2021年A類B類會(huì)議時(shí)間序列相關(guān)研究論文匯總
前些天學(xué)姐的學(xué)習(xí)交流群里面有同學(xué)提出學(xué)姐推的論文都太老了！可是前幾天推的都是經(jīng)典論文~肯定是舊一點(diǎn)的，但是新的論文也是學(xué)習(xí)了舊的論文發(fā)現(xiàn)了新的研究點(diǎn)寫出來的，所以經(jīng)典論文不可丟。
今天學(xué)姐整理了ACM CIKM、ICDE、ICML、KDD、IJCAI等期刊會(huì)議的時(shí)間序列2021年的論文給大家。記得收藏點(diǎn)贊哦！
CIKM
ACM CIKM（International Conference on Information and Knowledge Management）是信息知識(shí)管理和數(shù)據(jù)挖掘領(lǐng)域頂級(jí)學(xué)術(shù)會(huì)議之一，其目的是確定未來知識(shí)和信息系統(tǒng)發(fā)展面臨的具有挑戰(zhàn)性的問題，并通過征求和審查高質(zhì)量的應(yīng)用和理論研究成果來塑造未來的研究方向。被中國計(jì)算機(jī)協(xié)會(huì)CCF推薦為B類會(huì)議。
ACM CIKM 2021主要包含7個(gè)tracks，其中關(guān)注的三個(gè)tracks對(duì)應(yīng)的接收率為：
Full Paper Track（271/1251=21.7%），
Applied Paper Track（70/290=24.1%），
Short Paper Track（177/626=28.3%）。
下面整理了以上三個(gè)Tracks中時(shí)間序列相關(guān)的研究成果。主要研究方向包括：時(shí)間序列預(yù)測(cè)&異常檢測(cè)，時(shí)間序列分割，時(shí)間序列分類 及其交叉應(yīng)用領(lǐng)域。
01
論文標(biāo)題：
ClaSP - Time Series Segmentation
論文地址：
https://dl.acm.org/doi/10.1145/3459637.3482240
源碼鏈接：
https://github.com/alan-turing-institute/sktime
02
論文標(biāo)題
Learning to Learn the Future: Modeling Concept Drifts in Time Series Prediction
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482271
源碼鏈接
暫未開源
03
論文標(biāo)題
AdaRNN: Adaptive Learning and Forecasting of Time Series
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482315
源碼鏈接
https://github.com/jindongwang/transferlearning/tree/master/code/deep/adarnn
04
論文標(biāo)題
Learning Saliency Maps to Explain Deep Time Series Classifiers
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482446
源碼鏈接
https://github.com/kingspp/timeseries-explain
05
論文標(biāo)題
Actionable Insights in Urban Multivariate Time-series
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482410
源碼鏈接
https://github.com/AdityaLab/RaTSS
06
論文標(biāo)題
Integrating Static and Time-Series Data in Deep Recurrent Models for Oncology Early Warning Systems
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482441
源碼鏈接
暫未開源
07
論文標(biāo)題
Historical Inertia: A Neglected but Powerful Baseline for Long Sequence Time-series Forecasting
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482120
源碼鏈接
https://github.com/TakuyaShintate/tsts
08
論文標(biāo)題
Discovering Time-invariant Causal Structure from Temporal Data
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482086
源碼鏈接
暫未開源
09
論文標(biāo)題
AGCNT: Adaptive Graph Convolutional Network for Transformer-based Long Sequence Time-Series Forecasting
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482054
源碼鏈接
暫未開源
10
論文標(biāo)題
Improving Irregularly Sampled Time Series Learning with Time-Aware Dual-Attention Memory-Augmented Networks
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482079
源碼鏈接
暫未開源
11
論文標(biāo)題
BiCMTS: Bidirectional Coupled Multivariate Learning of Irregular Time Series with Missing Values
論文地址
https://dl.acm.org/doi/10.1145/3459637.3482064
源碼鏈接
暫未開源
ICDE
