Held in conjunction with KDD'17
Aug 14, 2017 - Halifax, Nova Scotia, Canada
3rd SIGKDD Workshop on
Mining and Learning from Time Series

Introduction

Time series data are ubiquitous. In domains as diverse as finance, entertainment, transportation and health care, we observe a fundamental shift away from parsimonious, infrequent measurement to nearly continuous monitoring and recording. Rapid advances in diverse sensing technologies, ranging from remote sensors to wearables and social sensing, are generating a rapid growth in the size and complexity of time series archives. Thus, although time series analysis has been studied extensively, its importance only continues to grow. What is more, modern time series data pose significant challenges to existing techniques (e.g., irregular sampling in hospital records and spatiotemporal structure in climate data). Finally, time series mining research is challenging and rewarding because it bridges a variety of disciplines and demands interdisciplinary solutions. Now is the time to discuss the next generation of temporal mining algorithms. The focus of MiLeTS workshop is to synergize the research in this area and discuss both new and open problems in time series analysis and mining. The solutions to these problems may be algorithmic, theoretical, statistical, or systems-based in nature. Further, MiLeTS emphasizes applications to high impact or relatively new domains, including but not limited to biology, health and medicine, climate and weather, road traffic, astronomy, and energy.
The MiLeTS workshop will discuss a broad variety of topics related to time series, including:

  • Time series pattern mining and detection, representation, searching and indexing, classification, clustering, prediction, forecasting, and rule mining.
  • BIG time series data.
  • Hardware acceleration techniques using GPUs, FPGAs and special processors.
  • Online, high-speed learning and mining from streaming time series.
  • Uncertain time series mining.
  • Privacy preserving time series mining and learning.
  • Time series that are multivariate, high-dimensional, heterogeneous, etc., or that possess other atypical properties.
  • Time series with special structure: spatiotemporal (e.g., wind patterns at different locations), relational (e.g., patients with similar diseases), hierarchical, etc.
  • Time series with sparse or irregular sampling, non-random missing values, and special types of measurement noise or bias.
  • Time series analysis using less traditional approaches, such as deep learning and subspace clustering.
  • Applications to high impact or relatively new time series domains, such as health and medicine, road traffic, and air quality.
  • New, open, or unsolved problems in time series analysis and mining.

Schedule

Room 200D
8:00 AM - 16:30 PM
August 14, 2017

 

Morning Session

08:00-08:10  Opening remarks
08:10-09:00  Invited Talk Albert Bifet
09:00-10:00  Paper Presentation 1: Deep Learning and Nonparametric Solutions for time series

      Reading the Tea Leaves: A Neural Network Perspective on Technical Trading. Sid Ghoshal and Stephen Roberts
      DECADE: A Deep Metric Learning Model for Multivariate Time Series. Zhengping Che, Xinran He, Ke Xu and Yan Liu
      Robust Parameter-Free Season Length Detection in Time Series. Maximilian Toller and Roman Kern

10:00-10:30  Coffee Break
10:30-11:30  Keynote Talk Jian Pei
     Tracking in Dynamic Networks
11:30-12:00  Poster Highlights (Short Presentations):

       Utilizing Artificial Neural Networks to Detect Compound Events in Spatio-Temporal Soccer Data. Keven Richly, Florian Moritz and Christian Schwarz
       Automatic Singular Spectrum Analysis and Forecasting. Michele Trovero, Michael Leonard and Bruce Elsheimer
       Short-Term Wind Energy Forecasting with Temporally Dependent Neural Network Models. Rui Li, Pu Wang, Jingrui Xie, Alex Chien and Mustafa Kabul
       Time Series Classification for Scrap Rate Prediction in Transfer Molding. Anna Mandli, Robert Palovics, Matyas Susits and Andras A. Benczur

12:00-13:00  Lunch Break

Afternoon Session

13:00-14:00  Invited Talk Benjamin Marlin
     Learning With Temporally Uncertain Labels
14:00-15:00  Paper Presentation 2: Streaming and Trajectory data analysis

