Sparseout controls sparsity in neural networks, improving performance in language modeling.
problem Controlling sparsity in neural networks to optimize performance.
method Sparseout is a variant of Dropout that controls sparsity, theoretically proven and empirically validated.
result Sparseout controls the desired level of sparsity in neural network activations, improving performance in language modeling.
This paper studies activation sparsity in large language models, finding key trends and implications.
problem Activation sparsity in large language models (LLMs) can be improved for efficiency and interpretability.
method Proposes PPL-p% sparsity, analyzes trends with training data, width-depth ratio, and parameter scale. result ReLU is more efficient for sparsity than SiLU, and deeper architectures can improve sparsity.
Sparsity helps reduce diffusion model costs.
problem High computational costs in diffusion models.
method Introduced sparsity concept to reduce input dimensionality.
result Sparsity reduces computational complexity to intrinsic data dimension.
Our work is focused on the joint sparsity recovery problem where the common sparsity pattern is corrupted by Poisson noise. We formulate the confidence-constrained optimization problem in both least squares (LS) and maximum likelihood (ML) frameworks and study the conditions for perfect reconstruction of the original r…
Sparse activations in neural models correlate with frequent words, suggesting sparsity is natural.
problem Interpretability and resource efficiency in neural language models.
method Used the Taxi-Euclidean norm to measure sparsity and analyzed gradients and activations of frequent words.
result Frequent input words are associated with sparse activations, while frequent target words are associated with dispersed activations.
Embeds sparsity in deep neural networks, allowing exact zero parameters.
problem Learning sparse structures in deep networks.
method Embeds sparsity into neural network structure, allowing exact zero parameters during training.
result Can learn both structured and unstructured sparsity.
The paper studies how regularization parameters affect sparsity in deep neural networks.
problem Reducing the complexity of deep neural networks by promoting sparsity.
method Derives ℓ1-norm sparsity-promoting models, characterizes sparsity levels, and develops algorithms for selecting optimal regularization parameters. result Developed algorithms to select regularization parameters for desired sparsity levels in neural networks.
Spark Transformer achieves high sparsity in FFN and attention without sacrificing model quality.
problem Inactivity of most neurons in FFN of Transformers, leading to inefficiency.
method Top-k masking, statistical top-k algorithm, and predictor for activated entries.
result Only 8% of FFN neurons are activated, achieving up to 2.5x wall-time speedup.
New method solves graph-structured sparsity problems efficiently.
problem Graph-structured sparsity optimization in complex models.
method Stochastic gradient-based approach for non-convex graph-structured sparsity.
result Linear convergence up to a constant error.
Spatially-sparse predictors are good models for brain decoding: they give accurate predictions and their weight maps are interpretable as they focus on a small number of regions. However, the state of the art, based on total variation or graph-net, is computationally costly. Here we introduce sparsity in the local neig…
Compressive sensing (CS) exploits sparsity to recover sparse or compressible signals from dimensionality reducing, non-adaptive sensing mechanisms. Sparsity is also used to enhance interpretability in machine learning and statistics applications: While the ambient dimension is vast in modern data analysis problems, the…
New method enforces encoder sparsity in HPF for more interpretable feature selection.
problem Lack of encoder sparsity in HPF leads to lack of column-clustering property.
method Enforces encoder sparsity using a generalized additive model (GAM).
result Gains ability to perform feature selection and relates each representation to original features.
STR reparameterizes DNN weights with soft thresholds for better sparsity and accuracy.
problem Improving sparsity in DNNs for better accuracy and lower inference cost.
method Soft Threshold Reparameterization (STR) using the soft-threshold operator on DNN weights.
result STR achieves state-of-the-art accuracy and reduces FLOPs by up to 50%.
New method targets sparsity to prevent overfitting in deep nets.
problem Overfitting in deep neural networks with small datasets.
method Targeted sparsity regularization to visualize and counteract overfitting.
result Significant increase in image classification performance without overfitting.
Many natural signals exhibit a sparse representation, whenever a suitable describing model is given. Here, a linear generative model is considered, where many sparsity-based signal processing techniques rely on such a simplified model. As this model is often unknown for many classes of the signals, we need to select su…
New methods solve graph sparsity optimization problems faster.
problem Complex graph sparsity optimization problems in disease outbreak monitoring and social network analysis.
method Stochastic variance-reduced gradient-based methods GraphSVRG-IHT and GraphSCSG-IHT.
result Our methods achieve linear convergence speed.
