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:
08:00-08:10 Opening remarks
08:10-09:10 Keynote Talk
09:15-10:00 Keynote Talk
10:00-10:30 Coffee Break
10:30-12:00 Contributed Talks
12:00-13:00 Lunch Break
13:00-14:00 Keynote Talk
14:00-14:30 Poster Spotlights
14:30-15:00 Coffee Break
15:00-16:00 Keynote Talk
16:00-16:45 Poster Session
16:45-17:00 Concluding Remarks
Professor, Amazon Fellow
University of Texas at Austin & Amazon
Forecasting high-dimensional time series plays a crucial role in many applications such as demand forecasting and financial predictions. Modern datasets can have millions of correlated time-series that evolve together, i.e they are extremely high dimensional (one dimension for each individual time-series). There is a need for exploiting global patterns and coupling them with local calibration for better prediction. However, most recent deep learning approaches in the literature are one-dimensional, i.e, even though they are trained on the whole dataset, during prediction, the future forecast for a single dimension mainly depends on past values from the same dimension. In this paper, we seek to correct this deficiency and propose DeepGLO, a deep forecasting model which thinks globally and acts locally. In particular, DeepGLO is a hybrid model that combines a global matrix factorization model regularized by a temporal convolution network, along with another temporal network that can capture local properties of each time-series and associated covariates. Our model can be trained effectively on high-dimensional but diverse time series, where different time series can have vastly different scales, without a priori normalization or rescaling. Empirical results demonstrate that DeepGLO can outperform state-of-the-art approaches; for example, we see more than 25% improvement in WAPE over other methods on a public dataset that contains more than 100K-dimensional time series. This is joint work with Rajat Sen and Hsiang-Fu Yu.Bio
Staff Research Scientist
Google Brain & CNRS
Transformer models have been used in a variety of fields and yield great results on many NLP tasks. But between the BERT, GPT-2, and many other variants, they can be inefficient and it can be hard to apply them. I will introduce a new efficient variant of Transformer called the Reformer. I'll take you through the code that implements it and I will show how it runs at high efficiency and addresses the main problems or high memory use and low performance on long sequences that limited the use of some Transformers before. I will finish with new applications of Reformer that open up.
University of Cambridge, UCLA, The Alan Turing Institute
Leader of Data Sciences Group
NASA Ames Research Center
The National Airspace, with all of its aircraft, airports, personnel, and related infrastructure, is an incredibly safe system. It is essential that we keep it safe through the changes that it is experiencing, including expected increases in traffic over time, increasing variety of traffic in the form of Unmanned Aerial Vehicles (UAVs), and sharp reductions and increases in traffic due to transient phenomena such as pandemics. The system is currently monitored for safety issues through a set of exceedances, which are rules describing various known safety issues. By definition, these rules cannot identify previously-unknown safety issues. Additionally, they do not identify precursors to these safety issues—states that may not represent safety issues by themselves, but are circumstances under which safety issues are more likely to occur in the near future. In this talk, I describe machine learning-based methods that we have developed for anomaly detection and precursor identification and the aviation safety results that we have obtained. I also describe the active learning algorithm that we have developed to mitigate the false alarm problem that is common to data-driven anomaly detection methods. Our ultimate aim is to allow for aviation safety analysts to discover new safety issues as they arise.
The Effectiveness of Discretization in Forecasting: An Empirical Study on Neural Time Series Models Stephan Rabanser, Tim Januschowski, Valentin Flunkert, David Salinas and Jan Gasthaus. [Video]
Driver2vec: Driver Identification from Automotive Data Jingbo Yang, Ruge Zhao, Meixian Zhu, Jaka Sodnik, David Hallac and Jure Leskovec. [Video]
Improving Robustness on Seasonality-Heavy Multivariate Time Series Anomaly Detection Farzaneh Khoshnevisan, Zhewen Fan and Vitor Carvalho. [Video]
Detection of Environment Transitions in Time Series Data for Responsive Science Ameya Daigavane, Kiri Wagstaff, Gary Doran, Corey Cochrane, Caitriona Jackman and Abigail Rymer. [Video]
Instance Explainable Temporal Network For Multivariate Timeseries Naveen Madiraju and Homa Karimabadi. [Video]
Modality Selection for Classification on Time-series Data Murchana Baruah and Bonny Banerjee. [Video]
Machine Learning Methods for Predictive Maintenance of Multifunctional Printers Wojciech Indyk, Zuzana Neverilova and Jakub Valcik. [Video]
Towards Deep Unsupervised Representation Learning from Accelerometer Time Series for Animal Activity Recognition Jacob Kamminga, Viet Duc Le, Nirvana Meratnia and Paul Havinga. [Video]
Statistical Evaluation of Anomaly Detectors for Sequences Erik Scharwächter and Emmanuel Müller. [Video]
Towards a framework for incorporating data acquisition cost in predictive time series models Adrian Stetco, Razvan Mosincat, Goran Nenadic and John Keane. [Video]
PastProp-RNN: improved predictions of the future by correcting the past André Baptista, Yassine Baghoussi, Carlos Soares, Miguel Arantes and João Mendes-Moreira. [Video]
RobustTAD: Robust Time Series Anomaly Detection via Decomposition and Convolutional Neural Networks Jingkun Gao, Xiaomin Song, Qingsong Wen, Pichao Wang, Liang Sun and Huan Xu. [Video]
Towards Automating Time Series Analysis for Paleogeosciences Deborah Khider, Pratheek Athreya, Varun Ratnakar, Yolanda Gil, Feng Zhu, Myron Kwan and Julien Emile-Geay. [Video]
Interpreting Deep Temporal Neural Networks by Selective Visualization of Internally Activated Nodes Sohee Cho, Wonjoon Chang, Ginkyeng Lee and Jaesik Choi. [Video]
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. Submissions are strongly recommended to be no more than 4 pages, excluding references or supplementary materials (all in a single pdf). The appropriateness of using additional pages over the recommended length will be judged by reviewers. All submissions must be in pdf format using theAll submissions must be in pdf format using the workshop template ( latex, word). Submissions will be managed via the MiLeTS 2020 EasyChair website: https://easychair.org/conferences/?conf=milets2020.
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.
COVID-19 Time Series Analysis Special Track: The COVID-19 pandemic is impacting almost everyone worldwide and is expected to have life-altering short and long-term effects. There are many potential applications of time series analysis and mining that can contribute to understanding of this pandemic. We encourage submission of high quality manuscripts describing original problems, time series datasets, and novel solutions for time series analysis and forecasting of COVID-19.
The review process is single-round and double-blind (submission files have to be anonymized). Concurrent submissions to other journals and conferences are acceptable. Accepted papers will be presented as posters during the workshop and list on the website. Besides, a small number of accepted papers will be selected to be presented as contributed talks.
Any questions may be directed to the workshop e-mail address: firstname.lastname@example.org.
Paper Submission Deadline: June 10th June 20th, 2020, 11:59PM Alofi Time
Author Notification: July 13th, 2020
Camera Ready Version: July 27th, 2020
Video Submission: July 30th, 2020
Workshop: August 24th, 2020