       Online Thinning for High Volume Streaming Data. Xin Hunt and Rebecca Willett
       Sub-string/Pattern Matching in Sub-linear Time Using a Sparse Fourier Transform Approach. Nagaraj Thenkarai Janakiraman, Avinash Vem, Krishna Narayanan and Jean-Francois Chamberland
       Coordination Event Detection and Initiator Identification in Time Series Data. Chainarong Amornbunchornvej, Ivan Brugere, Ariana Strandburg-Peshkin, Damien Farine, Margaret Crofoot and Tanya Berger-Wolf

15:00-15:30  Coffee Break and Poster Session
15:30-16:00  Poster Session
16:00-16:30  Panel Discussion and Concluding Remarks

 

Speakers

Jian Pei

Jian Pei

Professor
Simon Fraser University

Tracking in Dynamic Networks

In many application scenarios ranging from social networks to IoT, we need to process and analyze a huge amount of data, connected, evolving, linkages being more interesting than entities individually. Modeling such temporal data in nature as graphs provides a conceptually convenient way to support novel and meaningful intelligent applications. At the same time, it also posts grant challenges in many aspects, such as algorithm design and computing system development. In this talk, I will present some interesting and novel application scenarios where graphs play a central role, as well as the corresponding algorithms. Moreover, I will briefly introduce our on-going effort to build a distributed cloud-based graph computing engine that can query huge graphs and networks in seconds.
Bio
Jian Pei is currently a Canada Research Chair (Tier 1) in Big Data Science and a Professor in the School of Computing Science at Simon Fraser University. He is also an associate member of the Department of Statistics and Actuarial Science, Faculty of Science, and Faculty of Health Sciences. He is a well known leading researcher in the general areas of data science, big data, data mining, and database systems. His expertise is on developing effective and efficient data analysis techniques for novel data intensive applications. He is recognized as a Fellow of the Association of Computing Machinery (ACM) for his “contributions to the foundation, methodology and applications of data mining” and as a Fellow of the Institute of Electrical and Electronics Engineers (IEEE) for his “contributions to data mining and knowledge discovery”. At the same time, he is also renowned for his professional leadership.
Jian Pei is one of the most cited authors in data mining, database systems, and information retrieval. Since 2000, he has published one textbook, two monographs and over 200 research papers in refereed journals and conferences, which have been cited by more than 60,000 in literature. His H-index is 69 according to Google Scholar as of June 6, 2016. His research has generated remarkable impact substantially beyond academia. For example, his algorithms have been adopted by industry in production and popular open source software suites. More details can be found at: https://www.cs.sfu.ca/~jpei/main.htm

Albert Bifet

Albert Bifet

Associate Professor
Télécom ParisTech

Massive Online Analytics for the Internet of Things (IoT)

Big Data and the Internet of Things (IoT) have the potential to fundamentally shift the way we interact with our surroundings. The challenge of deriving insights from the Internet of Things (IoT) has been recognized as one of the most exciting and key opportunities for both academia and industry. Advanced analysis of big data streams from sensors and devices is bound to become a key area of data mining research as the number of applications requiring such processing increases. Dealing with the evolution over time of such data streams, i.e., with concepts that drift or change completely, is one of the core issues in stream mining. In this talk, I will present an overview of data stream mining, and I will introduce some popular open source tools for data stream mining.
Bio
Albert Bifet is Associate Professor at Telecom ParisTech and Honorary Research Associate at the WEKA Machine Learning Group at University of Waikato. Previously he worked at Huawei Noah's Ark Lab in Hong Kong, Yahoo Labs in Barcelona, University of Waikato and UPC BarcelonaTech. He is the author of a book on Adaptive Stream Mining and Pattern Learning and Mining from Evolving Data Streams. He is one of the leaders of MOA and Apache SAMOA software environments for implementing algorithms and running experiments for online learning from evolving data streams. He was serving as Co-Chair of the Industrial track of IEEE MDM 2016, ECML PKDD 2015, and as Co-Chair of BigMine (2015, 2014, 2013, 2012), and ACM SAC Data Streams Track (2017, 2016, 2015, 2014, 2013, 2012). . More details can be found at: http://albertbifet.com/

