Pruned neural networks' error scales predictably with architecture and task.
problem Understanding the predictability of pruning across different scales and architectures.
method Functionally approximated the error of pruned networks, showing it is predictable in terms of invariant tying width, depth, and pruning level.
result The error of pruned networks follows a scaling law with interpretable coefficients that depend on architecture and task.
Neural architecture search (NAS) is gaining more and more attention in recent years due to its flexibility and remarkable capability to reduce the burden of neural network design. To achieve better performance, however, the searching process usually costs massive computations that might not be affordable for researcher…
Gradient-based method prunes large models to create transferable architectures.
problem Creating transferable architectures from large models with limited fine-tuning data.
method Gradient-based algorithm for architecture pruning and subset selection.
result Successfully retrain architectures on new tasks with few fine-tuning data.
Structured pruning is a popular method for compressing a neural network: given a large trained network, one alternates between removing channel connections and fine-tuning; reducing the overall width of the network. However, the efficacy of structured pruning has largely evaded scrutiny. In this paper, we examine ResNe…
Network pruning is widely used for reducing the heavy inference cost of deep models in low-resource settings. A typical pruning algorithm is a three-stage pipeline, i.e., training (a large model), pruning and fine-tuning. During pruning, according to a certain criterion, redundant weights are pruned and important weigh…
New method grows deep networks efficiently by dynamically pruning and growing layers.
problem Training deep networks is computationally expensive and inefficient.
method Structured continuous sparsification starting from a small seed architecture.
result 49.7% inference FLOPs and 47.4% training FLOPs savings with 75.2% top-1 accuracy.
New pruning methods improve energy efficiency of neural networks.
problem Energy-efficient neural networks for devices with limited resources.
method Magnitude and Gradient based pruning at initialization and training of sparse architectures.
result Proposed novel pruning methods prevent full layer pruning and improve training.
Pruning FCNs reveals sub-networks that match CNNs' performance.
problem Understanding the inductive bias of pruning in neural networks.
method Iterative magnitude pruning of a simple FCN followed by analysis of the resulting architecture.
result Pruned FCNs exhibit key features of CNNs, suggesting new architectural biases.
Lookahead pruning extends single-layer optimization to multi-layer, outperforming magnitude-based pruning.
problem Pruning neural networks to reduce computational cost and memory usage.
method Developed a multi-layer optimization approach extending the single-layer optimization of magnitude-based pruning.
result Consistently outperforms magnitude-based pruning on various networks, especially in high sparsity.
A method to reduce memory usage in NAS by pruning the search space.
problem High GPU memory consumption in One-Shot NAS techniques.
method Utilising Zero-Shot NAS to prune the search space before applying One-Shot NAS.
result Reduces memory consumption by 81% while maintaining accuracy.
Pruning is one of the most effective model reduction techniques. Deep networks require massive computation and such models need to be compressed to bring them on edge devices. Most existing pruning techniques are focused on vision-based models like convolutional networks, while text-based models are still evolving. The…
i-SpaSP prunes neural networks by identifying important groups of parameters, improving pruning efficiency.
problem Pruning neural networks to reduce computational cost and improve performance.
method i-SpaSP uses sparse signal recovery principles to iteratively identify and threshold important parameter groups.
result i-SpaSP achieves strong empirical results and theoretical convergence guarantees, improving pruning efficiency.
A new pruning method improves neural network efficiency and accuracy.
problem Optimizing neural network efficiency and accuracy through pruning.
method Formulated as a Knapsack Problem, the method optimizes trade-off between neuron importance and computational cost. Channels are pruned while maintaining network structure, and fine-tuned using inner knowledge distillation from parent network.
result State-of-the-art pruning results on ImageNet, CIFAR-10, and CIFAR-100.
Dirichlet pruning compresses neural networks by removing unimportant units.
problem Compressing large neural network models without sacrificing performance.
method Assigns Dirichlet distribution over network layers' units and uses variational inference to estimate parameters.
result Achieves state-of-the-art compression performance on larger architectures like VGG and ResNet.
LTP learns per-layer thresholds for efficient pruning of deep networks.
problem Efficiently pruning deep neural networks to reduce computational cost and size.
method LTP learns thresholds via gradient descent, making pruning computationally efficient and scalable.
result LTP achieves competitive compression rates and maintains high accuracy on ImageNet networks.
