Information processing apparatus, information processing method, and recording medium

Patent metadata

Field	Value
Publication number	US-2019155267-A1
Application number	US-201815917030-A
Country	US
Kind code	A1
Filing date	Mar 9, 2018
Priority date	Nov 17, 2017
Publication date	May 23, 2019
Grant date	—

Abstract

An information processing apparatus has an input data acquisition unit that acquires a plurality of input data, a fault tolerance requirement acquisition unit that acquires a fault tolerance requirement for the plurality of input data, a training data definition unit that defines an output which satisfies the fault tolerance requirement, a fault pattern generation unit that generates a plurality of fault patterns which include at least one of the plurality of input data based on the plurality of input data and the fault tolerance requirement, a model update unit that updates the information processing model so as to tolerate a fault of the input data satisfying the fault tolerance requirement, and an execution control unit that applies the plurality of input data to the information processing model updated by the model update unit and executes the information processing model.

First claim

1 . An information processing apparatus comprising:
an input data acquisition unit that acquires a plurality of input data; a fault tolerance requirement acquisition unit that acquires a fault tolerance requirement for the plurality of input data; a training data definition unit that defines an output which satisfies the fault tolerance requirement when at least one input data among the plurality of input data is input to a predetermined information processing model, as training data; a fault pattern generation unit that generates a plurality of fault patterns which include at least one of the plurality of input data based on the plurality of input data and the fault tolerance requirement; a model update unit that updates the information processing model so as to tolerate a fault of the input data satisfying the fault tolerance requirement based on the plurality of fault patterns and the training data; and an execution control unit that applies the plurality of input data to the information processing model updated by the model update unit and executes the information processing model.

2 . The information processing apparatus according to claim 1 , wherein
the input data acquisition unit acquires a plurality of detection signals detected by a plurality of sensors as the plurality of input data, the fault tolerance requirement acquisition unit acquires an upper limit value of a number of multiple faults which indicates a number of faulty sensors among the plurality of sensors, as the fault tolerance requirement, the fault pattern generation unit generates the fault pattern indicating a combination of detection signals of the faulty sensors within a range of the number of faults equal to or less than the upper limit value of the number of the multiple faults, and the model update unit updates the information processing model so as to perform information processing that satisfies the fault tolerance requirement by using input data which is not included in the fault pattern.

3 . The information processing apparatus according to claim 2 , wherein
the fault pattern generation unit generates the plurality of fault patterns by searching combinations of the plurality of input data in descending order of the number of the combinations within the range of the number of faults equal to or less than the upper limit value of the number of the multiple faults.

4 . The information processing apparatus according to claim 2 , wherein
the fault pattern generation unit generates a fault pattern corresponding to the new information processing model, which uses a detection signal of a sensor other than the faulty sensor as an input signal within the range of the number of faults equal to or less than the upper limit value of the number of the multiple faults.

5 . The information processing apparatus according to claim 1 , wherein
the input data acquisition unit acquires a plurality of detection signals detected by a plurality of sensors as the plurality of input data, the fault tolerance requirement acquisition unit acquires a fault probability of an arbitrary combination of the plurality of sensors as the fault tolerance requirement, and the fault pattern generation unit generates the plurality of fault patterns with the fault probability equal to or lower than a predetermined upper limit value.

6 . The information processing apparatus according to claim 2 , wherein
the model update unit updates a plurality of the information processing models corresponding to the plurality of fault patterns, and the execution control unit comprises: a monitoring unit that monitors whether at least one of the plurality of sensors is faulty; a fault exclusion input data acquisition unit that acquires a detection signal detected by a sensor other than a faulty sensor as input data; and a combination output unit that combines output data obtained by executing the plurality of information processing models using the input data acquired by the fault exclusion input data acquisition unit and outputs the combined output data.

7 . The information processing apparatus according to claim 1 , wherein
the plurality of sensors output the plurality of detection signals relating to a demand prediction of a predetermined target object or target service.

8 . An information processing method comprising:
acquiring a plurality of input data; acquiring a fault tolerance requirement for the plurality of input data; defining an output which satisfies the fault tolerance requirement when at least one input data among the plurality of input data is input to a predetermined information processing model, as training data; generating a plurality of fault patterns which include at least one of the plurality of input data based on the plurality of input data and the fault tolerance requirement; updating the information processing model so as to tolerate a fault of the input data satisfying the fault tolerance requirement based on the plurality of fault patterns and the training data; and applying the plurality of input data to the information processing model updated by the model update unit and executing the information processing model.

9 . The information processing method according to claim 8 , wherein
the acquiring the input data acquires a plurality of detection signals detected by a plurality of sensors as the plurality of input data, the acquiring the fault tolerance requirement acquires an upper limit value of a number of multiple faults which indicates a number of faulty sensors among the plurality of sensors, as the fault tolerance requirement, the generating the plurality of fault patterns generates the fault pattern indicating a combination of detection signals of the faulty sensors within a range of the number of faults equal to or less than the upper limit value of the number of the multiple faults, and the updating the information updates the information processing model so as to perform information processing that satisfies the fault tolerance requirement by using input data which is not included in the fault pattern.

10 . The information processing method according to claim 9 , wherein
the generating the plurality of fault patterns generates the plurality of fault patterns by searching combinations of the plurality of input data in descending order of the number of the combinations within the range of the number of faults equal to or less than the upper limit value of the number of the multiple faults.

11 . The information processing method according to claim 9 , wherein
the generating the plurality of fault patterns generates a fault pattern corresponding to the new information processing model, which uses a detection signal of a sensor other than the faulty sensor as an input signal within the range of the number of faults equal to or less than the upper limit value of the number of the multiple faults.

12 . The information processing method according to claim 8 , wherein
the acquiring the plurality of input data acquires a plurality of detection signals detected by a plurality of sensors as the plurality of input data, the acquiring the fault tolerance requirement acquires a fault probability of an arbitrary combination of the plurality of sensors as the fault tolerance requirement, and the generating the plurality of fault patterns generates the plurality of fault patterns with the fault probability equal to or lower than a predetermined upper limit value.

13 . The information processing method according to claim 9 , wherein
the updating the information processing model updates a plurality of the information processing models corresponding

Assignees

• Toshiba Kk

Inventors

• Shin Hiromasa
• Kakimoto Mitsuru
• Shiga Yoshiaki

Classifications

• G05B23/0221 Primary

  Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods · CPC title

• G06F15/18

  Physics · mapped topic

• G05B23/0283

  Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] (preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06) · CPC title

• G06F11/0721

  within a central processing unit [CPU] · CPC title

• G06N20/00

  Machine learning · CPC title