Interactive artificial intelligence

• US 6,604,091 B2
• Filed: 09/10/1999
• Issued: 08/05/2003
• Est. Priority Date: 09/10/1999
• Status: Expired due to Fees

• Current Assignee: Yamaha Hatsudoki Kabushiki Kaisha (Yamaha Corporation)
• Original Assignee: Yamaha Hatsudoki Kabushiki Kaisha (Yamaha Corporation)
• Inventors: Sadakuni, Naoki
• Primary Examiner(s): Follansbee, John
• Assistant Examiner(s): Hirl, Joseph P.
• Application Number:
  US09/393,146
• Publication Number:
  US 20030069863A1
• Time in Patent Office: 1,425 Days
• Field of Search: 706/11, 706/25, 706/20, 706/14
• US Class Current: 706/14
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/008 based on physical entities controlled by simulated intelligence so as to replicate intelligent life forms, e.g. based on robots replicating pets or humans in their appearance or behaviour

Abstract

Behavior of a device is modified based on the device'"'"'s experience. The device comprises: (i) a sensing unit for sensing signals; (ii) an interest-generating unit programmed to generate interest-parameters in response to the sensed signals; (iii) an emotion-generating unit programmed to generate emotion-parameters in response to the sensed signals; and (iv) an actuating unit for actuating the device. The device collects information based on its interest and modifies its behavior when the device observes the same object as previously, based on variations of emotions generated when the device observed the object in the past.

Claims

1. A method for adjusting behavior of a device based on the device'"'"'s experience, said device comprising:

(i) a sensing unit for sensing signals;

(ii) an interest-generating unit programmed to generate interest-parameters;

(iii) an emotion-generating unit programmed to generate emotion-parameters; and

(iv) an actuating unit for actuating the device, the interest-generating unit and the emotion-generating unit are different units, and the interest-parameters and the emotion-parameters are different sets of parameters,

said method comprising the steps of;

(a) selecting an object of interest based on interest-parameters programmed into the device, said interest-parameters being generated by the interest-generating unit, said device having emotion-parameters (E_i) generated by the emotion-generating unit;

(b) upon sensing the object, extracting information, if any, on variations of emotions (Δ

E<sub>i−

1</sub>) associated with the object from a memory installed separately from the interest-generating unit and the emotion-generating unit, said memory storing under an index of the object, information on variations of emotions (Δ

E<sub>i−

1</sub>) previously generated by the device through past interaction with the object;

(c) modifying the initial emotions (E_i) of the device with the variations of emotions (Δ

E<sub>i−

1</sub>);

(d) actuating the device based on the modified emotions (E_i+Δ

E<sub>i−

1</sub>); and

(e) inputting in the memory, under the index of the object, variations of emotions (Δ

E_i) generated by the device during interaction with the object, thereby updating the memory.

2. The method according to claim 1, wherein the memory stores variations of emotions (Δ

E′

_i) under the index of the object, wherein Δ

E′

_i=α

Δ

E_i+(1−

α

)Δ

E<sub>i−

1</sub> (0<

α

<

1).

3. The method according to claim 1, wherein, in step (b), if no information on the object is stored in the memory, steps (d) and (e) are performed by regarding Δ

E<sub>i−

1</sub> as zero.

4. The method according to claim 3, where, in step (a), if no object of interest is selected because there is no record, an object sensed by the sensing unit is selected as an object of interest.

5. The method according to claim 1, wherein the interest-generating unit comprises a plurality of interest modules, each differently programmed to generate and output an interest-parameter, and the interest-parameter programmed into the device is that which has been assigned a highest score when the object is selected.

6. The method according to claim 1, wherein an input-output relationship of the interest-generating unit is predetermined, said interest-generating unit receiving pre-selected sensed signals and outputting the interest-parameters.

7. The method according to claim 1, wherein an input-output relationship of the emotion-generating unit is predetermined, said emotion-generating unit receiving pre-selected sensed signals and an output of the interest-generating unit, and outputting the emotion-parameters.

8. The method according to claim 1, wherein an input-output relationship of the actuating unit is predetermined, said actuating unit receiving pre-selected sensed signals and the information extracted from the memory, and actuating the device.

9. The method according to claim 1, further comprising a working memory which temporarily pools and stores information from the sensing unit, the interest-generating unit, the emotion-generating unit, and the memory used in step (b) until the device completes its action, and which outputs information to the interest-generating unit, the actuating unit, and the memory used in step (b).