@incollection{fdi:010085544,
  title = {{S}tochastic pairing for contrastive anomaly detection on time series},
  author = {{C}hambaret, {G}. and {B}erti-{E}quille, {L}aure and {B}ouchara, {F}. and {B}runo, {E}. and {M}artin, {V}. and {C}haillan, {F}.},
  editor = {},
  language = {{ENG}},
  abstract = {{A}nomaly detection for predictive maintenance is a significant concern for industry. {U}nanticipated failures cause high costs for experts involved in maintenance policy. {T}raditional reconstruction-based anomaly detection methods perform well on multivariate time series but they do not consider the diversity of samples in the training dataset. {A}n abrupt change of operating conditions, which is labeled as anomaly by experts, is often not detected due to the lack of sample diversity. {B}esides, obtaining large volumes of labeled training data is cumbersome and sometimes impossible in practice, whereas large amounts of unlabelled data are available and could be used by unsupervised learning techniques. {I}n this paper, we apply the principles of contrastive learning and augmentation in a self supervised way to improve feature representation of multivariate time series. {W}e model a large variety of operating conditions with an innovative distance based stochastic method to prepare an anomaly detection downstream task. {O}ur approach is tested on {NASA} {SMAP}/{MSL} public dataset and shows good performance close to the state-of-the-art anomaly detection methods.},
  keywords = {},
  booktitle = {{P}attern recognition and artificial intelligence : proceedings, part {II}},
  numero = {13364},
  pages = {306--317},
  address = {{C}ham},
  publisher = {{S}pringer},
  series = {{L}ecture {N}otes in {C}omputer {S}cience},
  year = {2022},
  DOI = {10.1007/978-3-031-09282-4_26},
  ISBN = {978-3-031-09281-7},
  URL = {https://www.documentation.ird.fr/hor/fdi:010085544},
}