Projects | Romain Tavenard

Machine Learning for Environmental Time Series

A lot of environmental data are timestamped. Designing ML techniques that can handle this time dimension can often lead to much improved performance. We have so far turned our focus on 2 different types of environmental data: chemistry data in streams and remote sensing data (such as satellite image time series).

Related papers

A. Bailly - D. Arvor - L. Chapel - R. Tavenard. Classification of MODIS Time Series with Dense Bag-of-Temporal-SIFT-Words: Application to Cropland Mapping in the Brazilian Amazon. IEEE International Geoscience and Remote Sensing Symposium 2016.
Show/Hide BiBTeX

@inproceedings{bailly:halshs-01343211,
  TITLE = {{Classification of MODIS Time Series with Dense Bag-of-Temporal-SIFT-Words: Application to Cropland Mapping in the Brazilian Amazon}},
  AUTHOR = {Bailly, Adeline and Arvor, Damien and Chapel, Laetitia and Tavenard, Romain},
  URL = {https://shs.hal.science/halshs-01343211},
  BOOKTITLE = {{IEEE International Geoscience and Remote Sensing Symposium}},
  ADDRESS = {Beijing, China},
  YEAR = {2016},
  MONTH = Jul,
  DOI = {10.1109/IGARSS.2016.7729594},
  KEYWORDS = {Time Series Classification ; MODIS ; Vegetation Index ; D-BoTSW},
  PDF = {https://shs.hal.science/halshs-01343211v1/file/classification-modis-time.pdf},
  HAL_ID = {halshs-01343211},
  HAL_VERSION = {v1},
}

R. Dupas - R. Tavenard - O. Fovet - N. Gilliet - C. Grimaldi - C. Gascuel. Identifying seasonal patterns of phosphorus storm dynamics with dynamic time warping. Water Resources Research 2015.
Show/Hide BiBTeX

@article{dupas:halshs-01228397,
  TITLE = {{Identifying seasonal patterns of phosphorus storm dynamics with dynamic time warping}},
  AUTHOR = {Dupas, R{\'e}mi and Tavenard, Romain and Fovet, Oph{\'e}lie and Gilliet, Nicolas and Grimaldi, Catherine and Gascuel, Chantal},
  URL = {https://shs.hal.science/halshs-01228397},
  JOURNAL = {{Water Resources Research}},
  PUBLISHER = {{American Geophysical Union}},
  VOLUME = {51},
  NUMBER = {11},
  PAGES = {8868--8882},
  YEAR = {2015},
  DOI = {10.1002/2015WR017338},
  KEYWORDS = {dynamic time warping ; watershed ; storm events ; phosphorus ; time series analysis},
  PDF = {https://shs.hal.science/halshs-01228397v1/file/article_WRR_accepte_avec_fig.pdf},
  HAL_ID = {halshs-01228397},
  HAL_VERSION = {v1},
}

A. Aubert - R. Tavenard - R. Emonet - A. de Lavenne - S. Malinowski - T. Guyet - R. Quiniou - J. M. Odobez - P. Mérot - C. Gascuel. Clustering Flood Events from Water Quality Time-Series using Latent Dirichlet Allocation Model. Water Resources Research 2013.
Show/Hide BiBTeX

@article{aubert:halshs-00906292,
  TITLE = {{Clustering Flood Events from Water Quality Time-Series using Latent Dirichlet Allocation Model}},
  AUTHOR = {Aubert, Alice and Tavenard, Romain and Emonet, R{\'e}mi and de Lavenne, Alban and Malinowski, Simon and Guyet, Thomas and Quiniou, Ren{\'e} and Odobez, Jean-Marc and M{\'e}rot, Philippe and Gascuel, Chantal},
  URL = {https://shs.hal.science/halshs-00906292},
  JOURNAL = {{Water Resources Research}},
  PUBLISHER = {{American Geophysical Union}},
  VOLUME = {49},
  NUMBER = {12},
  PAGES = {8187--8199},
  YEAR = {2013},
  DOI = {10.1002/2013WR014086},
  KEYWORDS = {temporal pattern extraction ; time series analysis ; probabilistic model ; flood events ; Water quality},
  PDF = {https://shs.hal.science/halshs-00906292v1/file/aubert_wrr_final.pdf},
  HAL_ID = {halshs-00906292},
  HAL_VERSION = {v1},
}

Machine Learning & Time Series

This section gathers Machine Learning tools dedicated to time series with no specific focus on environmental data.

