Cadence: A Practical Time-series Partitioning Algorithm for Unlabeled IoT Sensor Streams

Timeseries partitioning is an essential step in most machine-learning driven, sensor-based IoT applications. This paper introduces a sample-efficient, robust, time-series segmentation model and algorithm. We show that by learning a representation specifically with the segmentation objective based on maximummean discrepancy (MMD), our algorithm can robustly detect time-series events across different applications. Our loss function allows us to infer whether consecutive sequences of samples are drawn from the same distribution (null hypothesis) and determines the change-point between pairs that reject the null hypothesis (i.e., come from different distributions). We demonstrate its applicability in a real-world IoT deployment for ambient-sensing based activity recognition. Moreover, while many works on change-point detection exist in the literature, our model is significantly simpler and matches or outperforms state-of-the-art methods. We can fully train our model in 9–93 seconds on average with little variation in hyperparameters for data across different applications.

[1]  Sanglu Lu,et al.  Unsupervised Human Activity Representation Learning with Multi-task Deep Clustering , 2021, Proc. ACM Interact. Mob. Wearable Ubiquitous Technol..

[2]  Ali Ghodsi,et al.  Detecting Change-Points in Time Series by Maximum Mean Discrepancy of Ordinal Pattern Distributions , 2012, UAI.

[3]  Max Welling,et al.  Auto-Encoding Variational Bayes , 2013, ICLR.

[4]  Xiaonan Guo,et al.  Device-free Personalized Fitness Assistant Using WiFi , 2018, Proc. ACM Interact. Mob. Wearable Ubiquitous Technol..

[5]  M. Srivastava,et al.  SenseHAR: a robust virtual activity sensor for smartphones and wearables , 2019, SenSys.

[6]  Yanwen Wang,et al.  Modeling RFID Signal Reflection for Contact-free Activity Recognition , 2018, Proc. ACM Interact. Mob. Wearable Ubiquitous Technol..

[7]  Michèle Basseville,et al.  Detection of abrupt changes: theory and application , 1993 .

[8]  Jie Zhang,et al.  Deep Learning-Based Real-Time Building Occupancy Detection Using AMI Data , 2020, IEEE Transactions on Smart Grid.

[9]  Ruby B. Lee,et al.  Time Series Segmentation through Automatic Feature Learning , 2018, ArXiv.

[10]  Kenji Fukumizu,et al.  Characteristic Kernels and Infinitely Divisible Distributions , 2014, J. Mach. Learn. Res..

[11]  Marimuthu Palaniswami,et al.  Internet of Things (IoT): A vision, architectural elements, and future directions , 2012, Future Gener. Comput. Syst..

[12]  Eamonn J. Keogh,et al.  Clustering of time-series subsequences is meaningless: implications for previous and future research , 2004, Knowledge and Information Systems.

[13]  Yoshua Bengio,et al.  Why Does Unsupervised Pre-training Help Deep Learning? , 2010, AISTATS.

[14]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[15]  H. Vincent Poor,et al.  Nonparametric Detection of Anomalous Data via Kernel Mean Embedding , 2014, ArXiv.

[16]  Iqbal H. Sarker,et al.  Individualized Time-Series Segmentation for Mining Mobile Phone User Behavior , 2018, Comput. J..

[17]  Qing Liu,et al.  Exploring Transfer Learning to Reduce Training Overhead of HPC Data in Machine Learning , 2019, 2019 IEEE International Conference on Networking, Architecture and Storage (NAS).

[18]  Ryan P. Adams,et al.  Bayesian Online Changepoint Detection , 2007, 0710.3742.

[19]  Sivaraman Balakrishnan,et al.  Optimal kernel choice for large-scale two-sample tests , 2012, NIPS.

[20]  Nigel Collier,et al.  Change-Point Detection in Time-Series Data by Relative Density-Ratio Estimation , 2012, Neural Networks.

[21]  Sayantan Banerjee,et al.  Change‐point analysis in financial networks , 2019, Stat.

[22]  Pascal Vincent,et al.  Stacked Denoising Autoencoders: Learning Useful Representations in a Deep Network with a Local Denoising Criterion , 2010, J. Mach. Learn. Res..