 Benjamin Marlin

Benjamin Marlin

Assistant Professor
University of Massachusetts Amherst

Learning With Temporally Uncertain Labels

In this talk, I will begin with an overview of the challenges of learning time series event detection models in the emerging area of mobile health or mHealth. mHealth technologies, including on-body physiological sensors, have the potential to yield significant insights into health and behavior. They also offer compelling possibilities for informing the delivery of adaptive health interventions. However, the analysis of mHealth data is often subject to an array of complicating factors at the level of both sensor data and event labels. This talk will focus on the problem of learning time series detection models from temporally imprecise event labels. In this problem, the data consist of a set of input time series, and supervision is provided by a sequence of noisy event time stamps corresponding to the occurrence of positive class events. Such temporally imprecise labels occur in mHealth research both when observers are tasked with precisely labeling the occurrence of short duration events (such as individual puffs on a cigarette), as well as when study subjects are tasked with self-reporting events retrospectively (such as the start and end times of the last time they smoked).
Bio
Benjamin M. Marlin joined the College of Information and Computer Sciences at the University of Massachusetts Amherst in 2011. His research area is machine learning with a focus on the development of customized probabilistic models and approximate inference and learning algorithms. His current research centers on probabilistic models for time series with applications in health and behavioral science. His research has been supported by the National Science Foundation, the National Institutes of Health, the Patient-Centered Outcomes Research Institute, the Intelligence Advanced Research Projects Activity, and the US Army Research Laboratory. He is a 2014 NSF CAREER award recipient and a 2013 Yahoo! Faculty Research Engagement Program award recipient. He was previously a fellow of the Pacific Institute for the Mathematical Sciences and the Killam Trusts at the University of British Columbia. He completed his PhD in machine learning in the Department of Computer Science at the University of Toronto.

-->

Accepted Papers

 

Reading the Tea Leaves: A Neural Network Perspective on Technical Trading Sid Ghoshal and Stephen Roberts

DECADE: A Deep Metric Learning Model for Multivariate Time Series Zhengping Che, Xinran He, Ke Xu and Yan Liu

Robust Parameter-Free Season Length Detection in Time Series Maximilian Toller and Roman Kern

Online Thinning for High Volume Streaming Data Xin Hunt and Rebecca Willett

Sub-string/Pattern Matching in Sub-linear Time Using a Sparse Fourier Transform Approach Nagaraj Thenkarai Janakiraman, Avinash Vem, Krishna Narayanan and Jean-Francois Chamberland

Coordination Event Detection and Initiator Identification in Time Series Data Chainarong Amornbunchornvej, Ivan Brugere, Ariana Strandburg-Peshkin, Damien Farine, Margaret Crofoot and Tanya Berger-Wolf

 

Accepted Posters

 

Utilizing Artificial Neural Networks to Detect Compound Events in Spatio-Temporal Soccer Data. Keven Richly, Florian Moritz and Christian Schwarz

Automatic Singular Spectrum Analysis and Forecasting. Michele Trovero, Michael Leonard and Bruce Elsheimer

Short-Term Wind Energy Forecasting with Temporally Dependent Neural Network Models. Rui Li, Pu Wang, Jingrui Xie, Alex Chien and Mustafa Kabul

Time Series Classification for Scrap Rate Prediction in Transfer Molding. Anna Mandli, Robert Palovics, Matyas Susits and Andras A. Benczur

 

Call for Papers

Submissions should follow the SIGKDD formatting requirements and will be evaluated using the SIGKDD Research Track evaluation criteria. Preference will be given to papers that are reproducible, and authors are encouraged to share their data and code publicly whenever possible.

Note on open problem submissions: In order to promote new and innovative research on time series, we plan to accept a small number of high quality manuscripts describing open problems in time series analysis and mining. Such papers should provide a clear, detailed description and analysis of a new or open problem that poses a significant challenge to existing techniques, as well as a thorough empirical investigation demonstrating that current methods are insufficient.

Submissions will be managed via the MiLeTS 2017 EasyChair website.

Key Dates

 

Paper Submission Deadline: May 28, 2017, 11:59 PM PST

Author Notification: June 16, 2017, 11:59 PM PST

Camera Ready Version: June 23, 2017, 11:59 PM PST

Workshop: August 14, 2017

Workshop Organizers

 

Eamonn Keogh

University of California Riverside

 

Yan Liu

University of Southern California

 

Abdullah Mueen

University of New Mexico

 

Sanjay Purushotham

University of Southern California

 

Vijay Manikandan Janakiraman

NASA Ames Research Center