarial examples provoke weak reliability and potential security issues in deep neural networks. Although adversarial training has been widely studied to improve adversarial robustness, it works in an over-parameterized regime and requires high computations and large memory budgets. To bridge adversarial robustness and model compression, we propose a novel adversarial pruning method, Masking Adversarial Damage (MAD) that employs second-order information of adversarial loss. By using it, we can accurately estimate adversarial saliency for model parameters and determine which parameters can …
University of Washington（华盛顿大学）；University of Maryland（马里兰大学）；New York University（纽约大学）
Conventional saliency maps highlight input features to which neural network predictions are highly sensitive. We take a different approach to saliency, in which we identify and analyze the network parameters, rather than inputs, which are responsible for erroneous decisions. We find that samples which cause similar parameters to malfunction are semantically similar. We also show that pruning the most salient parameters for a wrongly classified sample often improves model behavior. Furthermore, fine-tuning a small number of the most salient parameters on a single sample results in error correct…
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）
IBM Research Zurich（苏黎世IBM研究公司）；EPFL Lausanne（洛桑联邦理工学院）；ETH Zurich（苏黎世联邦理工学院）；Fondazione Pascale（帕斯卡尔基金会）；University of Bern（伯尔尼大学）；Uppsala University（乌普萨拉大学）
Explainability of deep learning methods is imperative to facilitate their clinical adoption in digital pathology. However, popular deep learning methods and explainability techniques (explainers) based on pixel-wise processing disregard biological entities' notion, thus complicating comprehension by pathologists. In this work, we address this by adopting biological entity-based graph processing and graph explainers enabling explanations accessible to pathologists. In this context, a major challenge becomes to discern meaningful explainers, particularly in a standardized and quantifiable fashio…