One-shot neural architecture search limits depth search space and prunes networks for better performance and uncertainty.
problem Finding optimal depth in residual networks for efficient training and inference.
method Formulated a variational objective to approximate the depth distribution and pruned networks based on this distribution.
result Pruned networks achieve competitive accuracy with unpruned networks and better uncertainty calibration.
Dynamic Sparse Training finds efficient sparse networks from scratch.
problem Finding efficient sparse neural networks.
method Jointly optimizes network parameters and sparsity with trainable thresholds.
result Achieves state-of-the-art performance with minimal performance loss.
New statistical mechanics analysis shows edge pruning outperforms node pruning in neural networks.
problem Theoretical understanding of neural network pruning effectiveness is lacking.
method Statistical mechanics analysis of a teacher-student framework.
result DPP node pruning method is superior to other methods, but edge pruning is better overall.
This paper shows RBMs can maintain strong performance even after extreme pruning, but only if done early in training.
problem The computational and environmental costs of large neural networks.
method Investigating the performance of RBMs under extreme pruning conditions, inspired by the Lottery Ticket Hypothesis.
result RBMs can achieve high-quality generative performance even after 80% pruning, but performance degrades sharply above a critical point.
Paper combines regularization and pruning to reduce FLOPs in DNNs.
problem Efficient inference in deep learning models.
method Apply mixup and cutout regularizations and soft filter pruning to the ResNet architecture.
result Combining regularization and pruning reduces FLOPs more effectively than each technique alone.
Study finds differences in LTs across tasks and architectures, proposing a consensus-based method for generating refined lottery tickets.
problem Understanding the variability and uniqueness of Lottery Tickets across different image classification tasks and architectures.
method 28 combinations of image classification tasks and architectures, iterative pruning techniques, consensus-based method for generating refined lottery tickets.
result Disproves the uniqueness of Lottery Tickets and connects emergent mask structure to the choice of pruning.
FlipOut prunes neural networks by flipping weights' signs, achieving high sparsity.
problem Redundant weights in neural networks increase training time and resource usage.
method Uses sign flips during training to determine weight saliency for pruning.
result Competitive with existing methods, achieving state-of-the-art performance for high sparsity.
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,…
Paper proposes efficient pruning method for neural networks.
problem Compressing deep neural networks for resource-constrained devices.
method Adaptive sparsity loss for budget-aware optimization during training.
result Demonstrated effectiveness on various architectures and datasets.
PCNN prunes CNN weights efficiently for hardware acceleration.
problem Efficiently compressing CNN models for hardware acceleration.
method PCNN uses a novel Sparsity Pattern Mask (SPM) to encode and prune weights.
result PCNN achieves up to 8.4X compression with minimal accuracy loss.
New pruning method captures global correlations for efficient neural network inference.
problem Efficiently pruning neural networks for faster inference and reduced memory usage.
method Second-order structured pruning (SOSP-H) with innovative saliency-based approaches.
result SOSP-H scales to large-scale vision tasks and improves accuracy without compromising efficiency.
Predict accuracy of neural architectures using non-neural models.
problem Improving efficiency and effectiveness of neural architecture search.
method Used gradient boosting decision tree (GBDT) for accuracy prediction and pruned the search space based on features learned from GBDT.
result GBDT-based predictor achieves comparable or better accuracy than neural network-based predictors and is 22x more sample efficient on NASBench-101.
Greedy pruning reduces neural networks by a logarithmic number of tickets, improving accuracy.
problem Pruning large neural networks to reduce size while maintaining accuracy.
method Greedy optimization-based pruning method with exponential decay guarantee.
result The discrepancy between pruned and original networks decays exponentially with network size.
We present a provable, sampling-based approach for generating compact Convolutional Neural Networks (CNNs) by identifying and removing redundant filters from an over-parameterized network. Our algorithm uses a small batch of input data points to assign a saliency score to each filter and constructs an importance sampli…
Mixed integer programming identifies critical neurons in neural networks.
problem Identifying neurons critical for network performance and generalization.
method Developed a mixed integer program (MIP) to assign importance scores to neurons, guiding pruning decisions.
result The method identifies multiple 'lucky' sub-networks resulting in optimized architectures that generalize across datasets.