Related papers

R. Tavenard - S. Malinowski. Cost-Aware Early Classification of Time Series. European Conference on Machine Learning and Principles and Practice of Knowledge Discovery 2016. Supplementary material
Show/Hide BiBTeX

@inproceedings{tavenard:halshs-01339007,
  TITLE = {{Cost-Aware Early Classification of Time Series}},
  AUTHOR = {Tavenard, Romain and Malinowski, Simon},
  URL = {https://shs.hal.science/halshs-01339007},
  BOOKTITLE = {{European Conference on Machine Learning and Principles and Practice of Knowledge Discovery}},
  ADDRESS = {Riva del Garda, Italy},
  PAGES = {632-647},
  YEAR = {2016},
  MONTH = Sep,
  DOI = {10.1007/978-3-319-46128-1\_40},
  KEYWORDS = {time series classification  ;  early classification},
  PDF = {https://shs.hal.science/halshs-01339007v1/file/20160628_final-ecml_early_classification.pdf},
  HAL_ID = {halshs-01339007},
  HAL_VERSION = {v1},
}

A. Le Guennec - S. Malinowski - R. Tavenard. Data Augmentation for Time Series Classification using Convolutional Neural Networks. ECML/PKDD Workshop on Advanced Analytics and Learning on Temporal Data 2016.
Show/Hide BiBTeX

@inproceedings{leguennec:halshs-01357973,
  TITLE = {{Data Augmentation for Time Series Classification using Convolutional Neural Networks}},
  AUTHOR = {Le Guennec, Arthur and Malinowski, Simon and Tavenard, Romain},
  URL = {https://shs.hal.science/halshs-01357973},
  BOOKTITLE = {{ECML/PKDD Workshop on Advanced Analytics and Learning on Temporal Data}},
  ADDRESS = {Riva Del Garda, Italy},
  YEAR = {2016},
  MONTH = Sep,
  KEYWORDS = {time series ; convolutional neural networks},
  PDF = {https://shs.hal.science/halshs-01357973v1/file/AALTD16_paper_9.pdf},
  HAL_ID = {halshs-01357973},
  HAL_VERSION = {v1},
}

A. Bailly - S. Malinowski - R. Tavenard - L. Chapel - T. Guyet. Dense Bag-of-Temporal-SIFT-Words for Time Series Classification. Advanced Analysis and Learning on Temporal Data 2016.
Show/Hide BiBTeX

@incollection{bailly:hal-01252726,
  TITLE = {{Dense Bag-of-Temporal-SIFT-Words for Time Series Classification}},
  AUTHOR = {Bailly, Adeline and Malinowski, Simon and Tavenard, Romain and Chapel, Laetitia and Guyet, Thomas},
  URL = {https://hal.science/hal-01252726},
  BOOKTITLE = {{Advanced Analysis and Learning on Temporal Data}},
  PUBLISHER = {{Springer}},
  YEAR = {2016},
  DOI = {10.1007/978-3-319-44412-3\_2},
  KEYWORDS = {time series classification ; SIFT ; Bag-of-Words ; BoTSW},
  PDF = {https://hal.science/hal-01252726v4/file/dense-bag-temporal.pdf},
  HAL_ID = {hal-01252726},
  HAL_VERSION = {v4},
}

A. Bailly - S. Malinowski - R. Tavenard - T. Guyet - L. Chapel. Bag-of-Temporal-SIFT-Words for Time Series Classification. ECML/PKDD Workshop on Advanced Analytics and Learning on Temporal Data 2015.
Show/Hide BiBTeX

@inproceedings{bailly:halshs-01184900,
  TITLE = {{Bag-of-Temporal-SIFT-Words for Time Series Classification}},
  AUTHOR = {Bailly, Adeline and Malinowski, Simon and Tavenard, Romain and Guyet, Thomas and Chapel, Laetitia},
  URL = {https://shs.hal.science/halshs-01184900},
  BOOKTITLE = {{ECML/PKDD Workshop on Advanced Analytics and Learning on Temporal Data}},
  ADDRESS = {Porto, Portugal},
  YEAR = {2015},
  MONTH = Sep,
  KEYWORDS = {SIFT ; BoTSW ; time series classification ; Bag-of-Words},
  PDF = {https://shs.hal.science/halshs-01184900v1/file/bag-temporal-sift.pdf},
  HAL_ID = {halshs-01184900},
  HAL_VERSION = {v1},
}

S. Malinowski - T. Guyet - R. Quiniou - R. Tavenard. 1d-SAX: A Novel Symbolic Representation for Time Series. Advances in Intelligent Data Analysis 2013.
Show/Hide BiBTeX