Paper proposes a new sparsity scheme for high-dimensional VAR models.
problem Estimation of high-dimensional VAR models with sparsity assumptions.
method Regularized estimation procedures for sparse VAR models.
result Threholding extends consistency properties of regularized estimators.
Paper studies binary random projections with controllable sparsity patterns for computational and accuracy advantages.
problem Improving computational efficiency and accuracy in random projections.
method Proposes two sparse binary projection models with controllable sparsity patterns.
result Significant computational advantages and improved accuracies in empirical evaluations.
New measure SEV shows non-sparse models can still have low decision sparsity.
problem Non-sparse models can still make accurate decisions based on a few features.
method Introduced Sparse Explanation Value (SEV) to measure decision sparsity, not overall model sparsity.
result Many non-sparse models have low decision sparsity, as measured by SEV.
In this paper, we investigate a new compressive sensing model for multi-channel sparse data where each channel can be represented as a hierarchical tree and different channels are highly correlated. Therefore, the full data could follow the forest structure and we call this property as \emph{forest sparsity}. It exploi…
Working in high-dimensional latent spaces, the internal encoding of data in Variational Autoencoders becomes naturally sparse. We discuss this known but controversial phenomenon sometimes refereed to as overpruning, to emphasize the under-use of the model capacity. In fact, it is an important form of self-regularizatio…
DeepHoyer introduces differentiable, scale-invariant sparsity measures for neural networks.
problem Efficiently sparsifying neural networks with scale-invariant sparsity measures.
method Developed DeepHoyer, a set of differentiable, scale-invariant sparsity-inducing regularizers based on the Hoyer measure.
result DeepHoyer produces sparser neural networks than previous methods, maintaining similar accuracy.
Study finds economic data may not be as sparse as previously thought.
problem Modeling economic relations with many variables and prior sensitivity issues.
method Bayesian approach with Spike-and-Slab prior to evaluate variable selection and shrinkage.
result Prior distribution affects detection of sparsity patterns in economic data.
Adversarial robustness improved by sparsity in network weights.
problem How sparsity affects adversarial robustness in neural networks.
method Theoretical proof and experimental validation of adversarial pruning methods.
result Weights sparsity improves adversarial robustness, especially through inheritance from smaller networks.
New Bayesian method for joint sparse parameter inference.
problem Inference of jointly sparse parameter vectors from multiple measurements.
method Hierarchical Bayesian learning with joint sparsity-promoting priors.
result New algorithms consistently outperform existing methods in numerical experiments.
MuVI models multi-view data with structured sparsity, integrating domain knowledge.
problem Disentangling variation across multiple data views in complex systems.
method Multi-view latent variable model with structured sparsity using a modified horseshoe prior.
result MuVI outperforms state-of-the-art methods in structured sparsity modeling and integrates noisy domain expertise.
Paper introduces structured sparsity estimators for Generalized Linear Models.
problem Estimating structured sparsity in GLMs with debiased estimators.
method Extends Stucky and van de Geer's results to GLMs with structured sparsity.
result Proves oracle inequalities for structured sparsity estimators in GLMs.
New sparsity operator reduces variance reduction methods' computational cost.
problem Reduce computational cost of variance reduction methods.
method Introduce random-top-k operator to estimate gradient sparsity and reduce operations per update.
result Our algorithm consistently outperforms SpiderBoost in various tasks.
HALO learns to prune neural networks by adaptively shrinking weights.
problem Sparsity and model size in deep neural networks.
method Bayesian hierarchical models and trainable parameters for adaptive sparsification.
result HALO learns to create highly sparse networks with significant performance gains.
We propose a new sparsity-smoothness penalty for high-dimensional generalized additive models. The combination of sparsity and smoothness is crucial for mathematical theory as well as performance for finite-sample data. We present a computationally efficient algorithm, with provable numerical convergence properties, fo…
New layer sparsity concept improves neural networks.
problem Improving neural network efficiency and interpretability.
method Formulated layer sparsity, introduced regularization and refitting schemes.
result Generated more compact and accurate neural networks.
Study the effects of data parallelism and sparsity on neural network training.
problem Understanding the effects of data parallelism and sparsity on neural network training.
method Conducted extensive experiments and developed a theoretical analysis.
result Found a general scaling trend between batch size and number of training steps to convergence for the effect of data parallelism, and difficulty of training under sparsity.
