SparseRT accelerates sparse computations on GPUs for deep learning inference.
problem Efficiently handling unstructured sparsity patterns on GPUs for deep learning.
method SparseRT, a code generator that leverages unstructured sparsity for accelerating sparse linear algebra operations.
result Geometric mean speedups of 3.4x at 90% sparsity and 5.4x at 95% sparsity for 1x1 convolutions and fully connected layers.
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.
This paper accelerates sparse CNN layers on GPUs by using unstructured sparsity.
problem Efficiency of sparse CNN layers on GPUs.
method Direct sparse operation and reduced precision.
result Achieving up to 90% sparsity in deep CNN models improves efficiency.
Multi-cell cooperative processing with limited backhaul traffic is studied for cellular uplinks. Aiming at reduced backhaul overhead, a sparsity-regularized multi-cell receive-filter design problem is formulated. Both unstructured distributed cooperation as well as clustered cooperation, in which base station groups ar…
RMDA trains structured neural networks with regularization and variance reduction.
problem Training structured neural networks with desired properties.
method RMDA algorithm for structured NNs with regularization and variance reduction.
result RMDA achieves desired structures identical to regularizer's at stationary points.
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%.
Sparsity helps reduce the computational complexity of deep neural networks by skipping zeros. Taking advantage of sparsity is listed as a high priority in next generation DNN accelerators such as TPU. The structure of sparsity, i.e., the granularity of pruning, affects the efficiency of hardware accelerator design as w…
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.
Paper improves robustness and sparsity in adversarially trained DNNs.
problem Developing efficient compression algorithms for robustly trained DNNs.
method Pruning weights using relaxed augmented Lagrangian algorithms for both structured and unstructured levels, leveraging Feynman-Kac formalism.
result At least doubles channel sparsity of adversarially trained ResNet20 for CIFAR10 classification.
Recurrent Neural Networks (RNNs) are used in state-of-the-art models in domains such as speech recognition, machine translation, and language modelling. Sparsity is a technique to reduce compute and memory requirements of deep learning models. Sparse RNNs are easier to deploy on devices and high-end server processors. …
Sparse deep neural networks(DNNs) are efficient in both memory and compute when compared to dense DNNs. But due to irregularity in computation of sparse DNNs, their efficiencies are much lower than that of dense DNNs on regular parallel hardware such as TPU. This inefficiency leads to poor/no performance benefits for s…
MLP-Mixer achieves better performance through sparsity and wider architecture.
problem Understanding why MLP-Mixer outperforms conventional MLPs.
method Revealed sparseness as a key mechanism, showed effective expression as wider MLP, demonstrated quantitative similarities, and applied a guiding principle.
result MLP-Mixer's performance improvement through sparseness and wider architecture.
Proposes RBGP framework for efficient block sparse neural networks.
problem Efficiently exploit structured sparsity patterns for sparse neural networks on GPU.
method Uses Ramanujan Bipartite Graph Product to generate structured multi-level block sparse neural networks.
result Achieves 5-9x and 2-5x runtime gains over unstructured and block sparsity patterns respectively, while maintaining accuracy.
We present an extension of sparse PCA, or sparse dictionary learning, where the sparsity patterns of all dictionary elements are structured and constrained to belong to a prespecified set of shapes. This \emph{structured sparse PCA} is based on a structured regularization recently introduced by [1]. While classical spa…
Dropout-based regularization methods can be regarded as injecting random noise with pre-defined magnitude to different parts of the neural network during training. It was recently shown that Bayesian dropout procedure not only improves generalization but also leads to extremely sparse neural architectures by automatica…
A new method for optimizing hierarchical multi-objective problems.
problem Symmetry and neglect of objective hierarchy in existing multi-objective methods.
method Priority-Constrained Descent (PCD) framework exploiting hierarchical objective structures.
result Pareto dominance and better per-objective performance with secondary progress guarantees.