IEEE ICDE（IEEE International Conference on Data Engineering）是 IEEE 的旗艦會(huì)議，和 SIGMOD、VLDB 并稱數(shù)據(jù)庫領(lǐng)域三大頂會(huì)，旨在解決設(shè)計(jì)、構(gòu)建、管理和評(píng)估高級(jí)數(shù)據(jù)密集型系統(tǒng)和應(yīng)用程序中的研究問題，是研究人員、從業(yè)人員、開發(fā)人員和用戶探索前沿思想并交流技術(shù)、工具和經(jīng)驗(yàn)的領(lǐng)先論壇。被中國計(jì)算機(jī)協(xié)會(huì)CCF推薦為 A類會(huì)議。
IEEE ICDE 會(huì)議的多年平均接受率為 19.1%， 2021論文接收率暫無。
下面整理的是ICDE 2021 Research Track (Full & Short) & TKDE Posters中時(shí)間序列相關(guān)研究成果。
01
論文標(biāo)題
EnhanceNet: Plugin Neural Networks for Enhancing Correlated Time Series Forecasting
論文地址
https://ieeexplore.ieee.org/document/9458855
源碼鏈接
https://github.com/alan-turing-institute/sktime
02
論文標(biāo)題
Scalable Model-Based Management of Correlated Dimensional Time Series in ModelarDB+
論文地址
https://ieeexplore.ieee.org/document/9458830
源碼鏈接
https://github.com/skejserjensen/ModelarDB
03
論文標(biāo)題
TS-Benchmark: A Benchmark for Time Series Databases
論文地址
https://ieeexplore.ieee.org/document/9458659
源碼鏈接
https://github.com/timescale/tsbs
04
論文標(biāo)題
DAEMON: Unsupervised Anomaly Detection and Interpretation for Multivariate Time Series
論文地址
https://ieeexplore.ieee.org/document/9458835
源碼鏈接
暫無
05
論文標(biāo)題
GRAB: Finding Time Series Natural Structures via A Novel Graph-based Scheme
論文地址
https://ieeexplore.ieee.org/document/9458905
源碼鏈接
暫無
06
論文標(biāo)題
An Actor-Critic Ensemble Aggregation Model for Time-Series Forecasting
論文地址
https://ieeexplore.ieee.org/document/9458798
源碼鏈接
暫無
07
論文標(biāo)題
Efficient Shapelet Discovery for Time Series Classification
論文地址
https://ieeexplore.ieee.org/document/9458803
源碼鏈接
暫無
想要源碼的可以找原作者，給原作者發(fā)郵件申請(qǐng)嗷！
ICML
ICML（International Conference on Machine Learning, ICML）是由國際機(jī)器學(xué)習(xí)學(xué)會(huì)（IMLS）主辦的年度機(jī)器學(xué)習(xí)國際頂級(jí)會(huì)議，是機(jī)器學(xué)習(xí)領(lǐng)域發(fā)展的重要會(huì)議，且被中國計(jì)算機(jī)協(xié)會(huì)CCF推薦為 A類會(huì)議。
ICML 2021 的論文接收結(jié)果：共有5513篇有效投稿，其中1184篇論文被接收，接收率為21.4%。在被接收論文中，有166篇長文和1018篇短文。
以下整理了ICML 2021 Long Representation中時(shí)間序列相關(guān)研究成果。主要方向包括不局限于時(shí)序預(yù)測(cè)、時(shí)序異常檢測(cè)、時(shí)序分類、時(shí)序因果分析、多元時(shí)序等。
Papers
01
論文標(biāo)題
Voice2Series: Reprogramming Acoustic Models for Time Series Classification
論文地址
http://proceedings.mlr.press/v139/yang21j.html
源碼鏈接
https://github.com/huckiyang/Voice2Series-Reprogramming
02
論文標(biāo)題
Neural Rough Differential Equations for Long Time Series
論文地址
http://proceedings.mlr.press/v139/morrill21b.html
源碼鏈接
https://github.com/jambo6/neuralRDEs
03
論文標(biāo)題
Necessary and sufficient conditions for causal feature selection in time series with latent common causes
論文地址
http://proceedings.mlr.press/v139/mastakouri21a.html
源碼鏈接
暫未開源
04
論文標(biāo)題
Autoregressive Denoising Diffusion Models for Multivariate Probabilistic Time Series Forecasting
論文地址
http://proceedings.mlr.press/v139/rasul21a.html
源碼鏈接
暫未開源
05
論文標(biāo)題
Conformal prediction interval for dynamic time-series
論文地址
http://proceedings.mlr.press/v139/xu21h.html
源碼鏈接
https://github.com/hamrel-cxu/EnbPI
06
論文標(biāo)題
Z-GCNETs: Time Zigzags at Graph Convolutional Networks for Time Series Forecasting
