【重制版】ICML 2023 | 时间序列(Time Series)和时空数据(Spatial-Temporal)论文总结

2023ICML(International Conference on Machine Learning,国际机器学习会议)在2023年7月23日-29日在美国夏威夷举行。2023年ICML 共收到 6538 份投稿,其中 1827 份被接收,接收率约为 27.9%。(好像ICML24要开始第一轮rebuttal了,蹭蹭热度)

本文总结了ICML 23有关**时间序列(Time Series)时空(Spatial-temporal)**的相关论文,如有疏漏,欢迎大家补充。

时间序列Topic:插补,预测,因果,表示学习,无监督,对比学习,不确定性等

时空数据Topic:AI4Science,GeoAI,时空预测等

🌟【紧跟前沿】“时空探索之旅”与你一起探索时空奥秘!🚀
欢迎大家关注时空探索之旅时空探索之旅QRCode

目录

时间序列(Time Series)

  1. Probabilistic Imputation for Time-series Classification with Missing Data
  2. Provably Convergent Schrödinger Bridge with Applications to Probabilistic Time Series Imputation
  3. Deep Latent State Space Models for Time-Series Generation
  4. Neural Stochastic Differential Games for Time-series Analysis
  5. Context Consistency Regularization for Label Sparsity in Time Series
  6. Sequential Multi-Dimensional Self-Supervised Learning for Clinical Time Series
  7. SOM-CPC: Unsupervised Contrastive Learning with Self-Organizing Maps for Structured Representations of High-Rate Time Series
  8. Prototype-oriented unsupervised anomaly detection for multivariate time series
  9. Learning Deep Time-index Models for Time Series Forecasting
  10. [Oral]Neural Continuous-Discrete State Space Models for Irregularly-Sampled Time Series
  11. [Oral]Self-Interpretable Time Series Prediction with Counterfactual Explanations
  12. Learning Perturbations to Explain Time Series Predictions
  13. Feature Programming for Multivariate Time Series Prediction
  14. Theoretical Guarantees of Learning Ensembling Strategies with Applications to Time Series Forecasting
  15. Sequential Predictive Conformal Inference for Time Series
  16. Non-autoregressive Conditional Diffusion Models for Time Series Prediction
  17. Sequential Monte Carlo Learning for Time Series Structure Discovery
  18. Domain Adaptation for Time Series Under Feature and Label Shifts

时空数据(Spatial-temporal)

  1. NeuralStagger: Accelerating Physics-constrained Neural PDE Solver with Spatial-temporal Decomposition
  2. Spatial Implicit Neural Representations for Global-Scale Species Mapping
  3. Spatial-Temporal Graph Learning with Adversarial Contrastive Adaptation

时间序列(time series)

时序标题词云

1. Probabilistic Imputation for Time-series Classification with Missing Data

大会论文链接:https://icml.cc/virtual/2023/poster/23522

PMLR链接:https://proceedings.mlr.press/v202/kim23m

作者:SeungHyun Kim · Hyunsu Kim · EungGu Yun · Hwangrae Lee · Jaehun Lee · Juho Lee

机构:韩国科学技术院(KAIST),赛视智能(Saige Research),三星研究院

关键词:时间序列数据插补,概率模型,不确定性

2. Provably Convergent Schrödinger Bridge with Applications to Probabilistic Time Series Imputation

大会论文链接:https://icml.cc/virtual/2023/poster/24369

PMLR链接:https://proceedings.mlr.press/v202/chen23f.html

代码:https://github.com/morganstanley/MSML/tree/main/papers/Conditional_Schrodinger_Bridge_Imputation

作者:Yu Chen · Wei Deng · Shikai Fang · Fengpei Li · Tianjiao N Yang · Yikai Zhang · Kashif Rasul · Shandian Zhe · Anderson Schneider · Yuriy N evmyvaka

机构:摩根士丹利,犹他大学(Utah),埃默里大学(Emory)