Bayesian framework for encoding uncertainty and inducing sparsity.
problem Handling uncertainty and inducing sparsity in statistical models.
method General Bayesian framework with explicit encoding of uncertainty and sparsity-inducing approach.
result Effective in linear and logistic regression, and Bayesian neural networks.
Sparsity-promoting priors have become increasingly popular over recent years due to an increased number of regression and classification applications involving a large number of predictors. In time series applications where observations are collected over time, it is often unrealistic to assume that the underlying spar…
Paper proposes a unified sparsity-based framework for evaluating algorithmic fairness.
problem Ensuring fairness in machine learning across diverse domains.
method Unified sparsity-based framework for evaluating fairness.
result Demonstrates broad applicability and effectiveness of the framework.
In this paper, we consider the block-sparse signals recovery problem in the context of multiple measurement vectors (MMV) with common row sparsity patterns. We develop a new method for recovery of common row sparsity MMV signals, where a pattern-coupled hierarchical Gaussian prior model is introduced to characterize bo…
New model allows some connections to be zero, improving network analysis.
problem Networks with block structure and sparsity.
method Sparse Popularity Adjusted Stochastic Block Model (PABM).
result Allows some probabilities of connections to be zero.
Deep learning is finding its way into the embedded world with applications such as autonomous driving, smart sensors and aug- mented reality. However, the computation of deep neural networks is demanding in energy, compute power and memory. Various approaches have been investigated to reduce the necessary resources, on…
This paper describes a simple framework for structured sparse recovery based on convex optimization. We show that many structured sparsity models can be naturally represented by linear matrix inequalities on the support of the unknown parameters, where the constraint matrix has a totally unimodular (TU) structure. For …
LEWIS merges LLMs without training, improving performance on specific tasks.
problem Limited performance improvement of merged models on specific benchmarks.
method Guided model merging using layer-wise sparsity and task-vector pruning.
result Improved model performance by up to 11.3% on math-solving tasks.
Improves robustness of information bottleneck framework with sparsity-inducing prior.
problem Fixed-dimensional priors restrict flexibility and restrict robustness.
method Sparsity-inducing spike-slab categorical prior that learns dimension distribution per data point.
result Improves accuracy and robustness compared to traditional priors and other methods.
A new online learning algorithm for graph-structured sparsity.
problem Efficiently handling graph-structured sparsity constraints in online learning settings.
method Proposes extsc{GraphDA} algorithm that projects gradients and variables onto subspaces.
result Improves classification performance and captures graph-structured features effectively.
Proposes a neural network for sparsity regularization in inverse problems using Gaussian mixture.
problem Sparsity in inverse problems with limited significant components.
method Probabilistic sparsity prior as a mixture of degenerate Gaussians, trained with neural network.
result Neural network yields lower mean square error than LASSO, group LASSO, and iterative hard thresholding.
Proposes RT decomposition for better multi-relational link prediction.
problem Improving multi-relational link prediction in knowledge graphs.
method Relational Tucker3 (RT) decomposition, decouples entity and relation embeddings, allows parameter sharing, and learns sparsity patterns.
result RT decomposition can outperform existing sparse models in multi-relational link prediction.
This work introduces adversarial sparsity to measure robustness beyond adversarial accuracy.
problem Evaluating robustness to adversarial attacks beyond just accuracy.
method Adversarial sparsity, which quantifies the difficulty of finding perturbations.
result Sparsity provides valuable insights into neural networks and suggests improvements in robustness.
Sparsity-constrained optimization is an important and challenging problem that has wide applicability in data mining, machine learning, and statistics. In this paper, we focus on sparsity-constrained optimization in cases where the cost function is a general nonlinear function and, in particular, the sparsity constrain…
Grassmannian packings improve CNN kernels' diversity and reduce sparsity.
problem Kernel sparsity and lack of diversity in CNNs decrease model capacity.
method Initialize CNN kernels with Grassmannian packings to maximize diversity and minimize sparsity.
result Grassmannian packings lead to diverse features and improved classification accuracy.
New Max-Plus neural network exploits subgradient sparsity for efficient training.
problem Training Max-Plus neural networks is challenging due to dense subgradients.
method Proposes a sparse subgradient algorithm tailored to Max-Plus models.
result Achieves more efficient updates while retaining theoretical guarantees.