論文地址
http://proceedings.mlr.press/v139/chen21o.html
源碼鏈接
https://github.com/Z-GCNETs/Z-GCNETs
07
論文標(biāo)題
End-to-End Learning of Coherent Probabilistic Forecasts for Hierarchical Time Series
論文地址
http://proceedings.mlr.press/v139/rangapuram21a.html
源碼鏈接
https://github.com/awslabs/gluon-ts
08
論文標(biāo)題
Approximation Theory of Convolutional Architectures for Time Series Modelling
論文地址
http://proceedings.mlr.press/v139/jiang21d.html
源碼鏈接
暫未開源
09
論文標(biāo)題
Whittle Networks: A Deep Likelihood Model for Time Series
論文地址
http://proceedings.mlr.press/v139/yu21c.html
源碼鏈接
https://github.com/ml-research/WhittleNetworks
10
論文標(biāo)題
Explaining Time Series Predictions with Dynamic Masks
論文地址
http://proceedings.mlr.press/v139/crabbe21a.html
源碼鏈接
https://github.com/JonathanCrabbe/Dynamask
KDD
ACM SIGKDD（Conference on Knowledge Discovery and Data Mining, KDD）是全球最大規(guī)模的國際數(shù)據(jù)科學(xué)會(huì)議，將展示知識(shí)發(fā)現(xiàn)和數(shù)據(jù)挖掘方面的最新研究成果。SIG是ACM中關(guān)于知識(shí)發(fā)現(xiàn)和數(shù)據(jù)挖掘的特別興趣小組，KDD知識(shí)發(fā)現(xiàn)和數(shù)據(jù)挖掘年度國際會(huì)議是數(shù)據(jù)挖掘、數(shù)據(jù)科學(xué)和分析領(lǐng)域的首屈一指的跨學(xué)科會(huì)議，被中國計(jì)算機(jī)協(xié)會(huì)CCF推薦為A類會(huì)議。
KDD 2021 的論文接收結(jié)果：共有1541篇有效投稿，其中238篇論文被接收，接收率為15.44%。相比 KDD 2020 的接收率 16.9% (216/1279)有所下降。
以下主要整理了KDD 2021 Research Track & Applied Data Science Track中時(shí)間序列相關(guān)研究成果。
01
論文標(biāo)題
MiniRocket: A Very Fast (Almost) Deterministic Transform for Time Series Classification
論文地址
https://dl.acm.org/doi/abs/10.1145/3447548.3467231
源碼鏈接
https://github.com/angus924/minirocket
02
論文標(biāo)題
Deep Learning Embeddings for Data Series Similarity Search
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467317
源碼鏈接
https://github.com/qtwang/SEAnet
03
論文標(biāo)題
Fast and Accurate Partial Fourier Transform for Time Series Data
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467293
源碼鏈接
https://github.com/snudatalab/PFT
04
論文標(biāo)題
A Transformer-based Framework for Multivariate Time Series Representation Learning
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467401
源碼鏈接
https://github.com/gzerveas/mvts_transformer
05
論文標(biāo)題
ST-Norm: Spatial and Temporal Normalization for Multi-variate Time Series Forecasting
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467330
源碼鏈接
https://github.com/JLDeng/ST-Norm
06
論文標(biāo)題
Statistical Models Coupling Allows for Complex Local Multivariate Time Series Analysis
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467362
源碼鏈接
暫未開源
07
論文標(biāo)題
Causal and Interpretable Rules for Time Series Analysis
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467161
源碼鏈接
暫未開源
08
論文標(biāo)題
Multivariate Time Series Anomaly Detection and Interpretation using Hierarchical Inter-Metric and Temporal Embedding
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467075
源碼鏈接
https://github.com/zhhlee/InterFusion
09
論文標(biāo)題
Practical Approach to Asynchronous Multivariate Time Series Anomaly Detection and Localization