[23]  John A. Stankovic,et al.  Research Directions for the Internet of Things , 2014, IEEE Internet of Things Journal.

[24]  Geoffrey I. Webb,et al.  Characterizing concept drift , 2015, Data Mining and Knowledge Discovery.

[25]  Hideki Takayasu.  Basic methods of change-point detection of financial fluctuations , 2015, 2015 International Conference on Noise and Fluctuations (ICNF).

[26]  Seetha Hari,et al.  Learning From Imbalanced Data , 2019, Advances in Computer and Electrical Engineering.

[27]  Eamonn J. Keogh,et al.  Segmenting Time Series: A Survey and Novel Approach , 2002 .

[28]  Yiming Yang,et al.  Kernel Change-point Detection with Auxiliary Deep Generative Models , 2019, ICLR.

[29]  Yoshua Bengio,et al.  A Recurrent Latent Variable Model for Sequential Data , 2015, NIPS.

[30]  Marco Gruteser,et al.  HandSense: capacitive coupling-based dynamic, micro finger gesture recognition , 2019, SenSys.

[31]  Guokun Lai,et al.  Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks , 2017, SIGIR.

[32]  Alexander Bertrand,et al.  Change Point Detection in Time Series Data Using Autoencoders With a Time-Invariant Representation , 2020, IEEE Transactions on Signal Processing.

[33]  Yusha Liu,et al.  Classifier Two Sample Test for Video Anomaly Detections , 2018, BMVC.

[34]  Stephen J. Roberts,et al.  Anomaly Detection for Time Series Using VAE-LSTM Hybrid Model , 2020, ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[35]  Douglas D. O'Shaughnessy,et al.  A highly non-stationary noise tracking and compensation algorithm, with applications to speech enhancement and on-line ASR , 2012, 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[36]  Kenji Yamanishi,et al.  A unifying framework for detecting outliers and change points from non-stationary time series data , 2002, KDD.

[37]  Takafumi Kanamori,et al.  A Least-squares Approach to Direct Importance Estimation , 2009, J. Mach. Learn. Res..

[38]  Shikha Mehta,et al.  Concept drift in Streaming Data Classification: Algorithms, Platforms and Issues , 2017, ITQM.

[39]  Zaïd Harchaoui,et al.  A Fast, Consistent Kernel Two-Sample Test , 2009, NIPS.

[40]  Hae Young Noh,et al.  Fine-Grained Recognition of Activities of Daily Living through Structural Vibration and Electrical Sensing , 2019, BuildSys@SenSys.

[41]  Geoffrey E. Hinton,et al.  Reducing the Dimensionality of Data with Neural Networks , 2006, Science.

[42]  Yanchun Liang,et al.  Non-Parametric Change-Point Method for Differential Gene Expression Detection , 2011, PloS one.

[43]  F. Gustafsson.  The marginalized likelihood ratio test for detecting abrupt changes , 1996, IEEE Trans. Autom. Control..

[44]  Le Song,et al.  M-Statistic for Kernel Change-Point Detection , 2015, NIPS.

[45]  A. Willsky,et al.  A generalized likelihood ratio approach to the detection and estimation of jumps in linear systems , 1976 .

[46]  Charles W. Champ,et al.  A multivariate exponentially weighted moving average control chart , 1992 .

[47]  Yasuyuki Matsushita,et al.  Detecting State Changes of Indoor Everyday Objects using Wi-Fi Channel State Information , 2017, Proc. ACM Interact. Mob. Wearable Ubiquitous Technol..

[48]  Marc'Aurelio Ranzato,et al.  Building high-level features using large scale unsupervised learning , 2011, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[49]  Cecilia Mascolo,et al.  SelfHAR: Improving Human Activity Recognition through Self-training with Unlabeled Data , 2021, Proc. ACM Interact. Mob. Wearable Ubiquitous Technol..

[50]  Samaneh Aminikhanghahi,et al.  Real-Time Change Point Detection with Application to Smart Home Time Series Data , 2019, IEEE Transactions on Knowledge and Data Engineering.

[51]  Zhen Li,et al.  Daily life event segmentation for lifestyle evaluation based on multi-sensor data recorded by a wearable device , 2013, 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[52]  Mark Goadrich,et al.  The relationship between Precision-Recall and ROC curves , 2006, ICML.

