Update of a machine learning system

Patent metadata

Field	Value
Publication number	US-2017193402-A1
Application number	US-201615395244-A
Country	US
Kind code	A1
Filing date	Dec 30, 2016
Priority date	Dec 31, 2015
Publication date	Jul 6, 2017
Grant date	—

Abstract

The invention notably relates to a computer-implemented method for updating a model of a machine learning system. The method comprises providing a first set of observations of similar events, each observation being associated with one or more variables, each variable being associated with a value, and with a target value; indexing each observation of the first set with its corresponding one or more variables and target value; receiving, on the index, a query allowing a selection of a subset of the first set of observations; returning, as a result of the query, a subset of the first set of observations; providing a second model; training the provided second model with the returned subset of the first set of observations; and loading the trained second model.

First claim

1 . A computer-implemented method for updating a model of a machine learning system, comprising:
providing a first set of observations of similar events, each observation being associated with a target value and with one or more variables, each variable being associated with a value corresponding to the observation; indexing each observation of the first set with its corresponding one or more variables and target value; receiving, on the index, a query allowing a selection of a subset of the first set of observations; returning, as a result of the query, a subset of the first set of observations; providing a second model; training the provided second model with the returned subset of the first set of observations; and loading the trained second model.

2 . The computer-implemented method of claim 1 , further comprising:
providing a first model; training the first model with the first set of observations; and storing the trained first model.

3 . The computer-implemented method of claim 1 , further comprising, before receiving a query on the index:
collecting a second set of observations of current events, wherein receiving a query on the index further comprises: receiving, on the index, a query allowing a selection of a subset of the first set of observations, the query being performed with one or more variables of the second set of observations.

4 . The computer-implemented method of claim 3 , further comprising before receiving a query on the index:
identifying one or more variables of the second set of observations, wherein receiving a query on the index further comprises: receiving, on the index, a query allowing a selection of a subset of the first set of observations, the query being performed with the identified one or more variables of the second set of observations.

5 . The computer-implemented method of claim 4 , wherein identifying one or more variables of the second set of observations comprises:
identifying slow moving variables and/or identifying fast moving variables.

6 . The computer-implemented method of claim 4 , further comprising:
identifying one or more unknown variables among the variables of the second set of observations; and indexing each observation of the second set associated with the identified one or more unknown variables, the indexation of the said each observation of the second set being performed with the corresponding one or more variables and target value of the said each observation of the second set.

7 . The computer-implemented method of claim 3 , wherein collecting the second set of observations further comprises:
collecting in real time the set of second observations; storing the collected second set of observations; and providing an access to the collected second set of observations before a predetermined period of time elapsed.

8 . The computer-implemented method of claim 1 , further comprising after loading the trained second model:
providing one or more options by setting input variables of a set of input variables with a value; and computing an outcome for the trained second model by using the provided one or more options.

9 . The computer-implemented method of claim 8 , wherein providing one or more options further comprises:
computing combinations of the values of input variables, and wherein computing an outcome for the trained second model further comprises: computing an outcome for the trained second model by using the combinations computed.

10 . The computer-implemented method of claim 9 , further comprising
before receiving a query on the index: collecting a second set of observations of current events, wherein receiving a query on the index further comprises: receiving, on the index, a query allowing a selection of a subset of the first set of observations, the query being performed with one or more variables of the second set of observations; and after computing combinations of the values of input variables: receiving values of variables of the second set of observations, computing combinations of the values of input variables and values of variables of the second set of observations, and computing an outcome for the trained second model by using the combinations of the values of input variables and of variables of the set of second observations.

11 . A computer program comprising instructions for performing the method of claim 1 .

12 . A non-transitory computer readable storage medium having recorded thereon a computer program that when executed by a computer causes the computer to implement a method for updating a model of a machine learning system, the method comprising:
providing a first set of observations of similar events, each observation being associated with a target value and with one or more variables, each variable being associated with a value corresponding to the observation; indexing each observation of the first set with its corresponding one or more variables and target value; receiving, on the index, a query allowing a selection of a subset of the first set of observations; returning, as a result of the query, a subset of the first set of observations; providing a second model; training the provided second model with the returned subset of the first set of observations; and loading the trained second model.

13 . A server comprising:
processing circuitry coupled to a memory, the memory having recorded thereon the computer program for updating a model of a machine learning system, the processing circuitry implanting the computer program by being configured to: provide a first set of observations of similar events, each observation being associated with a target value and with one or more variables, each variable being associated with a value corresponding to the observation; index each observation of the first set with its corresponding one or more variables and target value; receive, on the index, a query allowing a selection of a subset of the first set of observations; return, as a result of the query, a subset of the first set of observations; provide a second model; train the provided second model with the returned subset of the first set of observations; and load the trained second model.

14 . The server of claim 13 , wherein the server is connected to a client computer from which the query on the index is generated.

Assignees

• Dassault Systemes

Inventors

• Grehant Xavier M

Classifications

• G06N99/005 Primary

  Physics · mapped topic

• G06N20/20 Primary

  Ensemble learning · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G06F16/35

  Clustering; Classification · CPC title