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467174
源碼鏈接
https://github.com/eBay/RANSynCoders
10
論文標(biāo)題
Time Series Anomaly Detection for Cyber-physical Systems via Neural System Identification and Bayesian Filtering
論文地址
https://dl.acm.org/doi/10.1145/3447548.3467137
源碼鏈接
https://github.com/NSIBF/NSIBF
IJCAI
國際人工智能聯(lián)合會(huì)議 IJCAI（International Joint Conference on Artificial Intelligence）主要由國際人工智能聯(lián)合會(huì)議組織和東道國國家人工智能學(xué)會(huì)聯(lián)合主辦（每兩年舉行一次），旨在通過會(huì)議記錄、書籍、錄像和教材的方式傳播人工智能在會(huì)議上提出了尖端的科學(xué)成果。被中國計(jì)算機(jī)協(xié)會(huì)CCF推薦為A類會(huì)議。
IJCAI 2021 論文接收成果：共收到 4204 篇投稿，其中 587 篇論文被接收，接收率為 13.9%。相比于 IJCAI 2020 接受率 12.6%（592/4717）有所上升。
以下主要梳理了IJCAI 2021中Main Track & Survey Papers & Journal Papers的時(shí)間序列相關(guān)前沿研究成果。
主要方向包括：時(shí)間序列預(yù)測(cè)，時(shí)間序列分類，時(shí)序因果挖掘，多元時(shí)序分析，時(shí)序表示學(xué)習(xí)等。
01
論文標(biāo)題
Time-Aware Multi-Scale RNNs for Time Series Modeling
論文地址
https://www.ijcai.org/proceedings/2021/315
源碼鏈接
https://github.com/qianlima-lab/TAMS-RNNs
02
論文標(biāo)題
Time-Series Representation Learning via Temporal and Contextual Contrasting
論文地址
https://www.ijcai.org/proceedings/2021/324
源碼鏈接
https://github.com/emadeldeen24/TS-TCC
03
論文標(biāo)題
Adversarial Spectral Kernel Matching for Unsupervised Time Series Domain Adaptation
論文地址
https://www.ijcai.org/proceedings/2021/378
源碼鏈接
https://github.com/jarheadjoe/Adv-spec-ker-matching
04
論文標(biāo)題
Two Birds with One Stone: Series Saliency for Accurate and Interpretable Multivariate Time Series Forecasting
論文地址
https://www.ijcai.org/proceedings/2021/397
源碼鏈接
暫未開源
05
論文標(biāo)題
TE-ESN: Time Encoding Echo State Network for Prediction Based on Irregularly Sampled Time Series Data
論文地址
https://www.ijcai.org/proceedings/2021/414
源碼鏈接
暫未開源
06
論文標(biāo)題
Multi-series Time-aware Sequence Partitioning for Disease Progression Modeling
論文地址
https://www.ijcai.org/proceedings/2021/493
源碼鏈接
暫未開源
07
論文標(biāo)題
Time Series Data Augmentation for Deep Learning: A Survey
論文地址
https://www.ijcai.org/proceedings/2021/631
源碼鏈接
暫未開源
08
論文標(biāo)題
Uncertain Time Series Classification
論文地址
https://www.ijcai.org/proceedings/2021/683
源碼鏈接
https://github.com/frankl1/ustc
09
論文標(biāo)題
Learning Temporal Causal Sequence Relationships from Real-Time Time-Series
論文地址
https://arxiv.org/abs/1905.12262
源碼鏈接
暫未開源
本次推薦的論文夠大家看一周了叭？本來學(xué)姐想把這幾個(gè)會(huì)議拆開來發(fā)的，又怕大家覺得推的太少了，就一次性做了匯總！
以下資料：
關(guān)注公眾號(hào)【學(xué)姐帶你玩AI】進(jìn)AI交流群&領(lǐng)取學(xué)習(xí)資料
①拿來就能跑通的論文復(fù)現(xiàn)代碼講解及數(shù)據(jù)集。?
②獲得視頻課程的課件及思維導(dǎo)圖
③加入up主AI學(xué)習(xí)交流群。?
④能寫進(jìn)簡歷的企業(yè)級(jí)項(xiàng)目實(shí)戰(zhàn)。
?⑤從0到1帶你一步步成為數(shù)據(jù)分析師。?
⑥了解各種比賽baseline，提分的trick技巧。?
?⑦快速入門人工智能，夯實(shí)算法和代碼基礎(chǔ)。?
?⑧全部135+篇論文復(fù)現(xiàn)講解視頻。?
?⑨專屬學(xué)習(xí)路徑規(guī)劃，學(xué)習(xí)困惑解答。
⑩不定期免費(fèi)前沿論文直播。
文章內(nèi)容來源網(wǎng)站：https://dreamhomes.top/
免責(zé)聲明：所載內(nèi)容來源互聯(lián)網(wǎng)，僅供參考。轉(zhuǎn)載稿件版權(quán)歸原作者和機(jī)構(gòu)所有，如有侵權(quán)，請(qǐng)聯(lián)系我們刪除。
