time-series-6

Data Compression

3260 papers • 126 benchmarks • 313 datasets

This task has no description! Would you like to contribute one?

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in data-compression-5

Trend

Dataset

Best Model

Actions

No benchmarks available.

Libraries

i

Use these libraries to find data-compression-5 models and implementations

dmlc/xgboost

2 papers 25,654

Datasets

No datasets available.

Subtasks

No subtasks available.

Most implemented papers

XGBoost: A Scalable Tree Boosting System

Tianqi Chen, Carlos Guestrin•Tue Mar 08 2016

This paper proposes a novel sparsity-aware algorithm for sparse data and weighted quantile sketch for approximate tree learning and provides insights on cache access patterns, data compression and sharding to build a scalable tree boosting system called XGBoost.

48813

Content

0

Paper Graph

DNABERT-2: Efficient Foundation Model and Benchmark For Multi-Species Genome

Zhihan Zhou, Han Liu, Yanrong Ji, Weijian Li, Pratik Dutta, R. Davuluri•Sun Jun 25 2023

This work proposes DNABERT-2, a refined genome foundation model that adapts an efficient tokenizer and employs multiple strategies to overcome input length constraints, reduce time and memory expenditure, and enhance model capability, and introduces Byte Pair Encoding (BPE), a statistics-based data compression algorithm that constructs tokens.

325 0

Paper Graph

Efficient Manifold and Subspace Approximations with Spherelets

Didong Li, D. Dunson, Minerva Mukhopadhyay•Sun Jun 25 2017

A simple and general alternative to approximating subspaces using a locally linear, and potentially multiscale, dictionary is proposed, which instead uses pieces of spheres, or spherelets, to locally approximate the unknown subspace.

6 0

Paper Graph

Transformer-based Transform Coding

Taco Cohen, Yinhao Zhu, Yang Yang•Fri Dec 31 2021

It is shown that nonlinear transforms built on Swin-transformers can achieve better compression efficiency than transforms built on convolutional neural networks (ConvNets), while requiring fewer parameters and shorter decoding time.

201 0

Paper Graph

Norm-Explicit Quantization: Improving Vector Quantization for Maximum Inner Product Search

XINYAN DAI, Xiao Yan, K. K. Ng, Jie Liu, James Cheng•Mon Nov 11 2019

It is shown that quantization errors in norm have much higher influence on inner products than quantizationerrors in direction, and small quantization error does not necessarily lead to good performance in maximum inner product search (MIPS), so norm-explicit quantization (NEQ) is proposed — a general paradigm that improves existing VQ techniques for MIPS.

32 0

Paper Graph

ReduNet: A White-box Deep Network from the Principle of Maximizing Rate Reduction

John Wright, Yaodong Yu, Chong You, Yi Ma, Haozhi Qi, Kwan Ho Ryan Chan•Thu May 20 2021

This work argues that for high-dimensional multi-class data, the optimal linear discriminative representation maximizes the coding rate difference between the whole dataset and the average of all the subsets and indicates that such a deep convolution network is significantly more efficient to construct and learn in the spectral domain.

142 0

Paper Graph

Supervised Compression for Resource-Constrained Edge Computing Systems

M. Levorato, Yoshitomo Matsubara, Ruihan Yang, Stephan Mandt•Fri Aug 20 2021

This paper adopts ideas from knowledge distillation and neural image compression to compress intermediate feature representations more efficiently and shows that the learned feature representations can be tuned to serve multiple downstream tasks.

71 0

Paper Graph

Towards Empirical Sandwich Bounds on the Rate-Distortion Function

Stephan Mandt, Yibo Yang•Mon Nov 22 2021

The first attempt at an algorithm for sandwiching the R-D function of a general (not necessarily discrete) source requiring only i.i.d. data samples is made, indicating theoretical room for improving state-of-the-art image compression methods by at least one dB in PSNR at various bitrates.

28 0

Paper Graph

BottleFit: Learning Compressed Representations in Deep Neural Networks for Effective and Efficient Split Computing

M. Levorato, Sameer Singh, Francesco Restuccia, Yoshitomo Matsubara, Davide Callegaro•Thu Jan 06 2022

This work proposes a new framework called BottleFit, which, in addition to targeted DNN architecture modifications, includes a novel training strategy to achieve high accuracy even with strong compression rates, and applies BottleFit on cutting-edge DNN models in image classification.

52 0

Paper Graph

An Introduction to Neural Data Compression

Lucas Theis, Stephan Mandt, Yibo Yang•Sun Feb 13 2022

The present article aims to introduce neural compression to a broader machine learning audience by reviewing the necessary background in information theory and computer vision, and providing a curated guide through the essential ideas and methods in the literature thus far.

148 0

Paper Graph

Adding a benchmark result helps the community track progress.