We rigorously evaluate three state-of-the-art techniques for inducing sparsity in deep neural networks on two large-scale learning tasks: Transformer trained on WMT 2014 English-to-German, and ResNet-50 trained on ImageNet. Across thousands of experiments, we demonstrate that complex techniques (Molchanov et al., 2017;…
We examine how recently documented, fundamental phenomena in deep learning models subject to pruning are affected by changes in the pruning procedure. Specifically, we analyze differences in the connectivity structure and learning dynamics of pruned models found through a set of common iterative pruning techniques, to …
As a result of the growing size of Deep Neural Networks (DNNs), the gap to hardware capabilities in terms of memory and compute increases. To effectively compress DNNs, quantization and connection pruning are usually considered. However, unconstrained pruning usually leads to unstructured parallelism, which maps poorly…
Improved regret bounds for structured linear contextual bandits with Gaussian noise.
problem Optimizing bandit learning algorithms for structured contexts with Gaussian perturbations.
method Proposed simple greedy algorithms for structured linear contextual bandits with Gaussian noise.
result Unified regret analysis for structured parameters with geometric quantities as bounds.
UQE uses LLMs to analyze unstructured data efficiently.
problem Efficient analytics on unstructured data.
method Proposes UQE, a query engine that uses LLMs to interpret UQL queries.
result Demonstrates efficient analytics on various unstructured data types.
Proposes a robust 3D classification method for sparse point clouds.
problem Invariance to rotation, positional shift, scaling, and robustness to point sparsity in point cloud classification.
method Introduces a graph-based feature learning approach with an end-to-end neural network.
result Significantly improves 3D object classification and retrieval tasks with sparse point clouds.
Paper provides conditions for reliable use of pre-trained embeddings in econometrics.
problem Uncertainty in using pre-trained embeddings for econometric tasks.
method Derives sufficient conditions and convergence rates for machine learning models with pre-trained embeddings.
result Establishes theoretical foundations for reliable use of pre-trained embeddings in econometrics.
Bayesian approach sparsifies neural networks efficiently.
problem Efficiently pruning neural networks to save resources.
method Sparsifiability via the Marginal likelihood (SpaM) framework.
result Prunes neural networks effectively without significant loss in performance.
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.
This paper explains how transformers learn from unstructured data in ICL.
problem Understanding how transformers learn from unstructured data in in-context learning.
method A simple transformer model with one or two attention layers and positional encoding is used to study the role of each component in ICL.
result A transformer with two attention layers and a look-ahead attention mask can learn from unstructured data.
Study introduces a benchmark suite for evaluating neural MI estimators on real-world unstructured datasets.
problem Lack of comprehensive evaluation methods for neural MI estimators on real-world unstructured datasets.
method Developed a benchmark suite using same-class sampling and a binary symmetric channel trick.
result Showed accurate manipulation of true MI values of real-world datasets.
LLMs learn new tasks from unstructured data, but it depends on word co-occurrence and positional information.
problem Understanding how LLMs can learn new tasks from unstructured data without explicit training.
method Examined the capabilities of LLMs trained on unstructured data, focusing on sequence model requirements and training data structure.
result Many ICL capabilities can emerge from word co-occurrence in unstructured data, but positional information is crucial for certain tasks.
Principal component analysis (PCA) is an exploratory tool widely used in data analysis to uncover dominant patterns of variability within a population. Despite its ability to represent a data set in a low-dimensional space, the interpretability of PCA remains limited. However, in neuroimaging, it is essential to uncove…
Develops Φ-DVAE for assimilating unstructured data into physical models.
problem Challenges in incorporating unstructured data into physical models.
method Physics-informed dynamical variational autoencoder (Φ-DVAE) combining latent state-space model and VAE. result Demonstrates data-efficient dynamics encoding with competitive performance and uncertainty quantification.
Healthcare data continues to flourish yet a relatively small portion, mostly structured, is being utilized effectively for predicting clinical outcomes. The rich subjective information available in unstructured clinical notes can possibly facilitate higher discrimination but tends to be under-utilized in mortality pred…
LLMs help automate extraction of actuarial variables from unstructured claims data.
problem Manual processing of unstructured claims data is time-consuming and inconsistent.
method Two-stage processing architecture using LLMs, modular Python pipeline.
result LLM-based extraction achieved high accuracy and practical actuarial value.