Recent advances in the sparse neural network literature have made it possible to prune many large feed forward and convolutional networks with only a small quantity of data. Yet, these same techniques often falter when applied to the problem of recovering sparse recurrent networks. These failures are quantitative: when…
Automatic methods for Neural Architecture Search (NAS) have been shown to produce state-of-the-art network models. Yet, their main drawback is the computational complexity of the search process. As some primal methods optimized over a discrete search space, thousands of days of GPU were required for convergence. A rece…
Deep neural networks are typically too computationally expensive to run in real-time on consumer-grade hardware and low-powered devices. In this paper, we investigate reducing the computational and memory requirements of neural networks through network pruning and quantisation. We examine their efficacy on large networ…
RocketStack integrates predictions from multiple base learners using a recursive stacking architecture up to ten levels.
problem Feature redundancy, complexity, and computational burden in deep stacking.
method Level-aware recursive stacking with pruning and compression techniques.
result Increasing accuracy with depth and outperforming standalone ensembles at later levels.
A new method prunes neural networks efficiently without losing effectiveness.
problem Efficient pruning of neural networks without sacrificing performance.
method Deterministic approximation of binary gates and L0 regularization. result Pruning neural networks significantly without loss in effectiveness.
Galen algorithm compresses neural networks for specific hardware with reduced latency.
problem Finding optimal compression policies for neural networks on specific hardware.
method Reinforcement learning using pruning and quantization to optimize inference latency.
result Compressed ResNet18 for ARM processor reduced inference latency by 80%.
Accelerating DNN execution on various resource-limited computing platforms has been a long-standing problem. Prior works utilize l1-based group lasso or dynamic regularization such as ADMM to perform structured pruning on DNN models to leverage the parallel computing architectures. However, both of the pruning dimensio…
Randomly initialized networks can perform as well as pruned networks.
problem Understanding and improving network pruning methods.
method Sanity checks on recent pruning methods, proposing random tickets.
result Randomly initialized networks can perform as well as pruned networks.
CPOT prunes deep networks by identifying redundant filters using optimal transport.
problem Redundant filters in deep neural networks make models hard to deploy on resource-limited platforms.
method CPOT uses optimal transport to find the mean of channel distributions, pruning redundant information.
result CPOT outperforms state-of-the-art methods in pruning ResNet models and image-to-image translation tasks.
Dynamic sample pruning speeds up spatio-temporal forecasting models.
problem Training deep learning models on large, redundant datasets is computationally expensive.
method Dynamic sample pruning based on real-time learning state.
result Significant acceleration of training speed with improved performance.
Proposes dynamic channel pruning during neural network training.
problem Pruning neural networks during training to reduce computational cost and improve efficiency.
method Dynamic channel propagation to update channel utility values and selectively prune channels.
result Our scheme trains and prunes neural networks simultaneously, achieving superior performance.
Despite the promising results of convolutional neural networks (CNNs), their application on devices with limited resources is still a big challenge; this is mainly due to the huge memory and computation requirements of the CNN. To counter the limitation imposed by the network size, we use pruning to reduce the network …
ResRep prunes CNNs without losing accuracy by separating remembering and forgetting.
problem Pruning CNNs to reduce FLOPs without sacrificing accuracy.
method Decoupling remembering and forgetting in CNNs, using SGD for remembering and a novel update rule for forgetting.
result Achieved lossless pruning with high compression ratio (76.15% accuracy on ImageNet with 45% FLOPs reduction).
The study finds a theoretical bound for pre-training iterations needed for pruning to yield good subnetwork performance.
problem Discovering efficient subnetworks within pre-trained dense networks.
method Mathematical analysis of a two-layer, fully-connected network, validating with a multi-layer perceptron trained on MNIST.
result A logarithmically dependent threshold on dataset size for successful pruning.
We present a filter pruning approach for deep model compression, using a multitask network. Our approach is based on learning a a pruner network to prune a pre-trained target network. The pruner is essentially a multitask deep neural network with binary outputs that help identify the filters from each layer of the orig…
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…
Paper proposes neural network for efficient MIMO channel estimation and pilot reduction.
problem High overhead from pilot transmission in wideband MIMO systems.
method Neural network architecture for frequency-aware pilot design and channel estimation, with pruning technique.
result Neural network outperforms linear minimum mean square error (LMMSE) estimation.
Neural networks have achieved dramatic improvements in recent years and depict the state-of-the-art methods for many real-world tasks nowadays. One drawback is, however, that many of these models are overparameterized, which makes them both computationally and memory intensive. Furthermore, overparameterization can als…