关键词:时间序列插补、概率模型

24369

3. Deep Latent State Space Models for Time-Series Generation

大会论文链接:https://icml.cc/virtual/2023/poster/24503

PMLR链接:https://proceedings.mlr.press/v202/zhou23i.html

作者:Linqi Zhou · Michael Poli · Winnie Xu · Stefano Massaroli · Stefano Ermon

机构:斯坦福大学(Stanford),多伦多大学(Toronto),MILA

关键词:时间序列生成

24503

4. Neural Stochastic Differential Games for Time-series Analysis

大会论文链接:https://icml.cc/virtual/2023/poster/25204

PMLR链接:https://proceedings.mlr.press/v202/park23j.html

代码:https://github.com/LGAI-AML/MaSDEs

作者:Sungwoo Park, Byoungwoo Park, Moontae Lee, Changhee Lee

机构:LG AI Research,韩国中央大学,伊利诺伊大学芝加哥分校(UIC)

关键词:时间序列分析

5. Context Consistency Regularization for Label Sparsity in Time Series

大会论文链接:https://icml.cc/virtual/2023/poster/24019

PMLR链接:https://proceedings.mlr.press/v202/shin23e.html

代码:https://github.com/kaist-dmlab/CrossMatch

作者:Yooju Shin · Susik Yoon · Hwanjun Song · Dongmin Park · Byunghyun Kim · Jae-Gil Lee · Byung Suk Lee

机构:KAIST,伊利诺伊大学香槟分校(UIUC),亚马逊AI Lab(AWS AI Lab),佛蒙特大学(Vermont)

关键词:稀疏性,正则化

CrossMatch

6. Sequential Multi-Dimensional Self-Supervised Learning for Clinical Time Series

大会论文链接:https://icml.cc/virtual/2023/poster/24132

PMLR链接:https://proceedings.mlr.press/v202/raghu23a.html

代码:https://github.com/aniruddhraghu/smd-ssl

作者:Aniruddh Raghu · Payal Chandak · Ridwan Alam · John Guttag · Collin Stultz

机构:MIT,哈佛-麻省理工健康科学技术项目(HST)

关键词:自监督、临床时序

SMD SSL

7. SOM-CPC: Unsupervised Contrastive Learning with Self-Organizing Maps for Structured Representations of High-Rate Time Series

大会论文链接:https://icml.cc/virtual/2023/poster/24856

PMLR链接:https://proceedings.mlr.press/v202/huijben23a.html

代码:https://github.com/IamHuijben/SOM-CPC

作者:SeungHyun Kim · Hyunsu Kim · EungGu Yun · Hwangrae Lee · Jaehun Lee · Juho Lee

机构:埃因霍芬理工大学(tue),Onera Health,Sleep Medicine Center Kempenhaeghe

关键词:无监督,对比学习,高频时序

SOM-CPC

8. Prototype-oriented unsupervised anomaly detection for multivariate time series

大会论文链接:https://icml.cc/virtual/2023/poster/24139

PMLR链接:https://proceedings.mlr.press/v202/li23d.html

代码:https://github.com/LiYuxin321/PUAD

作者:yuxin li · Wenchao Chen · Bo Chen · Dongsheng Wang · Long Tian · Mingyuan Zhou

机构:西安电子科技大学,德克萨斯大学奥斯汀分校(utexas)

关键词:异常检测,多元时序,无监督

PUAD

9. Learning Deep Time-index Models for Time Series Forecasting

大会论文链接:https://icml.cc/virtual/2023/poster/24424

PMLR链接:https://proceedings.mlr.press/v202/woo23b.html

代码:https://github.com/salesforce/DeepTime

作者:Gerald Woo · Chenghao Liu · Doyen Sahoo · Akshat Kumar · Steven Hoi

机构:Salesforce,新加坡管理大学(SMU)

关键词:时间序列预测,元学习

DeepTime

10. [Oral]Neural Continuous-Discrete State Space Models for Irregularly-Sampled Time Series

大会论文链接:https://icml.cc/virtual/2023/poster/24923

PMLR链接:https://proceedings.mlr.press/v202/ansari23a.html

代码:https://github.com/clear-nus/NCDSSM

作者:Abdul Fatir Ansari · Alvin Heng · Andre Lim · Harold Soh

关键词:不规则采样时间序列

NCDSSM

11. [Oral]Self-Interpretable Time Series Prediction with Counterfactual Explanations

大会论文链接:https://icml.cc/virtual/2023/poster/23975

PMLR链接:https://proceedings.mlr.press/v202/yan23d.html

代码:https://github.com/Wang-ML-Lab/self-interpretable-time-series

作者:Jingquan Yan · Hao Wang

机构:罗格斯大学(Rutgers)

关键词:反事实,时间序列预测,可解释性

12. Learning Perturbations to Explain Time Series Predictions

大会论文链接:https://icml.cc/virtual/2023/poster/24182

PMLR链接:https://proceedings.mlr.press/v202/enguehard23a.html

代码:https://github.com/josephenguehard/time_interpret

作者:Joseph Enguehard(独立一个作者太牛了!)