@inproceedings{malinowski:halshs-00912512,
  TITLE = {{1d-SAX: A Novel Symbolic Representation for Time Series}},
  AUTHOR = {Malinowski, Simon and Guyet, Thomas and Quiniou, Ren{\'e} and Tavenard, Romain},
  URL = {https://shs.hal.science/halshs-00912512},
  BOOKTITLE = {{International Symposium on Intelligent Data Analysis}},
  ADDRESS = {United Kingdom},
  NUMBER = {12},
  PAGES = {273-284},
  YEAR = {2013},
  DOI = {10.1007/978-3-642-41398-8\_24},
  KEYWORDS = {SAX ; time series retrieval},
  PDF = {https://shs.hal.science/halshs-00912512v1/file/paper_IDA.pdf},
  HAL_ID = {halshs-00912512},
  HAL_VERSION = {v1},
}

Related source code

Time-Sensitive Graphical Models

We have been using time-sensitive topic models (such as Probabilistic Latent Semantic Motifs or Hierarchical Dirichlet Latent Semantic Motifs) to perform action recognition in videos. We are still investigating the design of richer models to better capture information from streams of numerical features.

Related papers

R. Tavenard - R. Emonet - J. M. Odobez. Time-sensitive Topic Models For Action Recognition In Videos. ICIP - International Conference on Image Processing 2013.
Show/Hide BiBTeX

@inproceedings{tavenard:hal-00872048,
  TITLE = {{Time-Sensitive Topic Models for Action Recognition in Videos}},
  AUTHOR = {Tavenard, Romain and Emonet, R{\'e}mi and Odobez, Jean-Marc},
  URL = {https://inria.hal.science/hal-00872048},
  BOOKTITLE = {{ICIP - International Conference on Image Processing}},
  ADDRESS = {Melbourne, Australia},
  YEAR = {2013},
  MONTH = Sep,
  PDF = {https://inria.hal.science/hal-00872048v1/file/hal-00872048.pdf},
  HAL_ID = {hal-00872048},
  HAL_VERSION = {v1},
}

@article{aubert:halshs-00906292,
  TITLE = {{Clustering Flood Events from Water Quality Time-Series using Latent Dirichlet Allocation Model}},
  AUTHOR = {Aubert, Alice and Tavenard, Romain and Emonet, R{\'e}mi and de Lavenne, Alban and Malinowski, Simon and Guyet, Thomas and Quiniou, Ren{\'e} and Odobez, Jean-Marc and M{\'e}rot, Philippe and Gascuel, Chantal},
  URL = {https://shs.hal.science/halshs-00906292},
  JOURNAL = {{Water Resources Research}},
  PUBLISHER = {{American Geophysical Union}},
  VOLUME = {49},
  NUMBER = {12},
  PAGES = {8187--8199},
  YEAR = {2013},
  DOI = {10.1002/2013WR014086},
  KEYWORDS = {temporal pattern extraction ; time series analysis ; probabilistic model ; flood events ; Water quality},
  PDF = {https://shs.hal.science/halshs-00906292v1/file/aubert_wrr_final.pdf},
  HAL_ID = {halshs-00906292},
  HAL_VERSION = {v1},
}

Indexing & IR [Past]

Our main goal in this project was to introduce new indexing schemes that were able to efficiently deal with time series. One contribution in this field was iSAX+, an approximate-lower-bound-based indexing scheme for DTW. Some works about vector data indexing are also cited here.

Related papers

R. Tavenard - L. Amsaleg. Improving the Efficiency of Traditional DTW Accelerators. Knowledge and Information Systems (KAIS) 2015.
Show/Hide BiBTeX

@article{tavenard:hal-00862176,
  TITLE = {{Improving the Efficiency of Traditional DTW Accelerators}},
  AUTHOR = {Tavenard, Romain and Amsaleg, Laurent},
  URL = {https://inria.hal.science/hal-00862176},
  JOURNAL = {{Knowledge and Information Systems (KAIS)}},
  PUBLISHER = {{Springer}},
  VOLUME = {42},
  NUMBER = {1},
  PAGES = {215--243},
  YEAR = {2015},
  MONTH = Jan,
  DOI = {10.1007/s10115-013-0698-7},
  KEYWORDS = {Dynamic Time Warping ; Indexing Trees ; Upper Bounds ; Indexing ; Lower Bounds},
  PDF = {https://inria.hal.science/hal-00862176v1/file/hal-00862176.pdf},
  HAL_ID = {hal-00862176},
  HAL_VERSION = {v1},
}

