Program

The workshop will take place on Monday 18th of September in room Aula 5i.

Invited talk session (chair: Tony Bagnall)

• 9am: Introduction
• 9:10-10am: Invited Talk by Geoff Webb, “Convolutional kernels for effective and scalable time series analytics”

Human Activity Segmentation Challenge Session (chair: Arik Ermshaus)

• 10am-11am: Presentation of the winning solutions of the Human Activity Segmentation Challenge (15’ presentation per solution + 5’ questions)
  • Change points detection in multivariate signal applied to human activity segmentation, Grzegorz Haranczyk
  • Change Point Detection via Synthetic Signals, Ting-Ji Huang (Nanjing University); Qi-le Zhou (Nanjing University); Han-Jia Ye (Nanjing University); De-Chuan Zhan (Nanjing University)

Coffee break

Poster session (chair: Romain Tavenard)

• 11:30-1pm: Poster session (including 2’ spotlight presentation per poster) [link to spotlight slides]
  • Temporal Performance Prediction for Deep Convolutional Long Short-Term Memory Networks, Laura Fieback (Volkswagen AG); Bidya Dash (Volkswagen AG); Jakob Spiegelberg (Volkswagen AG); Hanno Gottschalk (University of Wuppertal)
  • Evaluating Explanation Methods for Multivariate Time Series Classification, Davide Italo DI Serramazza (University College Dublin); Thu Trang Nguyen (University College Dublin); Thach Le Nguyen (University College Dublin); Georgiana Ifrim (University College Dublin)
  • tGLAD: A sparse graph recovery based approach for multivariate time series segmentation, Shima Imani (Microsoft Research); Harsh Shrivastava (Microsoft Research)
  • Designing a New Search Space for Multivariate Time-Series Neural Architecture Search, Christopher J MacKinnon (University of Strathclyde); Robert Atkinson (University of Strathclyde)
  • Back to Basics: A Sanity Check on Modern Time Series Classification Algorithms, Bhaskar Dhariyal (University College Dublin); Thach Le Nguyen (Insight Centre); Georgiana Ifrim (University College Dublin)
  • Do Cows Have Fingerprints? Using Time Series Techniques and Milk Flow Profiles to Characterise Cow Behaviours and Detect Health Issues, Changhong Jin (University College Dublin); John Upton (Teagasc); Brian Mac Namee (University College Dublin)
  • Exploiting Context and Attention with Recurrent Neural Network for Sensor Time Series Prediction, Rashmi Dutta Baruah (Universidad Carlos III de Madrid); Mario Munoz Organero (Universidad Carlos III de Madrid)
  • Rail Crack Propagation Forecasting Using Multi-horizons RNNs, SARA YASMINE OUERK (IRT SystemX); Olivier Vo Van (SNCF SA); Mouadh Yagoubi (IRT SYSTEMX)
  • Electricity Load and Peak Forecasting: Feature Engineering, Probabilistic LightGBM and Temporal Hierarchies, Nicolo’ Rubattu (IDSIA USI-SUPSI); Gabriele Maroni (IDSIA USI_SUPSI); Giorgio Corani (IDSIA USI_SUPSI)
  • Time-aware Predictions of Moments of Change in Longitudinal User Posts on Social Media, Anthony R Hills (Queen Mary University of London); Adam Tsakalidis (Queen Mary University of London); Maria Liakata (Queen Mary University of London)

Lunch break

“Forecasting & Generation” oral session (chair: Georgiana Ifrim)

• 2:30-4pm: “Forecasting & Generation” oral session (15’ presentation per paper + 5’ questions)
  • Explainable Parallel RCNN with Novel Feature Representation for Time Series Forecasting, Jimeng Shi (Florida International University); Rukmangadh Myana (Florida International University); Vitalii Stebliankin (FIU); Azam Shirali (Florida International University); Giri Narasimhan (Florida International University)
  • Deep Long Term Prediction for Semantic Segmentation in Autonomous Driving (supplementary material), Bidya Dash (Volkswagen AG); Shreyas Bilagi (Volkswagen AG); Jasmin Breitenstein (Technische Universität Braunschweig); Volker Schomerus (Volkswagen AG); Thorsten Bagdonat (Volkswagen AG); Tim Fingscheidt ( Technische Universität Braunschweig)
  • Distribution-aware Evaluation of Multimodal Trajectory Predictions with Energy Score, Novin Shahroudi (University of Tartu); Meelis Kull (University of Tartu)
  • ShapeDBA: Generating Effective Time Series Prototypes using ShapeDTW Barycenter Averaging, Ali Ismail-Fawaz (IRIMAS, Université de Haute-Alsace); Hassan Ismail Fawaz (University of Haute Alsace); François Petitjean (Department of Data Science and Artificial Intelligence, Monash University); Maxime Devanne (Université de Haute Alsace); Jonathan Weber (University of Haute Alsace); Stefano Berretti (University of Florence, Italy); Geoffrey I Webb (Monash); Germain Forestier (University of Haute Alsace)

Coffee break

“Classification & Clustering” oral session (chair: Romain Tavenard)

• 4:30-6pm: “Classification & Clustering” oral session (15’ presentation per paper + 5’ questions)
  • Clustering time series with k-medoids based algorithms, Christopher L Holder (University of East Anglia); David Guijo-Rubio (Universidad de Córdoba); Anthony Bagnall (University of East Anglia)
  • RED CoMETS: an ensemble classifier for symbolically represented multivariate time series, Luca A Bennett (Awerian); Zahraa Abdallah (University of Bristol)
  • Extracting Features from Random Subseries: A Hybrid Pipeline for Time Series Classification and Extrinsic Regression, Matthew Middlehurst (University of East Anglia); Anthony Bagnall (University of East Anglia)
• 6pm: Conclusion of the workshop