Information processing apparatus and method, and program for teaching an action to a device in a time-series pattern
View Patent ↗An information processing apparatus including a learning unit that learns a predetermined time-series pattern. An output unit outputs a time-series pattern corresponding to the result of learning by the learning unit. An adjusting unit supplied with a time-series pattern obtained from an action by an action unit on the basis of a time-series pattern supplied from the output unit and external teaching for the action adjusts a time-series pattern supplied from the output unit correspondingly to the input time-series pattern. The learning unit learns the time-series pattern supplied from the output unit and adjusted by the adjusting unit.
1. An information processing apparatus comprising:
a learning means for learning a predetermined time-series pattern;
an output means for outputting a time-series pattern learned by the learning means; and
an adjusting means supplied with a time-series pattern obtained from an action made by an action unit on the basis of a previously learned time-series pattern supplied from the output means and external teaching, for the action to adjust a time-series pattern supplied from the output means correspondingly to an input time-series pattern, the learning means learning the time-series pattern supplied from the output means and adjusted by the adjusting means,
wherein the external teaching is given during a period of time when, given one or more new starting or ending positions, the action to move from the given starting position to the ending position made by the action unit on the basis of a previously learned time-series pattern is made, correspondingly to a user's touch on the action unit fails,
wherein the time-series pattern learned by the learning means is previously learned time-series pattern adjusted by the adjusting means based upon external teaching, and
wherein the action unit incorporates an image capture unit that captures a picture for obtaining a time-series pattern, the image capture unit is installed in a predetermined place of the action unit.
2. The apparatus according to claim 1 , being formed from a recurrent type neural network, wherein:
the learning means acquires, as the result of learning the time-series pattern, a weighting factor and parametric bias parameter in the recurrent type neural network;
the output means calculates the parametric bias parameter on the basis of the weighting factor and output a time-series pattern corresponding to the parametric bias parameter resulted from the calculation; and
the adjusting means can be supplied with a time-series pattern obtained from an action made by the action unit on the basis of the time-series pattern supplied from the output means and time-series pattern obtained external teaching for the action to change the parametric bias parameter correspondingly to the input time-series pattern.
3. An information processing method, executed by a processor, comprising the steps of:
learning, utilizing a learning unit, a predetermined time-series pattern;
outputting, utilizing an output unit, a time-series pattern corresponding to the result of learning in the learning step; and
receiving, at a receiving unit, a time-series pattern obtained based on an action made by an action unit on the basis of a previously learned time-series pattern supplied from the outputting step and external teaching, for the action to adjust a time-series pattern supplied from the outputting step correspondingly to an input time-series pattern, in the learning step, there being learned the time-series pattern supplied from the outputting means and adjusted by the adjusting means,
wherein the external teaching is given during a period of time when, given one or more new starting or ending positions the action to move from the given starting position to the ending position made by the action unit on the basis of a previously learned time-series pattern is made, correspondingly to a user's touch on the action unit fails,
wherein the time-series pattern learned in the learning step is previously learned time-series pattern adjusted by the adjusting means based upon external teaching, and
wherein the action unit incorporates an image capture unit that captures a picture for obtaining a time-series pattern, the image capture unit is installed in a predetermined place of the action unit.
4. A non-transitory computer-readable medium encoded with a computer-readable program that when executed on a processor, cause the processor to perform a method, the method comprising the steps of:
learning, utilizing a learning unit, a predetermined time-series pattern;
outputting, utilizing an outputting unit, a time-series pattern corresponding to the result of learning in the learning step; and
receiving, at a receiving unit, a time-series pattern obtained based on an action made by an action unit on the basis of a previously learned time-series pattern supplied from the outputting step and external teaching, for the action to adjust a time-series pattern supplied from the outputting step correspondingly to an input time-series pattern, in the learning step, there being learned the time-series pattern supplied from the outputting step and adjusted by the adjusting means,
wherein the external teaching is given during a period of time when, given one or more new starting or ending positions, the action to move from the given starting position to the ending position made by the action unit on the basis of a previously learned time-series pattern is made, correspondingly to a user's touch on the action unit fails,
wherein the time-series pattern learned in the learning step is previously learned time-series pattern adjusted by the adjusting means based upon external teaching, and
wherein the action unit incorporates an image capture unit that captures a picture for obtaining a time-series pattern, the image capture unit is installed in a predetermined place of the action unit.
5. An information processing apparatus comprising:
a learning means for learning a predetermined time-series pattern;
an output means for outputting a time-series pattern corresponding to the result of learning by the learning means; and
an adjusting means supplied with a time-series pattern obtained from an action made by an action unit on the basis of a previously learned time-series pattern supplied from the output means and external teaching, for the action to adjust a time-series pattern supplied from the output means correspondingly to an input time-series pattern,
wherein the external teaching is given during a period of time when, given one more new starting or ending positions, the action to move from the given starting position to the ending position made by the action unit on the basis of a previously learned time-series pattern is made, correspondingly to a user's touch on the action unit fails,
wherein the time-series pattern learned by the learning means is previously learned time-series pattern adjusted by the adjusting means based upon external teaching, and
wherein the action unit incorporates an image capture unit that captures a picture for obtaining a time-series pattern, the image capture unit is installed in a predetermined place of the action unit.