机构:Babylon Health,Skippr, 99 Milton Keynes Business Centre

关键词:可解释性,时间序列预测

time_interpret

13. Feature Programming for Multivariate Time Series Prediction

大会论文链接:https://icml.cc/virtual/2023/poster/23862

PMLR链接:https://proceedings.mlr.press/v202/reneau23a.html

代码:https://github.com/SirAlex900/FeatureProgramming

作者:Alex Reneau · Jerry Yao-Chieh Hu · Ammar Gilani · Han Liu

机构:西北大学(Northwestern)

关键词:特征工程,多元时间序列预测

FeatureProgramming

14. Theoretical Guarantees of Learning Ensembling Strategies with Applications to Time Series Forecasting

大会论文链接:https://icml.cc/virtual/2023/poster/24910

PMLR链接:https://proceedings.mlr.press/v202/hasson23a.html

作者:Hilaf Hasson · Danielle Robinson · Yuyang Wang · Gaurav Gupta · Youngsuk Park

关键词:集成学习,时间序列预测

24910

15. Sequential Predictive Conformal Inference for Time Series

大会论文链接:https://icml.cc/virtual/2023/poster/24627

PMLR链接:https://proceedings.mlr.press/v202/xu23r

代码:https://github.com/hamrel-cxu/SPCI-code

作者:Chen Xu · Yao Xie

机构:佐治尼亚理工学院(Gatech)

关键词:共形预测,不确定性

SPCI

16. Non-autoregressive Conditional Diffusion Models for Time Series Prediction

大会论文链接:https://icml.cc/virtual/2023/poster/25084

PMLR链接:https://proceedings.mlr.press/v202/shen23d.html

作者:Lifeng Shen · James Kwok

机构:香港科技大学(HKUST)

关键词:扩散模型,时间序列预测

TimeDiff

17. Sequential Monte Carlo Learning for Time Series Structure Discovery

大会论文链接:https://icml.cc/virtual/2023/poster/24520

PMLR链接:https://proceedings.mlr.press/v202/saad23a.html

代码:https://github.com/probsys/AutoGP.jl(代码语言是Julia)

项目地址:https://probsys.github.io/AutoGP.jl/stable/

作者:Feras Saad · Brian Patton · Matthew Hoffman · Rif Saurous · Vikash Mansinghka

机构:卡耐基梅隆大学(CMU),Google,MIT

关键词:蒙特卡洛方法,高斯过程,时间序列结构发现

AutoGP

18. Domain Adaptation for Time Series Under Feature and Label Shifts

大会论文链接:https://icml.cc/virtual/2023/poster/23456

PMLR链接:https://proceedings.mlr.press/v202/he23b.html

代码:https://github.com/mims-harvard/Raincoat

作者:Huan He, Owen Queen, Teddy Koker, Consuelo Cuevas, Theodoros Tsiligkaridis, Marinka Zitnik

机构:哈佛大学(Harvard),MIT林肯实验室,

关键词:域适应,标签偏移,数据偏移,分布偏移

23456

时空数据(spatial-temporal data)

时空的论文偏少,且3篇内容差异较大,就不显示词云了

19. NeuralStagger: Accelerating Physics-constrained Neural PDE Solver with Spatial-temporal Decomposition

大会论文链接:https://icml.cc/virtual/2023/poster/23962

PMLR链接:https://proceedings.mlr.press/v202/huang23m.html

作者:Xinquan Huang · Wenlei Shi · Qi Meng · Yue Wang · Xiaotian Gao · Jia Zhang · Tie-Yan Liu

机构:阿卜杜拉国王科技大学(KAUST),微软亚洲研究院(MSRA)

关键词:神经偏微分方程,时空分解,AI4Science

20. Spatial Implicit Neural Representations for Global-Scale Species Mapping

大会论文链接:https://icml.cc/virtual/2023/poster/23767

PMLR链接:https://proceedings.mlr.press/v202/cole23a.html

代码:https://github.com/elijahcole/sinr

作者:Elijah Cole · Grant Horn · Christian Lange · Alexander Shepard · Patrick Leary · Pietro Perona · Scott Loarie · Oisin Mac Aodha

