methodology-10

Imputation

3260 papers • 126 benchmarks • 313 datasets

Substituting missing data with values according to some criteria.

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in imputation-41

Trend

Dataset

Best Model

Actions

Sprites

PhysioNet Challenge 2012

HMNIST

Libraries

i

Use these libraries to find imputation-41 models and implementations

WenjieDu/PyPOTS

11 papers 524

Datasets

WebKB

Sprites

MotionSense

PhysioNet Challenge 2012

AIDS Antiviral Screen

Ecoli

Subtasks

Multivariate Time Series Imputation

Most implemented papers

GAIN: Missing Data Imputation using Generative Adversarial Nets

M. Schaar, Jinsung Yoon, James Jordon•Wed Jun 06 2018

This work proposes a novel method for imputing missing data by adapting the well-known Generative Adversarial Nets (GAN) framework and calls it GAIN, which significantly outperforms state-of-the-art imputation methods.

1261

Content

Adult Data Set

xinychen/transdim

5 papers 1,082

FDF

METR-LA Point Missing

PEMS-BAY Point Missing

0

Paper Graph

Recurrent Neural Networks for Multivariate Time Series with Missing Values

Kyunghyun Cho, D. Sontag, Yan Liu, Zhengping Che, S. Purushotham•Sun Jun 05 2016

Novel deep learning models are developed based on Gated Recurrent Unit, a state-of-the-art recurrent neural network that takes two representations of missing patterns, i.e., masking and time interval, and effectively incorporates them into a deep model architecture so that it not only captures the long-term temporal dependencies in time series, but also utilizes the missing patterns to achieve better prediction results.

2219 0

Paper Graph

Unsupervised Data Imputation via Variational Inference of Deep Subspaces

Adrian V. Dalca, M. Sabuncu, J. Guttag•Thu Mar 07 2019

This work introduces a general probabilistic model that describes sparse high dimensional imaging data as being generated by a deep non-linear embedding and introduces sparsity-aware network building blocks that explicitly model observed and missing data.

14 0

Paper Graph

A Transformer-based Framework for Multivariate Time Series Representation Learning

George Zerveas, Srideepika Jayaraman, Dhaval Patel, A. Bhamidipaty, Carsten Eickhoff•Mon Oct 05 2020

A novel framework for multivariate time series representation learning based on the transformer encoder architecture, which can offer substantial performance benefits over fully supervised learning on downstream tasks, both with but even without leveraging additional unlabeled data, i.e., by reusing the existing data samples.

1246 0

Paper Graph

Deep Learning in Single-cell Analysis

Jiliang Tang, Jiayuan Ding, Dylan Molho, Zhaoheng Li, Haifang Wen, Wenzhuo Tang, Yixin Wang, Julian Venegas, Wei Jin, Renming Liu, Runze Su, P. Danaher, Robert Yang, Y. Lei, Yuying Xie•Fri Oct 21 2022

A comprehensive survey on deep learning in single-cell analysis, including seven popular tasks spanning different stages of the single-cell analysis pipeline, including multimodal integration, imputation, clustering, spatial domain identification, cell-type deconvolution, cell segmentation, and cell-type annotation.

29 0

Paper Graph

Imaging Time-Series to Improve Classification and Imputation

Zhiguang Wang, T. Oates•Sun May 31 2015

This work proposes a novel framework for encoding time series as different types of images, namely, Gramian Angular Summation/Difference Fields (GASF/GADF) and Markov Transition Fields (MTF), which enables the use of techniques from computer vision for time series classification and imputation.

798 0

Paper Graph

BRITS: Bidirectional Recurrent Imputation for Time Series

Lei Li, Hao Zhou, Wei Cao, Jian Li, Yitan Li•Mon Apr 30 2018

BRITS is a novel method based on recurrent neural networks for missing value imputation in time series data that directly learns the missing values in a bidirectional recurrent dynamical system, without any specific assumption.

797 0

Paper Graph

CSDI: Conditional Score-based Diffusion Models for Probabilistic Time Series Imputation

Stefano Ermon, Y. Tashiro, Jiaming Song, Yang Song•Tue Jul 06 2021

Conditional Score-based Diffusion models for Imputation (CSDI), a novel time series imputation method that utilizes score-based diffusion models conditioned on observed data, which improves by 40-65% over existing probabilistic imputation methods on popular performance metrics.

847 0

Paper Graph

Input Convex Neural Networks

J. Z. Kolter, Lei Xu, Brandon Amos•Wed Aug 31 2016

This paper presents the input convex neural network architecture, which are scalar-valued (potentially deep) neural networks with constraints on the network parameters such that the output of the network is a convex function of (some of) the inputs.

755 0

Paper Graph

Adding a benchmark result helps the community track progress.