R. Tavenard - L. Amsaleg - G. Gravier. Model-based similarity estimation of multidimensional temporal sequences. Annals of Telecommunications - annales des télécommunications 2009.
Show/Hide BiBTeX

@article{tavenard:inria-00567877,
  TITLE = {{Model-based similarity estimation of multidimensional temporal sequences}},
  AUTHOR = {Tavenard, Romain and Amsaleg, Laurent and Gravier, Guillaume},
  URL = {https://inria.hal.science/inria-00567877},
  JOURNAL = {{Annals of Telecommunications - annales des t{\'e}l{\'e}communications}},
  PUBLISHER = {{Springer}},
  VOLUME = {64},
  NUMBER = {5},
  PAGES = {381-390},
  YEAR = {2009},
  MONTH = Jun,
  DOI = {10.1007/s12243-009-0091-4},
  KEYWORDS = {Multidimensional feature sequences ; Support vector regression ; Temporal aspects ; Similarity estimation in a model space},
  HAL_ID = {inria-00567877},
  HAL_VERSION = {v1},
}

R. Tavenard - H. Jégou - M. Lagrange. Efficient Cover Song Identification using approximate nearest neighbors. Research Report, 2013.
Show/Hide BiBTeX

@unpublished{tavenard:hal-00672897,
  TITLE = {{Efficient Cover Song Identification using approximate nearest neighbors}},
  AUTHOR = {Tavenard, Romain and J{\'e}gou, Herv{\'e} and Lagrange, Mathieu},
  URL = {https://inria.hal.science/hal-00672897},
  NOTE = {working paper or preprint},
  YEAR = {2013},
  MONTH = Feb,
  PDF = {https://inria.hal.science/hal-00672897v1/file/hal2012.pdf},
  HAL_ID = {hal-00672897},
  HAL_VERSION = {v1},
}

R. Tavenard. Indexing feature sequences. Thesis, 2011.
Show/Hide BiBTeX

@phdthesis{tavenard:tel-00639225,
  TITLE = {{Indexation de s{\'e}quences de descripteurs}},
  AUTHOR = {Tavenard, Romain},
  URL = {https://theses.hal.science/tel-00639225},
  SCHOOL = {{Universit{\'e} Rennes 1}},
  YEAR = {2011},
  MONTH = Jul,
  KEYWORDS = {s{\'e}quences ; indexation},
  TYPE = {Theses},
  PDF = {https://theses.hal.science/tel-00639225v1/file/thesis.pdf},
  HAL_ID = {tel-00639225},
  HAL_VERSION = {v1},
}

H. Jégou - R. Tavenard - M. Douze - L. Amsaleg. Searching in one billion vectors: re-rank with source coding. ICASSP 2011 - International Conference on Acoustics, Speech and Signal Processing 2011.
Show/Hide BiBTeX

@inproceedings{jegou:inria-00566883,
  TITLE = {{Searching in one billion vectors: re-rank with source coding}},
  AUTHOR = {J{\'e}gou, Herv{\'e} and Tavenard, Romain and Douze, Matthijs and Amsaleg, Laurent},
  URL = {https://inria.hal.science/inria-00566883},
  BOOKTITLE = {{ICASSP 2011 - International Conference on Acoustics, Speech and Signal Processing}},
  ADDRESS = {Prague, Czech Republic},
  PUBLISHER = {{IEEE}},
  PAGES = {861-864},
  YEAR = {2011},
  MONTH = May,
  DOI = {10.1109/ICASSP.2011.5946540},
  KEYWORDS = {approximate nearest neighbors ; large scale indexing ; source coding ; similarity search},
  PDF = {https://inria.hal.science/inria-00566883v1/file/paper.pdf},
  HAL_ID = {inria-00566883},
  HAL_VERSION = {v1},
}

R. Tavenard - H. Jégou - L. Amsaleg. Balancing clusters to reduce response time variability in large scale image search. International Workshop on Content-Based Multimedia Indexing (CBMI 2011) 2011.
Show/Hide BiBTeX

@inproceedings{tavenard:inria-00576886,
  TITLE = {{Balancing clusters to reduce response time variability in large scale image search}},
  AUTHOR = {Tavenard, Romain and J{\'e}gou, Herv{\'e} and Amsaleg, Laurent},
  URL = {https://inria.hal.science/inria-00576886},
  BOOKTITLE = {{International Workshop on Content-Based Multimedia Indexing (CBMI 2011)}},
  ADDRESS = {Madrid, Spain},
  YEAR = {2011},
  MONTH = Jun,
  PDF = {https://inria.hal.science/inria-00576886v2/file/cbmi2011_preprint.pdf},
  HAL_ID = {inria-00576886},
  HAL_VERSION = {v2},
}