[53]  Robert C. Wilson,et al.  An Approximately Bayesian Delta-Rule Model Explains the Dynamics of Belief Updating in a Changing Environment , 2010, The Journal of Neuroscience.

[54]  J. F. Baldwin.  Future Directions for Machine Learning , 2000 .

[55]  Takafumi Kanamori,et al.  Density Ratio Estimation in Machine Learning , 2012 .

[56]  V. Moskvina,et al.  An Algorithm Based on Singular Spectrum Analysis for Change-Point Detection , 2003 .

[57]  Carl E. Rasmussen,et al.  Gaussian Process Change Point Models , 2010, ICML.

[58]  Geoffrey E. Hinton,et al.  Visualizing Data using t-SNE , 2008 .

[59]  Roman Garnett,et al.  A Machine Learning Approach to Pattern Detection and Prediction for Environmental Monitoring and Water Sustainability , 2011 .

[60]  Lida Xu,et al.  The internet of things: a survey , 2014, Information Systems Frontiers.

[61]  Qinfeng Shi,et al.  Attend and Discriminate , 2020, Proc. ACM Interact. Mob. Wearable Ubiquitous Technol..

[62]  Zoran A. Salcic,et al.  Adaptive sliding window segmentation for physical activity recognition using a single tri-axial accelerometer , 2017, Pervasive Mob. Comput..

[63]  Ricardo Cao,et al.  Nonparametric Density Estimation , 2013 .

[64]  Anthony Widjaja,et al.  Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond , 2003, IEEE Transactions on Neural Networks.

[65]  Roi Livni,et al.  On the Computational Efficiency of Training Neural Networks , 2014, NIPS.

[66]  Zaïd Harchaoui,et al.  Kernel Change-point Analysis , 2008, NIPS.

[67]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[68]  Motoaki Kawanabe,et al.  Direct Importance Estimation with Model Selection and Its Application to Covariate Shift Adaptation , 2007, NIPS.

[69]  Lina Yao,et al.  Adversarial Multi-view Networks for Activity Recognition , 2020, Proc. ACM Interact. Mob. Wearable Ubiquitous Technol..

[70]  Ruohong Zhang,et al.  Explainable Unsupervised Change-point Detection via Graph Neural Networks , 2020, ArXiv.

[71]  Nicholas D. Lane,et al.  Recurrent network based automatic detection of chronic pain protective behavior using MoCap and sEMG data , 2019, UbiComp.

[72]  G. Fasshauer.  Positive definite kernels: past, present and future , 2011 .

[73]  Graham J. Williams,et al.  On-Line Unsupervised Outlier Detection Using Finite Mixtures with Discounting Learning Algorithms , 2000, KDD '00.

[74]  Diane J. Cook,et al.  A survey of methods for time series change point detection , 2017, Knowledge and Information Systems.

[75]  Fernando Pérez-Cruz,et al.  Kullback-Leibler divergence estimation of continuous distributions , 2008, 2008 IEEE International Symposium on Information Theory.

[76]  Robert Weigel,et al.  Fusion of Nonintrusive Environmental Sensors for Occupancy Detection in Smart Homes , 2018, IEEE Internet of Things Journal.

[77]  G. Abowd,et al.  IMUTube , 2020, Proc. ACM Interact. Mob. Wearable Ubiquitous Technol..

[78]  Margaret Hamilton,et al.  Human occupancy recognition with multivariate ambient sensors , 2016, 2016 IEEE International Conference on Pervasive Computing and Communication Workshops (PerCom Workshops).

[79]  Bernhard Schölkopf,et al.  A Kernel Two-Sample Test , 2012, J. Mach. Learn. Res..

[80]  Jian Sun,et al.  Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[81]  MascoloCecilia,et al.  What Will You Do for the Rest of the Day , 2018 .

[82]  Lei Zhang,et al.  A Heterogeneous Ensemble Approach for Activity Recognition with Integration of Change Point-Based Data Segmentation , 2018, Applied Sciences.

[83]  Tsuyoshi Idé,et al.  Change-Point Detection using Krylov Subspace Learning , 2007, SDM.

[84]  Diane J. Cook,et al.  Using change point detection to automate daily activity segmentation , 2017, 2017 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops).