Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

169,341 papers · 148 categories

Trend · papers per month

175350524699 · Jun 202019922001200920182026
48 results for sample compression size

Learnable multiclass hypothesis classes don't always have a sample compression scheme of fixed size.

problem The limitation of sample compression schemes for multiclass hypothesis classes.
method Analysis of DS dimension and sample compression schemes.
result Learnable multiclass hypothesis classes do not always have a sample compression scheme of fixed size.

BDC compresses both sample size and dimensionality of large datasets.

problem Large datasets in both sample size and dimensionality.
method Two-stage framework using Decoded MMD, Reconstruction MMD, and Encoded MMD.
result BDC achieves comparable or superior performance with lower cost and higher compression rates.

New method compresses large sample data for faster discriminant analysis.

problem Large sample sizes in discriminant analysis increase computational burden.
method Proposes a new compression approach for reducing training samples.
result Significant computational gains and superior predictive ability compared to random sub-sampling.

New bounds found for agnostic learning with sample compression schemes.

problem Finding optimal rates of convergence for agnostic learning.
method Established tight characterization of worst-case rates for agnostic learning with sample compression schemes.
result Optimal rates of convergence for size-kk agnostic sample compression schemes are klog(n/k)n\sqrt{\frac{k \log(n/k)}{n}}.

Reduces policy space complexity for reinforcement learning.

problem Efficiency in exploring vast policy spaces in reinforcement learning.
method Uses Rényi divergence and l1l_1 norm to determine sample size for accurate policy approximation.
result Established error bounds for sample size requirements in model-based and model-free settings.

Adaptive sampling method optimizes DNN compression for resource-constrained platforms.

problem Efficiently compressing DNNs for resource-constrained platforms with high accuracy.
method Adaptive sampling using genetic algorithm-inspired operations to optimize hyperparameters.
result Adaptive sampling outperforms rule-based and reinforcement learning methods in compression rate and accuracy.

Clapping reduces memory usage in distributed optimization by reusing data samples.

problem Significant communication overhead and impractical memory overhead in pipeline-parallel distributed optimization.
method Lazy sampling strategy to reuse data samples across steps, supporting convergence without unbiased gradient assumptions.
result Clapping achieves convergence in few-epoch or online training regimes without sample-size memory overhead.

A new method for high-dimensional data classification reduces misclassification errors.

problem High-dimensional data classification with limited samples.
method Compressive Regularized Discriminant Analysis (CRDA) using joint-sparsity promoting hard thresholding and regularized covariance matrix estimators.
result CRDA gives fewer misclassification errors than competitors and accurately selects features.

A new framework for efficient large-scale learning using sketching of moments.

problem Efficiently learning from large datasets with limited computational resources.
method Compressing the training data into a low-dimensional sketch and solving a nonlinear least squares problem.
result Sufficient sketch sizes to control the generalization error of the procedure.

This paper automates deep model compression using reinforcement learning.

problem Efficiently compressing deep neural networks without sacrificing accuracy.
method Reinforcement learning-based actor-critic structure for automated compression.
result 4-fold reduction in FLOP with 2.8% higher accuracy for VGG-16.

CTE improves explanation estimation with less data and faster computation.

problem Inefficient and inaccurate explanation estimation in machine learning models.
method Distribution compression through kernel thinning to reduce sample size.
result CTE significantly improves accuracy and stability of explanation estimation.

Compress++ speeds up distribution compression to near-linear time.

problem Accurately summarize a probability distribution using a small number of points efficiently.
method Introduces Compress++, a meta-procedure to speed up any thinning algorithm.
result Achieves n\sqrt{n} points with O(logn/n)\mathcal{O}(\sqrt{\log n/n}) integration error in O(nlog3n)\mathcal{O}(n \log^3 n) time and O(nlog2n)\mathcal{O}( \sqrt{n} \log^2 n ) space.

A CAE improves DNN's outlier and adversary defense.

problem Improving DNN's robustness against outliers and adversaries.
method Proposes a classification-autoencoder (CAE) that compresses samples into disjoint spaces and uses a decoder to classify and defend against adversaries.
result The CAE achieves state-of-the-art outlier recognition and near-lossless classification of adversaries.

Study robust regression learning under adversarial attacks.

problem Understanding which function classes are learnable in the presence of adversarial attacks.
method Introduced a novel agnostic sample compression scheme and used fat-shattering dimension to construct adversarially robust sample compression schemes.
result Finite fat-shattering dimension classes are learnable in both realizable and agnostic settings.

