JP2016006615A - Animal-related information processing system and animal-related information processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress the occurrence of the problems that it takes a time, that cost is expensive, that human motivation affects, and that attention has to be paid to personal data handling problem.SOLUTION: An animal related information processing system includes a client device (1), a front stage processor (2) and an animal related information processor (3). A control part of the front stage processor (2) determines a task on the basis of a rule with reference to a state DB (26a) storing states of an animal in accordance with the reception of a task request from the client device (1), and transmits the task to the animal related information processor (3) to make the animal perform the task.

Description

  The present invention relates to an animal-related information processing system and an animal-related information processing method for causing an animal to perform a task.

  As a crowdsourcing service for requesting work from an unspecified number of people, a microtask type in which tasks are subdivided and a competition type in competitive bidding are known (as an example of a microtask type, for example, Non-Patent Document 1). . As a method for causing an animal to perform a task, a method of applying operant conditioning is known in the field of animal recognition. Operant boxes and skinner boxes are known as devices that enable efficient operant conditioning (for example, Non-Patent Documents 2 and 3).

Satoshi Koyama et al., “Highly accurate label integration using worker confidence in 2M5-OS-07b-2 crowdsourcing”, 2M5-OS-07b Organized Session OS-07 Human Computation and Crowdsourcing-2 Journal, June 5, 2013 B. F. SKINNER, “CASE HISTORY IN SCIENTIFIC METHOD”, (USA), 1956, p. 221-233 Oskar Pineno, “Arduino Box: A Low-Cost and Open-Source SkinxerUx” with a low-cost and open source skinner box using iPod Touch® and an Arduino Microcontroller an iPod Touch and an Arduino Microcontroller), [Online], 2013, [May 27, 2014 search], Internet <http://people.hofstra.edu/oskar_pineno/ArduiPod.htm>

  However, in the method shown in Non-Patent Document 1, an animal is not included in the worker who performs the work, and the worker who performs the work is only a human. For this reason, there are problems such as time consuming, high cost, depending on human motivation, and attention to the problem of handling personal data (information including information that is not widely personally identifiable about individuals). is there. On the other hand, some tasks such as crowdsourcing can be done not only by computers and humans but also by animals other than humans. In addition, there are things that can only be done by animals, not computers and humans.

  The methods shown in Non-Patent Documents 2 and 3 are not intended for crowdsourcing using animals.

  The object of the present invention made in view of such points is that problems such as time-consuming, expensive, dependent on human motivation, and the problem of handling personal data occur in crowdsourcing. An object of the present invention is to provide an animal-related information processing system and an animal-related information processing method.

  In order to solve the above problems, an animal-related information processing system according to the present invention includes a client device, a pre-processing device, and an animal-related information processing device, and the control unit of the pre-processing device performs a task from the client device. In response to receiving the request, the task is determined based on the rule with reference to the state DB storing the state of the animal, and the task is transmitted to the animal-related information processing apparatus to cause the animal to perform the task. It is characterized by that.

  In order to solve the above problems, an animal-related information processing method according to the present invention is an animal-related information processing method by an animal-related information processing system including a client device, a pre-processing device, and an animal-related information processing device. Determining a task based on a rule by referring to a state DB storing an animal state in response to receiving a task request from the client device by the control unit of the upstream processing device; And a step of transmitting the task to the animal-related information processing apparatus and causing the animal to perform the task by a control unit of a processing apparatus.

  According to the present invention, it is possible to suppress the occurrence of problems in crowdsourcing, such as time consuming, expensive, dependent on human motivation, and having to be aware of personal data handling problems. It becomes possible.

It is a figure showing a schematic structure of an animal related information processing system concerning one embodiment of the present invention. It is a functional block diagram of the client apparatus in the animal related information processing system of FIG. (A) is a functional block diagram of the pre-processing apparatus in the animal related information processing system of FIG. 1, and (b) is a functional block diagram of the multiple animal grouping apparatus in the animal related information processing system of FIG. FIG. 2 is a functional block diagram of an animal-related information processing apparatus in the animal-related information processing system of FIG. 1. It is a functional block diagram of the back | latter stage processing apparatus in the animal related information processing system of FIG. It is a figure which shows an example of the mode of a learning phase. It is a figure which shows the operation | movement flow in the state monitoring phase of the animal related information processing system which concerns on Embodiment 1 of this invention. It is a figure which shows the operation | movement flow in the learning phase of the animal related information processing system which concerns on Embodiment 1 of this invention. It is a figure which shows the operation | movement flow in the task phase of the animal related information processing system which concerns on Embodiment 1 of this invention. It is a figure which shows the operation | movement flow in the task phase of the animal related information processing system which concerns on Embodiment 2 of this invention. It is a figure which shows an example of a mode that a task is performed. It is a figure which shows the operation | movement flow in the task phase of the animal related information processing system which concerns on Embodiment 3 of this invention. It is a figure which shows the operation | movement flow in the state monitoring phase of the animal related information processing system which concerns on the modification of this invention.