Unstructured data refers to information that does not have a predefined data model or is not organized in a pre-defined manner. Loosely speaking, unstructured data refers to text data that is generated by humans. In after-sales service businesses, there are two main sources of unstructured data: customer complaints, wh…
Data fields sampled on irregularly spaced points arise in many applications in the sciences and engineering. For regular grids, Convolutional Neural Networks (CNNs) have been successfully used to gaining benefits from weight sharing and invariances. We generalize CNNs by introducing methods for data on unstructured poi…
GPI uses GenAI models to infer causal and predictive effects from unstructured data.
problem Estimating causal and predictive effects from unstructured data like text and images.
method Leverages open-source GenAI models to generate and represent unstructured data, applying machine learning to these representations.
result GPI efficiently estimates causal and predictive effects with quantified uncertainty, without fine-tuning.
Improved pruning method using iterative sensitivity ranking before training.
problem Improper sensitivity propagation in existing pruning methods.
method Iterative application of SNIP criterion before training.
result State-of-the-art sparsity-performance trade-offs achieved.
This paper presents a framework for efficiently learning feature selection policies which use less features to reach a high classification precision on large unstructured data. It uses a Deep Convolutional Autoencoder (DCAE) for learning compact feature spaces, in combination with recently-proposed Reinforcement Learni…
Sparse covariance estimation in the vertical-split model achieves exponential improvement over dense estimates.
problem Minimax estimation error for distributed covariance matrix estimation in the vertical-split setting.
method Elementwise s-sparsity is shown to reduce communication and sample complexity. result Minimax lower bounds for 1-sparse cross-covariance estimation are established. Enhanced regime shifts detection using unstructured text and financial data.
problem Detecting regime shifts in financial markets is challenging due to noisy and multicollinear data.
method Combines LLM reasoning on unstructured text and statistical validation on financial time series.
result Framework achieves F1 score of 0.82, outperforming pure data-driven methods.
Develops methods to improve demand counterfactuals from imperfect proxies.
problem Imperfect proxies in demand models lead to biased counterfactuals and invalid inference.
method Practical toolkit for market-level and individual data, requiring minimal computation.
result Improves substitution prediction and counterfactual performance.
A new method monitors unstructured 3D shapes without registration.
problem Error-prone registration and mesh reconstruction steps in PCD monitoring.
method Intrinsic geometric properties of shapes, using Laplacian and geodesic distances.
result Effective monitoring of defects without registration and mesh reconstruction.
Deep neural networks predict prostate motion from MR images.
problem Predicting prostate motion during ultrasound-guided interventions.
method Biomechanically-trained deep neural networks on unstructured nodes.
result Trained networks yield near real-time inference with 0.017 mm error.
This paper compares unstructured and structured EM-based semi-supervised learning methods.
problem Semi-supervised learning with EM algorithm for structured prediction.
method Comparative study between unstructured and structured EM-based semi-supervised learning methods.
result Structured EM is more robust to class confusion in flood mapping datasets.
We introduce a new Bayesian multi-class support vector machine by formulating a pseudo-likelihood for a multi-class hinge loss in the form of a location-scale mixture of Gaussians. We derive a variational-inference-based training objective for gradient-based learning. Additionally, we employ an inducing point approxima…
We propose Sparse Neural Network architectures that are based on random or structured bipartite graph topologies. Sparse architectures provide compression of the models learned and speed-ups of computations, they can also surpass their unstructured or fully connected counterparts. As we show, even more compact topologi…
New insights on pruning deep networks by preserving function locality.
problem Designing effective pruning methods for deep neural networks.
method Revisited loss modeling using first and second order Taylor expansions, emphasizing locality.
result Both first and second order Taylor expansions can achieve similar performance in pruning.
Proposes RTL model for sentiment classification and key word detection in online reviews.
problem Sentiment classification and key word detection in online reviews for hospitality industry.
method Regularized Text Logistic (RTL) regression model.
result RTL model achieves satisfactory classification performance and identifies key word features.
Automated tests detect interactions in unstructured data.
problem Detecting interactions between latent variables in low-dimensional systems.
method Derive two interaction tests based on pairwise interventions and integrate them into an active learning pipeline.
result Tests can identify more known biological interactions than random search and standard active learning baselines.