Abstract: Accuracy of unsupervised anomalous sound detection is improved using a small number of pieces of anomalous sound data. A threshold deciding part (13) calculates an anomaly score for each of a plurality of pieces of anomalous sound data, using a normal model learned with normal sound data and an anomaly model expressing the pieces of anomalous sound data, and decides a minimum value among the anomaly scores as a threshold. A weight updating part (14) updates, using a plurality of pieces of normal sound data, the pieces of anomalous sound data and the threshold, weights of the anomaly model so that all the pieces of anomalous sound data are judged as anomalous, and probability of the pieces of normal sound data being judged as anomalous is minimized.

Abstract: A possible region of encoding results of anomalous samples is limited. An encoder storage unit 14 stores an encoder for projecting an input feature value into a latent space in which the latent space is a closed manifold, a normal distribution obtained by learning normal data and an anomalous distribution obtained by learning anomalous data are held on the manifold, and a decoder for reconstructing the output of the encoder. An encoding unit 15 obtains a reconstruction result output by the decoder when a feature value of target data is input to the encoder. An anomaly score calculation unit 16 calculates an anomaly score of the target data based on distances between the reconstruction result and the normal distribution and distances between the reconstruction result and the anomalous distribution.

Abstract: The present disclosure relates to a method of machine learning regardless of the number of dimensions of the samples. The method provides model learning of a variational auto-encoder that uses AUC optimization criteria. The method includes learning parameters ?{circumflex over (?)} and ?{circumflex over (?)} of the a variational auto-encoder. The variational auto-encoder includes an encoder for constructing a latent variable from an observed variable and a decoder for reconstructing the observed variable. The method uses learning data set defined using based on normal data generated from sounds observed during normal operation and abnormal data generated from sounds observed during abnormal operation. The AUC value is based in part on a reconstruction probability. Incorporating aspects of the reconstruction error into the AUC value prevents the variational auto-encoder from divergence of the abnormality degree regarding the abnormal data.

Abstract: There is provided a model learning technique for learning a model which performs classification into three values by model learning using an AUC optimization criterion. A model learning unit is included which learns a parameter ?{circumflex over (?)} of a model by using a learning data set based on a criterion which uses a predetermined AUC value, the learning data set being defined using normal data generated from sound observed in a normal state and abnormal data generated from sound observed in an abnormal state, and the AUC value is defined from a difference between an abnormality degree of the normal data and an abnormality degree of the abnormal data using a two-stage step function T(x).

Abstract: Accuracy of unsupervised anomalous sound detection is improved using a small number of pieces of anomalous sound data. A threshold deciding part (13) calculates an anomaly score for each of a plurality of pieces of anomalous sound data, using a normal model learned with normal sound data and an anomaly model expressing the pieces of anomalous sound data, and decides a minimum value among the anomaly scores as a threshold. A weight updating part (14) updates, using a plurality of pieces of normal sound data, the pieces of anomalous sound data and the threshold, weights of the anomaly model so that all the pieces of anomalous sound data are judged as anomalous, and probability of the pieces of normal sound data being judged as anomalous is minimized.