EAST compresses deep ConvNets for tiny memory nodes.
problem Memory constraints in tiny devices for deep ConvNets.
method Encoding-Aware Sparse Training (EAST) with adaptive group pruning and LZ4 weight encoding.
result EAST achieves deep memory compression with lower sparsity and higher accuracy.
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.
SSVI efficiently trains sparse Bayesian neural networks with minimal compression and performance loss.
problem Efficiently training Bayesian neural networks with uncertainty quantification.
method SSVI optimizes a sparse subspace basis selection and its parameters alternately, guided by weight distribution statistics.
result SSVI achieves significant compression (10-20x model size reduction) with minimal performance drop (under 3%) and FLOPs reduction (up to 20x) compared to dense Variational Inference.
ADAPT method controls balance between sparse generator and discriminator in GANs.
problem High computational complexity in GANs training.
method Balanced Dynamic Sparse Training (ADAPT) method.
result ADAPT method achieves a good trade-off between performance and computational cost.
We demonstrate the possibility of what we call sparse learning: accelerated training of deep neural networks that maintain sparse weights throughout training while achieving dense performance levels. We accomplish this by developing sparse momentum, an algorithm which uses exponentially smoothed gradients (momentum) to…
Campfire simplifies sparse training for CNNs, achieving high accuracy with minimal pruning.
problem Training CNNs with sparse weights to maintain high accuracy and robustness.
method Gradual pruning technique with simplified sparsity structure, exploring various tradeoffs.
result Sparse ResNet-50 and ResNet-50 v1.5 achieve 75% top-1 accuracy with 70% sparsity, <1% accuracy loss.
The sizes of deep neural networks (DNNs) are rapidly outgrowing the capacity of hardware to store and train them. Research over the past few decades has explored the prospect of sparsifying DNNs before, during, and after training by pruning edges from the underlying topology. The resulting neural network is known as a …
SWAT reduces training efficiency and speed for CNNs.
problem Efficient training of neural networks on sparse hardware.
method SWAT modifies backpropagation to eliminate small weights and activations.
result SWAT reduces FLOPS by 80% on ResNet-50 for ImageNet, improving training speed by 3.3x.
Many applications require sparse neural networks due to space or inference time restrictions. There is a large body of work on training dense networks to yield sparse networks for inference, but this limits the size of the largest trainable sparse model to that of the largest trainable dense model. In this paper we int…
We analyze the computational complexity of Quantum Sparse Support Vector Machine, a linear classifier that minimizes the hinge loss and the L1 norm of the feature weights vector and relies on a quantum linear programming solver instead of a classical solver. Sparse SVM leads to sparse models that use only a small fr…
NeuroFabric proposes a method to optimize sparse network training topologies.
problem Long training times in deep neural networks due to high memory and compute requirements.
method Developed a new sparse neural network initialization scheme and evaluated various topologies.
result Identified a single optimal topology that maximizes accuracy across different datasets.
New pruning methods improve dynamic sparse training performance.
problem Improving dynamic sparse training performance.
method Design and empirical analysis of pruning criteria.
result Most pruning methods yield similar results, but magnitude-based pruning performs best in low-density regimes.
FedDST trains sparse sub-networks to improve efficiency in federated learning.
problem Efficiently train large models on resource-limited edge devices with limited network bandwidth.
method Dynamic sparse training to reduce model size and communication.
result FedDST outperforms existing methods in federated learning, especially in non-i.i.d. settings.
The sizes of deep neural networks (DNNs) are rapidly outgrowing the capacity of hardware to store and train them. Research over the past few decades has explored the prospect of sparsifying DNNs before, during, and after training by pruning edges from the underlying topology. The resulting neural network is known as a …
This work introduces a method to compare sparse neural network topologies using graph theory.
problem Comparing and understanding sparse neural network topologies, especially during training.
method Introducing Neural Network Sparse Topology Distance (NNSTD) to measure distances between different sparse neural networks.
result Sparse neural networks can outperform over-parameterized models without further structure optimization.
