Information processing apparatus and method, robot controlling apparatus and method, and non-transitory computer-readable storage medium

1. An information processing apparatus comprising:

a first acquiring unit configured to acquire, from an external apparatus which can communicate with the information processing apparatus via a network, a first learning model which outputs recognition information in response to input of image information;

a learning unit configured to cause the first learning model to learn a result of control by the information processing apparatus using recognition information that is output from a second learning model in response to input of the image information in an execution environment;

an evaluation unit configured to evaluate the first learning model based on image information in the execution environment and recognition information that is output as ground truth in response to input of the image information to a learning model;

a replacing unit configured to replace the second learning model with the first learning model that has been learned by the learning unit, in a case where the result of evaluation by the evaluation unit meets the a requirement; and

an output unit configured to cause the first learning model learned by the learning unit to output recognition information by inputting image information in the execution environment to the first learning model.

2. The information processing apparatus according to claim 1 , wherein the learning unit updates a coupled weighting coefficient for the first learning model based on the evaluation result of the first learning model obtained by the evaluation unit and recognition information as ground truth associated with the image information.

3. The information processing apparatus according to claim 1 , wherein the recognition information as ground truth associated with the image information is learning data obtained by performing a task based on the second learning model, which receives image information as an input and outputs recognition information, and

the learning unit updates the first learning model based on the learning data in a case where performance of the first learning model evaluated by the evaluation unit does not meet a predetermined performance level.

4. The information processing apparatus according to claim 1 , wherein the replacing unit replaces the second learning model with the first learning model in a case where performance of the first learning model evaluated by the evaluation unit meets performance that is set as performance required to perform a task.

5. The information processing apparatus according to claim 1 , wherein the replacing unit replaces the second learning model with the first learning model in a case where performance of the first learning model evaluated by the evaluation unit meets performance of the second learning model.

6. The information processing apparatus according to claim 1 , wherein each of the first learning model and the second learning model has a different neural network structure to each other.

7. The information processing apparatus according to claim 1 , wherein the recognition information that is output as ground truth in response to input of image information in the execution environment to a learning model is obtained by adding, to image information in the execution environment, a result of controlling the information processing apparatus based on recognition information output from the second model in response to image information in the execution environment.

8. An information processing method that is performed by an information processing apparatus, the method comprising:

(a) acquiring, from an external apparatus which can communicate with the information processing apparatus via a network, a first learning model which outputs recognition information in response to input of image information;

(b) causing the first model to learn a result of control by the information processing apparatus using recognition information that is output from a second learning model in response to input of the image information in an execution environment;

(c) evaluating the first learning model based on image information in the execution environment and recognition information that is output as ground truth in response to input of the image information to a learning model;

(d) replacing the second learning model with the first learning model that has been learned by the learning unit, in a case where the result of evaluation by an evaluation unit meets a requirement; and

e) causing the first learning model learned in the learning (b) to output recognition information by inputting image information in the execution environment to the first learning model.

9. A non-transitory computer-readable storage medium storing a program which, when read and executed by a computer, cause the computer to executes the step of an information processing method, the method comprising:

(a) acquiring, from an external apparatus which can communicate with the information processing apparatus via a network, a first learning model which outputs recognition information in response to input of image information;

(b) causing the first model to learn a result of control by the information processing apparatus using recognition information that is output from a second learning model in response to input of the image information in an execution environment; and

(c) evaluating the first learning model based on image information in the execution environment and recognition information that is output as ground truth in response to input of the image information to a learning model;

(d) replacing the second learning model with the first learning model that has been learned by the learning unit, in a case where the result of evaluation by an evaluation unit meets a requirement; and

(e) causing the first learning model learned in the learning (b) to output recognition information by inputting image information in the execution environment to the first learning model.