New method compresses neural networks using random code, improving efficiency.

problem Large memory footprint of deep neural networks.
method Training a variational distribution over weights, encoding using Kullback-Leibler divergence.
result Achieves state-of-the-art compression rates and test performance.

New study reveals how heavy-tailed SGD dynamics lead to compressible neural networks.

problem Understanding why large neural networks can be compressed effectively.
method Linking SGD dynamics to compressibility properties of neural networks.
result Large step-size/batch-size ratios and overparametrization lead to heavy-tailed SGD dynamics, making networks compressible.

Autoencoders achieve image compression without needing multiple transforms.

problem Learning transforms for image compression with varying quantization steps.
method Use a single learned transform for multiple rate-distortion points at test time.
result Comparable performance can be achieved with a single learned transform.

The paper bounds information losses in neural classifiers from sampling.

problem Information losses in neural classifiers from finite datasets.
method Proves a relationship between information losses and expected total variation of the estimated neural model, bounds this expected total variation as a function of dataset size.
result Obtains bounds on information losses that are less sensitive to input compression and much smaller than existing bounds.

Boosting improves accuracy by combining weak learners into a voting classifier.

problem Boosting's theoretical performance is sub-optimal, especially for voting classifiers.
method Proposes a randomized boosting algorithm that outputs voting classifiers with a single logarithmic dependency on sample size.
result Randomized boosting achieves a generalization error with a single logarithmic dependency on the sample size.

DeepThin compresses deep neural networks, improving performance and reducing resource usage.

problem Efficiently compressing large neural networks for mobile devices.
method Combining rank factorization with a reshaping process to add nonlinearity.
result DeepThin achieves significant improvements in word error rates and test loss compared to existing methods.

We compress large neural networks for quick adaptation to specific contexts.

problem How to quickly adapt a pretrained large neural network to specific contexts.
method Propose a Bayesian hypernetwork framework to compress the network and encourage sparsity.
result Generated compressed networks are significantly smaller than baseline methods.

We learn sparse precision matrices from compressed data sketches.

problem Learning a graph from high-dimensional data with limited storage.
method Estimate a sparse precision matrix from a sketch of the data using non-linear random features.
result It is possible to estimate a sparse precision matrix from a sketch of size $m=Ω\left((d+2k)\log(d) ight)$.

This paper uses deep reinforcement learning to compress CNN models, reducing size and maintaining accuracy.

problem Reducing model size for efficient deployment on limited hardware resources.
method Two-stage compression pipeline: pruning and quantization using deep reinforcement learning.
result Significant reduction in model size with minimal loss in accuracy.

Paper proposes compressive ICA algorithms for ICA model.

problem Efficiently solving ICA model with reduced memory and computational complexity.
method Compressive learning approach to ICA model, proving existence of compressive ICA scheme, proposing two algorithms (IPG and ASD).
result Proposed algorithms achieve substantial memory gains over well-known ICA algorithms.

We propose a method for inferring the conditional indepen- dence graph (CIG) of a high-dimensional discrete-time Gaus- sian vector random process from finite-length observations. Our approach does not rely on a parametric model (such as, e.g., an autoregressive model) for the vector random process; rather, it only assu…

2013-11-13abs ↗pdf ↗

We extend quantization-aware training to extreme model compression.

problem Maximizing model accuracy with minimal model size.
method Quantize a random subset of weights during training, allowing unbiased gradients through other weights.
result Established new state-of-the-art compromises between accuracy and model size.

IMPACT optimizes LLM compression by focusing on activation importance, reducing model size up to 55.4%.

problem Resource constraints in deploying large language models (LLMs).
method IMPACT integrates activation importance into low-rank compression, optimizing for both size and accuracy.
result IMPACT achieves up to 55.4% greater model size reduction while maintaining comparable or better accuracy.

Saec compresses recommendation system embeddings by clustering similar features.

problem Large embedding matrix in recommendation systems consumes excessive memory.
method Saec clusters similar features within a field to reduce embedding matrix size.
result Saec reduces embedding size by ~27x with no performance loss.

Paper optimizes privacy-preserving distribution estimation for sparse data.

problem Sparse distribution estimation under local differential privacy constraints.
method Compressive sensing approaches for privacy-preserving estimation.
result Significant reduction in sample complexity for approximately sparse distributions.