V. Claveau - R. Tavenard - L. Amsaleg. Vectorisation des processus d'appariement document-requête. 7e conférence en recherche d'informations et applications, CORIA'10 2010.
Show/Hide BiBTeX

@inproceedings{claveau:inria-00561797,
  TITLE = {{Vectorisation des processus d'appariement document-requ{\^e}te}},
  AUTHOR = {Claveau, Vincent and Tavenard, Romain and Amsaleg, Laurent},
  URL = {https://inria.hal.science/inria-00561797},
  BOOKTITLE = {{7e conf{\'e}rence en recherche d'informations et applications, CORIA'10}},
  ADDRESS = {Sousse, Tunisia},
  YEAR = {2010},
  MONTH = Mar,
  KEYWORDS = {Vectorization ; vector space ; language modeling ; pairing complexity},
  HAL_ID = {inria-00561797},
  HAL_VERSION = {v1},
}

Time Series Mining for Smart Environments [Past]

The growing use of lots of low-level sensors instead of few higher-level ones implies the use of dedicated pattern extraction methods. To do so, we have worked on the already existing T-patterns algorithm so that it can efficiently scale up to larger volumes of data.

Related papers

A. Ali Salah - E. Pauwels - R. Tavenard - T. Gevers. T-Patterns Revisited: Mining for Temporal Patterns in Sensor Data. Sensors 2010.
Show/Hide BiBTeX

@article{salah:halshs-01138500,
  TITLE = {{T-Patterns Revisited: Mining for Temporal Patterns in Sensor Data}},
  AUTHOR = {Salah, Albert Ali and Pauwels, Eric and Tavenard, Romain and Gevers, Theo},
  URL = {https://shs.hal.science/halshs-01138500},
  JOURNAL = {{Sensors}},
  PUBLISHER = {{MDPI}},
  VOLUME = {10},
  NUMBER = {8},
  PAGES = {7496-7513},
  YEAR = {2010},
  DOI = {10.3390/s100807496},
  KEYWORDS = {MERL motion data ; Gaussian mixture model ; Lempel-Ziv ; T-patterns ; temporal pattern extraction ; sensor networks},
  PDF = {https://shs.hal.science/halshs-01138500v1/file/sensors-10-07496.pdf},
  HAL_ID = {halshs-01138500},
  HAL_VERSION = {v1},
}

E. Pauwels - A. Ali Salah - R. Tavenard. Sensor Networks for Ambient Intelligence. IEEE Workshop on Multimedia Signal Processing 2007.
Show/Hide BiBTeX

@inproceedings{pauwels:halshs-01138508,
  TITLE = {{Sensor Networks for Ambient Intelligence}},
  AUTHOR = {Pauwels, Eric and Salah, Albert Ali and Tavenard, Romain},
  URL = {https://shs.hal.science/halshs-01138508},
  BOOKTITLE = {{IEEE Workshop on Multimedia Signal Processing}},
  ADDRESS = {Chania, Greece},
  YEAR = {2007},
  MONTH = Oct,
  DOI = {10.1109/MMSP.2007.4412806},
  KEYWORDS = {Sensors ; Ambient Intelligence ; Data mining},
  PDF = {https://shs.hal.science/halshs-01138508v1/file/mmsp_07_pauwels.pdf},
  HAL_ID = {halshs-01138508},
  HAL_VERSION = {v1},
}

R. Tavenard - A. Ali Salah - E. Pauwels. Searching for Temporal Patterns in AmI Sensor Data. Constructing Ambient Intelligence 2007.
Show/Hide BiBTeX

@inproceedings{tavenard:halshs-01138512,
  TITLE = {{Searching for Temporal Patterns in AmI Sensor Data}},
  AUTHOR = {Tavenard, Romain and Salah, Albert Ali and Pauwels, Eric},
  URL = {https://shs.hal.science/halshs-01138512},
  BOOKTITLE = {{Constructing Ambient Intelligence}},
  ADDRESS = {Darmstadt, Germany},
  PAGES = {53-62},
  YEAR = {2007},
  MONTH = Nov,
  DOI = {10.1007/978-3-540-85379-4\_7},
  KEYWORDS = {Data mining ; Ambient Intelligence},
  PDF = {https://shs.hal.science/halshs-01138512v1/file/ami_07_tavenard.pdf},
  HAL_ID = {halshs-01138512},
  HAL_VERSION = {v1},
}