6. The apparatus according to claim 5 , being formed from a recurrent type neural network, wherein:
the learning means acquires, as the result of learning the time-series pattern, a weighting factor and parametric bias parameter in the recurrent type neural network;
the output means calculates the parametric bias parameter on the basis of the weighting factor and output a time-series pattern corresponding to the parametric bias parameter resulted from the calculation; and
the adjusting means can be supplied with a time-series pattern obtained from an action made by the action unit on the basis of the time-series pattern supplied from the output means and time-series pattern obtained external teaching for the action to change the parametric bias parameter correspondingly to the input time-series pattern.
7. An information processing method, executed by a processor, comprising the steps of:
learning, utilizing a learning unit, a predetermined time-series pattern;
outputting, utilizing an outputting unit, a time-series pattern corresponding to the result of learning in the learning step; and
receiving, at a receiving unit, a time-series pattern obtained based on an action made by an action unit on the basis of a previously learned time-series pattern supplied from the outputting step and external teaching, for the action to adjust a time-series pattern supplied from the outputting step correspondingly to an input time-series pattern,
wherein the external teaching is given during a period of time when given one or more new starting or ending positions, the action to move from the given starting position to the ending position made by the action unit on the basis of a previously learned time-series pattern is made, correspondingly to a user's touch on the action unit fails,
wherein the time-series pattern learned in the learning step is previously learned time-series pattern adjusted by the adjusting means based upon external teaching, and
wherein the action unit incorporates an image capture unit that captures a picture for obtaining a time-series pattern, the image capture unit is installed in a predetermined place of the action unit.
8. A non-transitory computer-readable medium encoded with a computer-readable program that when executed on a processor, cause the processor to perform a method, the method comprising the steps of:
learning, utilizing a learning unit, a predetermined time-series pattern;
outputting, utilizing an outputting unit, a time-series pattern corresponding to the result of learning in the learning step; and
receiving, at a receiving unit, a time-series pattern obtained based on an action made by an action unit on the basis of a previously learned time-series pattern supplied from the outputting step and external teaching, for the action to adjust a time-series pattern supplied from the outputting step correspondingly to an input time-series pattern,
wherein the external teaching is given during a period of time when given one or more new starting or ending positions, the action to move from the given starting position to the ending position made by the action unit on the basis of a previously learned time-series pattern is made, correspondingly to a user's touch on the action unit fails,
wherein the time-series pattern learned in the learning step is previously learned time-series pattern adjusted by the adjusting means based upon external teaching, and
wherein the action unit incorporates an image capture unit that captures a picture for obtaining a time-series pattern, the image capture unit is installed in a predetermined place of the action unit.
9. An information processing apparatus comprising:
a learning unit for learning a predetermined time-series pattern;
an output unit for outputting a time-series pattern corresponding to the result of learning by the learning unit; and
an adjusting unit supplied with a time-series pattern obtained from an action made by an action unit on the basis of a previously learned time-series pattern supplied from the output unit and external teaching, for the action to adjust a time-series pattern supplied from the output unit correspondingly to an input time-series pattern, the learning unit learning the time-series pattern supplied from the output unit and adjusted by the adjusting unit,
wherein the external teaching is given during a period of time when, given one or more new starting or ending positions, the action to move from the given starting position to the ending position made by the action unit on the basis of a previously learned time-series pattern is made, correspondingly to a user's touch on the action unit fails,
wherein the time-series pattern learned in the learning step is previously learned time-series pattern adjusted by the adjusting means based upon external teaching, and
wherein the action unit incorporates an image capture unit that captures a picture for obtaining a time-series pattern, the image capture unit is installed in a predetermined place of the action unit.
10. An information processing apparatus comprising:
a learning unit for learning a predetermined time-series pattern;
an output unit for outputting a time-series pattern corresponding to the result of learning by the learning unit; and
an adjusting unit supplied with a time-series pattern obtained from an action made by an action unit on the basis of a previously learned time-series pattern supplied from the output unit and external teaching, for the action to adjust a time-series pattern supplied from the output unit correspondingly to an input time-series pattern,
wherein the external teaching is given during a period of time when, given one or more new starting or ending positions, the action to move from the given starting position to the ending position made by the action unit on the basis of a previously learned time-series pattern is made, correspondingly to a user's touch on the action unit fails,
wherein the time-series pattern learned by the learning unit is previously learned time-series pattern adjusted by the adjusting unit based upon external teaching, and
wherein the action unit incorporates an image capture unit that captures a picture for obtaining a time-series pattern, the image capture unit is installed in a predetermined place of the action unit.