Inducing sparseness while training neural networks has been shown to yield models with a lower memory footprint but similar effectiveness to dense models. However, sparseness is typically induced starting from a dense model, and thus this advantage does not hold during training. We propose techniques to enforce sparsen…
SnAp approximates RTRL for online training of sparse recurrent networks.
problem Training large sparse recurrent networks online is computationally expensive.
method Sparse n-step Approximation (SnAp) of the RTRL influence matrix.
result SnAp with n=2 remains tractable for highly sparse networks and outperforms backpropagation through time.
We investigate the difficulties of training sparse neural networks and make new observations about optimization dynamics and the energy landscape within the sparse regime. Recent work of \citep{Gale2019, Liu2018} has shown that sparse ResNet-50 architectures trained on ImageNet-2012 dataset converge to solutions that a…
Graph neural networks have become increasingly popular in recent years due to their ability to naturally encode relational input data and their ability to scale to large graphs by operating on a sparse representation of graph adjacency matrices. As we look to scale up these models using custom hardware, a natural assum…
New algorithms reduce communication in GNN training.
problem Higher communication costs in GNNs due to sparse connectivity.
method Parallel algorithms for sparse-dense matrix multiplication.
result Asymptotic reduction in communication compared to previous methods.
Artificial Neural Networks (ANNs) have emerged as hot topics in the research community. Despite the success of ANNs, it is challenging to train and deploy modern ANNs on commodity hardware due to the ever-increasing model size and the unprecedented growth in the data volumes. Particularly for microarray data, the very-…
Improvements in the performance of deep neural networks have often come through the design of larger and more complex networks. As a result, fast memory is a significant limiting factor in our ability to improve network performance. One approach to overcoming this limit is the design of sparse neural networks, which ca…
We study the problem of estimating high dimensional models with underlying sparse structures while preserving the privacy of each training example. We develop a differentially private high-dimensional sparse learning framework using the idea of knowledge transfer. More specifically, we propose to distill the knowledge …
Sparse neural networks can be robust to adversarial attacks.
problem Finding robust neural networks that are also sparse.
method Empirical evaluation and analysis of the Lottery Ticket Hypothesis with adversarial training.
result Sparse, robust neural networks can be found using adversarial training.
Neuromorphic hardware tends to pose limits on the connectivity of deep networks that one can run on them. But also generic hardware and software implementations of deep learning run more efficiently for sparse networks. Several methods exist for pruning connections of a neural network after it was trained without conne…
Modern deep neural networks are typically highly overparameterized. Pruning techniques are able to remove a significant fraction of network parameters with little loss in accuracy. Recently, techniques based on dynamic reallocation of non-zero parameters have emerged, allowing direct training of sparse networks without…
New method uses randomized sparse neural networks to solve time-dependent PDEs more accurately and efficiently.
problem Numerical challenges in training neural networks sequentially in time to solve time-dependent PDEs.
method Introduces Neural Galerkin schemes that update randomized sparse subsets of network parameters at each time step.
result Up to two orders of magnitude more accurate and two orders of magnitude faster than dense update schemes.
Pruning at initialization fails to find sparse subnetworks, revealing information-theoretic barriers.
problem Difficulty in finding sparse subnetworks without training the full model.
method Analysis of effective parameter count and mutual information between sparsity mask and data.
result Pruning at initialization cannot find sparse subnetworks due to high mutual information.
New model leads to optimal test loss in sparse linear regression.
problem Sparse linear regression with low test loss despite interpolating training data.
method Developed a new parametrization of the model that combines benefits of ℓ1 and ℓ2 norms.
result Training via gradient descent leads to an interpolator with near-optimal test loss.
In this paper, we propose a semi-supervised dictionary learning method that uses both the information in labelled and unlabelled data and jointly trains a linear classifier embedded on the sparse codes. The manifold structure of the data in the sparse code space is preserved using the same approach as the Locally Linea…
New method learns sparse distributions by thresholding samples, improving performance and efficiency.
problem Sparse coding optimization in high-dimensional problems is computationally expensive and inefficient.
method Proposes a new variational sparse coding approach that learns sparse distributions by thresholding samples.
result Shows superior performance, statistical efficiency, and gradient estimation compared to other sparse distributions.
Paper develops efficient variational inference for sparse deep learning with theoretical guarantees.
problem Sparse deep learning's challenge of huge storage consumption and sparse structure recovery.
method Bayesian treatment with spike-and-slab priors and continuous relaxation of Bernoulli distribution for computationally efficient variational inferences.
result Provides variational posterior contraction rate, justifying consistency of the proposed method.