机构:加州理工学院(CalTech),康奈尔大学(Cornell),爱丁堡大学(Edinburgh),iNaturalist

关键词:地理信息,稀疏性,隐式神经表示

sinr

21. Spatial-Temporal Graph Learning with Adversarial Contrastive Adaptation

(可能是最像时空数据挖掘的一篇了)

大会论文链接:https://icml.cc/virtual/2023/poster/24346

PMLR链接:https://proceedings.mlr.press/v202/zhang23p.html

代码:https://github.com/HKUDS/GraphST

作者:Qianru Zhang · Chao Huang · Lianghao Xia · Zheng Wang · Siu Ming Yiu · Ruihua Han

机构:香港大学(HKU),南洋理工大学(NTU)

关键词:时空图预测、对比学习,图表示学习

GraphST

相关链接

ICML2023接受论文:ICML 2023 Papers

ICML 2023 | 时间序列(Time Series)和时空数据(Spatial-Temporal)论文总结

ICLR 2024投稿时空数据论文汇总

NeurIPS 2023 时间序列(Time Series)论文总结

NeurIPS 2023 时空数据论文总结

KDD 2023 时空数据论文

VLDB 2023 时空&时序论文汇总

ICDE 2023 时空数据论文

WWW 2023 时空数据论文

ECML PKDD 2023 时空数据论文

CIKM 2023 时空数据论文总结

KDD 2023 时空数据论文

VLDB 2023 时空&时序论文汇总

ICDE 2023 时空数据论文

WWW 2023 时空数据论文

ECML PKDD 2023 时空数据论文

CIKM 2023 时空数据论文总结

🌟【紧跟前沿】“时空探索之旅”与你一起探索时空奥秘!🚀
欢迎大家关注时空探索之旅时空探索之旅QRCode

  • 13
    点赞
  • 24
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
以下为 MATLAB 代码实现 Soft-DTW 算法计算两个时间序列间的距离: ```matlab function dist = softdtw(x, y, gamma) % SOFTDTW Computes the Soft-DTW distance between two time series % dist = softdtw(x, y, gamma) computes the Soft-DTW distance between time % series x and y with regularization parameter gamma. % The time series can either be row vectors or column vectors. % % This code is based on the original Soft-DTW paper: % M. Cuturi, M. Blondel "Soft-DTW: a Differentiable Loss Function for % Time-Series," ICML 2017. % Code available at: https://github.com/mblondel/soft-dtw % % Modified by: Yiyang Wang % Email: yiyang.wang@outlook.com % Date: 2022-06-24 if size(x, 1) > 1 x = x.'; end if size(y, 1) > 1 y = y.'; end szx = size(x); szy = size(y); B = pdist2(x, y).^2; F = zeros(szx(2)+1, szy(2)+1); F(1,:) = inf; F(:,1) = inf; F(1,1) = 0; for i = 2:szx(2)+1 for j = 2:szy(2)+1 u = [F(i-1,j), F(i-1,j-1), F(i,j-1)]; F(i,j) = B(i-1,j-1) + min(u); end end dist = F(end,end); if gamma > 0 G = zeros(szx(2)+1, szy(2)+1); G(1,:) = inf; G(:,1) = inf; G(1,1) = 0; for i = 2:szx(2)+1 for j = 2:szy(2)+1 u = [G(i-1,j), G(i-1,j-1), G(i,j-1)]; v = [F(i-1,j)-F(i,j), F(i-1,j-1)-F(i,j), F(i,j-1)-F(i,j)]; G(i,j) = B(i-1,j-1) + min(u + gamma*v.^2); end end dist = G(end,end); end end ``` 其中 `x` 和 `y` 分别是两个时间序列,可以是行向量或列向量。`gamma` 是正则化参数。需要注意的是,在计算之前,需要先将时间序列转成行向量或者列向量的形式。 调用该函数可以计算出两个时间序列之间的 Soft-DTW 距离,例如: ```matlab x = [1, 3, 4, 9]; y = [1, 4, 5, 6, 9]; gamma = 0.1; dist = softdtw(x, y, gamma); % 计算 x 和 y 之间的 Soft-DTW 距离 disp(dist); % 输出结果 ``` 输出结果为 `1.1178`,表示时间序列 `x` 和 `y` 之间的距离为 `1.1178`。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值