Methods and systems for improving controlling of a robot

A method for training a machine learning (ML) algorithm, the method comprising:

• (i) assigning a level of realism to a simulation environment;
• (ii) receiving input information associated with the simulation environment to which the level of realism is assigned;
• (ii) associating the input information with a realism value representing the level of realism;
• (iii) modifying one or more parameters of the ML algorithm based on the input information and the realism value;
• (iv) controlling, based on the modified one or more parameters of the ML algorithm, a robot;
• (v) evaluating a performance of the robot;
• and
• (vi) determining, based on a result of the evaluation of the performance of the robot, whether to repeat the steps (i)-(v).

An apparatus for controlling a robot, the apparatus comprising:

• a memory;
• and
• processing circuitry coupled to the memory, wherein the apparatus is configured to:
• assign a first level of realism to a first simulation environment;
• assign a second level of realism to a second simulation environment;
• obtain first input information associated with the first simulation environment to which the first level of realism is assigned;
• obtain second input information associated with the second simulation environment to which the second level of realism is assigned, wherein the first level of realism is different from the second level of realism;
• associate the first input information with a first realism value representing the first level of realism;
• associate the second input information with a second realism value representing the second level of realism;
• modify one or more parameters of a machine learning (ML) process used for controlling the robot based on:
• (i) the first input information;
• (ii) the second input information;
• (iii) the first realism value representing the first level of realism;
• and
• (iv) the second realism value representing the second level of realism;
• and
• control the robot based on the modified one or more parameters of the ML process.

An apparatus for training a machine learning (ML) algorithm, the apparatus comprising:

• a memory;
• and
• processing circuitry coupled to the memory, wherein the apparatus is configured to:
• (i) assign a level of realism to a simulation environment;
• (ii) receive input information associated with the simulation environment to which the level of realism is assigned;
• (ii) associate the input information with a realism value representing the level of realism;
• (iii) modify one or more parameters of the ML algorithm based on the input information and the realism value;
• (iv) control, based on the modified one or more parameters of the ML algorithm, a robot;
• (v) evaluate a performance of the robot;
• and
• (vi) determine, based on a result of the evaluation of the performance of the robot, whether to repeat the steps (i)-(v).