New method finds efficient sparse neural networks.
problem Finding efficient, sparse deep neural network models.
method Continuous Sparsification, approximating ℓ0 regularization. result Surpasses state-of-the-art for pruning and ticket search.
New method improves deep learning models robustness to label noise.
problem Improving deep learning models' robustness to corrupted labels.
method Sparse over-parameterization and implicit regularization.
result State-of-the-art test accuracy against label noise on various datasets.
Deep weight factorization improves neural network training through smooth optimization of sparse penalties.
problem Challenges in applying sparse regularization in neural networks due to non-differentiability of penalties.
method Introduces deep weight factorization, decomposing weights into multiple factors for smooth optimization of L1-penalized networks. result Deep weight factorization outperforms shallow factorization and pruning methods consistently across various architectures and datasets.
Law derived for neural networks with sparse connections.
problem Understanding the behavior of neural networks with sparse connections.
method Law of large numbers for empirical distribution of parameters derived.
result Law for neural networks with sparse connections derived.
Sparse routing networks with co-training prevent catastrophic forgetting in continual learning.
problem Catastrophic forgetting in neural networks trained on a sequence of tasks.
method Sparse routing networks with co-training to minimize interference between dissimilar tasks.
result Sparse routing networks with co-training outperform densely connected networks on benchmarks.
Improved trajectory prediction for team sports using sparse outputs.
problem Challenging to train deep learning models for player trajectory prediction.
method Sparse trajectory prediction and constant acceleration interpolation.
result Interpolation improves performance for all tested models.
Evidential Softmax preserves multimodality in sparse probability distributions for generative models.
problem Sparse probability distributions in deep generative models make exact marginalization computationally intractable.
method Introduce ev-softmax, a sparse normalization function that preserves multimodality and can be trained with probabilistic loss functions.
result ev-softmax outperforms existing techniques in distributional accuracy and dimensionality reduction.
Sparse coding has been popularly used as an effective data representation method in various applications, such as computer vision, medical imaging and bioinformatics, etc. However, the conventional sparse coding algorithms and its manifold regularized variants (graph sparse coding and Laplacian sparse coding), learn th…
Develops method to train classifiers on incomplete feature datasets.
problem Training classifiers on datasets with missing feature subsets.
method Simultaneous training of neural networks and sparse coding.
result Classifier trained on incomplete features correctly separates original data.
A new method selects features efficiently for high-dimensional data.
problem High computational costs and memory requirements in high-dimensional data.
method QuickSelection uses the strength of neurons in sparse autoencoders to select features.
result QuickSelection achieves the best trade-off of accuracy, speed, and memory usage.
Bayesian approach improves sparse PCE for high-dimensional problems.
problem Sparse PCE struggles with high-dimensional uncertainty and underdetermined situations.
method Joint shrinkage priors and MCMC for sparse PCE with uncertainty estimation.
result Bayesian PCE achieves sparse representations with higher polynomial degrees.
Lottery tickets find good initializations for IMP with sparse training.
problem Finding good initializations for iterative magnitude pruning (IMP) in sparse networks.
method Empirical study of IMP performance with varying pre-training data and iterations.
result Training on a small fraction of data suffices to obtain good initializations for IMP.
In presence of sparse noise we propose kernel regression for predicting output vectors which are smooth over a given graph. Sparse noise models the training outputs being corrupted either with missing samples or large perturbations. The presence of sparse noise is handled using appropriate use of ℓ1-norm along-wi…
Renormalized pruning improves neural network accuracy.
problem Over-parameterized neural networks waste many parameters.
method Propose renormalizing sparse neural networks.
result Renormalized pruning converges to zero error.
Powerpropagation makes neural networks inherently sparse.
problem Training sparse neural networks to reduce computational footprint and model size.
method Introduces a new weight-parameterisation technique exploiting gradient descent dynamics.
result Models trained with Powerpropagation have a higher density of zero weights, allowing for more efficient pruning.
Deep neural networks (DNNs) have emerged as key enablers of machine learning. Applying larger DNNs to more diverse applications is an important challenge. The computations performed during DNN training and inference are dominated by operations on the weight matrices describing the DNN. As DNNs incorporate more layers a…