Back to results
US 8,527,434 B2Granted
Information processing device, information processing method, and program
AssigneeSony Corporation
InventorsKuniaki Noda, Kohtaro Sabe, Yukiko Yoshiike, Kenta Kawamoto, Kenichi Hidai
Life of this patent
Estimating the term…
Abstract
An information processing device includes: a learning section configured to learn a state transition probability model defined by state transition probability for each action of a state making a state transition due to an action performed by an agent capable of performing action and observation probability of a predetermined observed value being observed from the state, using an action performed by the agent and an observed value observed in the agent when the agent has performed the action.
Claims20
An information processing device comprising:
- at least one processor;
- and
- a storage device storing instructions that, when executed by the at least one processor, cause the information processing device to perform operations comprising:
- learning a state transition probability model defined by
- state transition probability for each action of a state making a state transition due to an action performed by an agent capable of performing action, and
- observation probability of a predetermined observed value being observed from said state, using an action performed by said agent and an observed value observed in said agent when said agent has performed the action.
3 more elements
An information processing method of an information processing device, said information processing method comprising a step of:
- learning, by at least one processor, a state transition probability model defined by
- state transition probability for each action of a state making a state transition due to an action performed by an agent capable of performing action, and
- observation probability of a predetermined observed value being observed from said state, using an action performed by said agent and an observed value observed in said agent when said agent has performed the action.
A program embodied on a non-transitory computer-readable medium comprising instructions that, when executed by at least one processor, cause a computer to perform operations comprising:
- learning a state transition probability model defined by
- state transition probability for each action of a state making a state transition due to an action performed by an agent capable of performing action, and
- observation probability of a predetermined observed value being observed from said state, using an action performed by said agent and an observed value observed in said agent when said agent has performed the action.
An information processing device comprising:
- at least one processor;
- a storage device storing instructions that, when executed by the at least one processor, cause the information processing device perform operations, said instructions comprising:
- state recognizing instructions for recognizing present conditions of an agent capable of performing action using an action performed by said agent and an observed value observed in said agent when said agent has performed the action on the basis of a state transition probability model obtained by learning said state transition probability model defined by
- state transition probability for each action of a state making a state transition due to an action performed by said agent, and
- observation probability of a predetermined observed value being observed from said state,
- instructions for using an action performed by said agent and an observed value observed in said agent when said agent has performed the action, and obtaining a present state as a state of said state transition probability model, the state of said state transition probability model corresponding to the present conditions;
- goal determining instructions for determining one of states of said state transition probability model as a goal state set as a goal;
- and
- action determining instructions for calculating an action plan as an action series that maximizes likelihood of state transition from said present state to said goal state on the basis of said state transition probability model, and determining an action to be performed next by said agent according to the action plan.
6 more elements
An information processing method of an information processing device, said information processing method comprising the steps of:
- recognizing, by at least one processor, present conditions of an agent capable of performing action using an action performed by said agent and an observed value observed in said agent when said agent has performed the action on the basis of a state transition probability model obtained by learning said state transition probability model defined by
- state transition probability for each action of a state making a state transition due to an action performed by said agent, and
- observation probability of a predetermined observed value being observed from said state, using an action performed by said agent and an observed value observed in said agent when said agent has performed the action, and obtaining a present state as a state of said state transition probability model, the state of said state transition probability model corresponding to the present conditions;
- determining one of states of said state transition probability model as a goal state set as a goal;
- and
- calculating an action plan as an action series that maximizes likelihood of state transition from said present state to said goal state on the basis of said state transition probability model, and determining an action to be performed next by said agent according to the action plan.
3 more elements
A program embodied on a non-transitory computer-readable medium comprising instructions that, when executed by at least one processor, cause a computer to perform operations comprising:
- recognizing present conditions of an agent capable of performing action using an action performed by said agent and an observed value observed in said agent when said agent has performed the action on the basis of a state transition probability model obtained by learning said state transition probability model defined by
- state transition probability for each action of a state making a state transition due to an action performed by said agent, and
- observation probability of a predetermined observed value being observed from said state, using an action performed by said agent and an observed value observed in said agent when said agent has performed the action, and obtaining a present state as a state of said state transition probability model, the state of said state transition probability model corresponding to the present conditions;
- determining one of states of said state transition probability model as a goal state set as a goal;
- and
- calculating an action plan as an action series that maximizes likelihood of state transition from said present state to said goal state on the basis of said state transition probability model, and determining an action to be performed next by said agent according to the action plan.
3 more elements
An information processing device comprising:
- at least one processor;
- a storage device storing learning instructions that, when executed by the at least one processor, cause the information processing device to learn a state transition probability model defined by
- state transition probability for each action of a state making a state transition due to an action performed by an agent capable of performing action, and
- observation probability of a predetermined observed value being observed from said state, using an action performed by said agent and an observed value observed in said agent when said agent has performed the action.
1 more element
An information processing device comprising:
- at least one processor;
- a storage device storing instructions that, when executed by the at least one processor, cause the information processing device to perform operations, said instructions comprising:
- state recognizing instructions configured to recognize present conditions of an agent capable of performing action using an action performed by said agent and an observed value observed in said agent when said agent has performed the action on the basis of a state transition probability model obtained by learning said state transition probability model defined by
- state transition probability for each action of a state making a state transition due to an action performed by said agent, and
- observation probability of a predetermined observed value being observed from said state, using an action performed by said agent and an observed value observed in said agent when said agent has performed the action, and obtaining a present state as a state of said state transition probability model, the state of said state transition probability model corresponding to the present conditions;
- goal determining instructions configured to determine one of states of said state transition probability model as a goal state set as a goal;
- and
- action determining instructions configured to calculate an action plan as an action series that maximizes likelihood of state transition from said present state to said goal state on the basis of said state transition probability model, and determine an action to be performed next by said agent according to the action plan.
5 more elements
Dependencies are inferred from the claim text.