Training system for training a generator neural network

A training system for training multiple separate generator neural networks each arranged to transform measured sensor data into generated sensor data, the system comprising:

• a communication interface for accessing a training set of measured sensor data;
• and
• a processor system arranged to train the multiple generator networks together with a discriminator neural network, wherein the generator networks are optimized to transform measured sensor data into generated sensor data, and the discriminator network is optimized to distinguish between the measured sensor data and the generated sensor data;
• wherein each one of the multiple generator networks is associated with a respective transformation goal, and arranged to receive, as input, sensor data and to transform the sensor data it received as input according to the respective transformation goal;
• and
• wherein the discriminator network is arranged to receive, as input, sensor data and a transformation goal selected from the plurality of transformation goals and to determine if the sensor data it received as input is measured sensor data satisfying the transformation goal it received as input.

A generator system for a generator neural network arranged to transform measured sensor data into generated sensor data, the system comprising:

• a communication interface for receiving, as input, measured sensor data and a transformation goal;
• and
• a processor system configured to apply a trained generator network to the received input measured sensor data and the received transformation goal;
• wherein the generator network is trained by a training system, the training system including a second communication interface for accessing a training set of measured sensor data, and a second processor system configured to train the generator network together with a discriminator neural network, wherein the generator network is optimized to transform the measured sensor data into generated sensor data, and the discriminator network is optimized to distinguish between the measured sensor data and the generated sensor data, wherein the generator network is configured to receive, as input, the measured sensor data and a transformation goal selected from a plurality of transformation goals and is configured to transform the measured sensor data it received as input into the generated sensor data according to the transformation goal it received as input, wherein the discriminator network is configured to receive, as input, sensor data and a transformation goal selected from the plurality of transformation goals and to determine if the sensor data it received as input is measured sensor data satisfying the transformation goal.

A generator method for a generator neural network arranged to transform measured sensor data into generated sensor data, the method comprising the following steps:

• receiving, as input, measured sensor data and a transformation goal;
• and
• applying a trained generator network to the received input measured sensor data and the received transformation goal;
• wherein the generator network has been trained by:
• accessing a training set of measured sensor data;
• training the generator network together with a discriminator neural network, including optimizing the generator network to transform the measured sensor data into generated sensor data, and optimizing the discriminator network to distinguish between the measured sensor data and the generated sensor data, wherein the training includes obtaining, as input, the measured sensor data and a transformation goal selected from a plurality of transformation goals and transforming the measured input sensor data into the generated sensor data according to the input transformation goal using the generator network, wherein the discriminator network is configured to receive, as input, sensor data and a transformation goal selected from the plurality of transformation goals and to determine if the sensor data it received as input is measured sensor data satisfying the transformation goal.

A non-transitory computer readable medium on which is stored data representing instructions for a generator neural network arranged to transform measured sensor data into generated sensor data, the data, when executed by a computer, causing the computer to perform the following steps:

• receiving, as input, measured sensor data and a transformation goal;
• and
• applying a trained generator network to the received input measured sensor data and the received transformation goal;
• wherein the generator network has been trained by:
• accessing a training set of measured sensor data;
• and
• training the generator network together with a discriminator neural network, including optimizing the generator network to transform the measured sensor data into generated sensor data, and optimizing the discriminator network to distinguish between the measured sensor data and the generated sensor data, wherein the training includes obtaining, as input, the measured sensor data and a transformation goal selected from a plurality of transformation goals and transforming the measured input sensor data into the generated sensor data according to the input transformation goal using the generator network, wherein the discriminator network is configured to receive, as input, sensor data and a transformation goal selected from the plurality of transformation goals and to determine if the sensor data it received as input is measured sensor data satisfying the transformation goal.

A training system for training a generator neural network arranged to transform measured sensor data into generated sensor data, the system comprising:

• a communication interface for accessing a training set of measured sensor data;
• and
• a processor system configured to train the generator network together with a discriminator neural network, wherein the generator network is optimized to transform the measured sensor data into generated sensor data, and the discriminator network is optimized to distinguish between the measured sensor data and the generated sensor data;
• wherein the generator network is configured to receive, as input, the measured sensor data and a transformation goal selected from a plurality of transformation goals and is configured to transform the measured sensor data it received as input into the generated sensor data according to the transformation goal it received as input, and wherein the plurality of transformation goals include a plurality of time differences, the training data being labeled with a timestamp, the generator network being configured to transform the sensor data it received as input from a first timestamp to a second timestamp according to the time difference, the discriminator network being configured to receive as input a first sensor data, a second sensor data and a time difference and to determine if the first sensor data and the second sensor data satisfy the time difference.