  Hereinafter, each embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing a schematic configuration of an animal-related information processing system according to an embodiment of the present invention. The schematic configuration illustrated in FIG. 1 is an example, and each configuration can be arbitrarily combined. A storage unit described as being inside each configuration may be provided outside each configuration and referred to by each device when necessary. Each function of the information processing system according to the present invention will be described, but it is not intended to exclude other functions of the information processing system.

  In one embodiment, the client apparatus 1, the pre-stage processing apparatus 2, the multiple animal grouping apparatus 2a, the animal-related information processing apparatus 3 (also referred to as “animal engine”), and the post-stage processing apparatus 4 are connected via a network. The number of devices connected to the network may be one or plural. Each device may be a stationary terminal or a portable terminal. In the present invention, “animal” refers to animals other than humans, including insects and fungi.

  FIG. 2 is a functional block diagram of the client device 1 in the animal-related information processing system of FIG. The client device 1 includes a control unit 11, a task transmission unit 12, and a task result reception unit 13. Various operations of the client device 1 are controlled by the control unit 11.

  The task transmission unit 12 creates a task based on manual input from the user of the client device 1 or external input acquired from another device. A task is a problem to be solved whose correct answer is unknown. The task transmission unit 12 transmits the created task to the pre-processing device 2.

  The task result receiving unit 13 receives a task result from the post-processing device 4.

  FIG. 3A is a functional block diagram of the pre-processing device 2 according to the embodiment of the present invention. The pre-processing device 2 includes a control unit 21, a state management unit 22, a learning problem management unit 23, a results management unit 24, a task management unit 25, and a storage unit 26. The storage unit 26 includes a state DB 26a, a learning problem DB 26b, a results DB 26c, and a task DB 26d. Various operations of the pre-processing device 2 are controlled by the control unit 21.

  The state management unit 22 requests the animal-related information processing device 3 to transmit animal state information. Further, the state management unit 22 receives the state information and stores it in the state DB 26a.

  The learning problem management unit 23 extracts learning problems from the learning problem DB 26b using each parameter as a key. Learning means giving a pseudo-task whose correct answer is known in advance and improving the ability to perform (execute) the task. The learning problem management unit 23 determines and transmits a learning problem to be transmitted to the animal-related information processing device 3 with reference to the results DB 26c. The learning problem management unit 23 can adjust the frequency of transmitting learning problems.

  The grade management unit 24 receives the transmitted learning grade and stores it in the grade DB 26c.

  The task management unit 25 receives the task request from the client device 1 and stores it in the task DB 26d. The task management unit 25 refers to information in the results DB 26c (current learning results, learning curve, etc.) and information in the status DB 26a (animal type, current animal status, etc.), and based on the rules, which task Whether to transmit to the animal-related information processing device 3 is determined. The task management unit 25 may refer to only the information in the state DB 26a and determine which task is to be transmitted to the animal related information processing device 3 based on the rule. The task management unit 25 transmits input information (task type, requested score, current score, etc.) to the multiple animal grouping device 2a. The task management unit 25 transmits the task to the animal-related information processing device 3 corresponding to the ID transmitted from the multiple animal grouping device 2a.

  FIG.3 (b) is a functional block diagram of the multiple animal grouping apparatus 2a in the animal related information processing system of FIG. The multiple animal grouping device 2a includes a control unit 21a and a multiple animal grouping unit 22a. Various operations of the multiple animal grouping device 2a are controlled by the control unit 21a. The multiple animal grouping unit 22a is based on the input information received from the task management unit 25, information in the results DB 26c (current learning results, learning curves, etc.), information in the state DB 26a (type of animals, current state), etc. The ID (single or plural) of the animal-related information processing apparatus 3 suitable for the task transmission destination is output as output information. The multiple animal grouping unit 22a outputs the output information to the pre-processing device 2. The multiple animal grouping unit 22a may be provided in the pre-treatment device 2.

  FIG. 4 is a functional block diagram of the animal-related information processing apparatus 3 according to an embodiment of the present invention. The animal-related information processing apparatus 3 includes a control unit 31, a state acquisition unit 32, a learning / task management unit 33, a learning / task presentation unit 34, and a learning / task buffer unit 35. Various operations of the animal-related information processing device 3 are controlled by the control unit 31.

  The state acquisition unit 32 acquires animal state information. The state acquisition unit 32 acquires the body temperature of the animal, the brightness of the place where the animal exists, the image or video of the animal, the voice of the animal, and the like using, for example, a sensor, a camera, and a microphone.

  The learning / task management unit 33 receives data from the pre-processing device 2, determines whether the received data is a learning problem or a task, and stores it in the learning / task buffer unit 35. The learning / task management unit 33 determines whether or not the animal's answer is successful, and transmits it to the pre-processing device 2 as a learning result.

  The learning / task presenting section 34 extracts a learning problem or task from the learning / task buffer section 35 and presents it to the animal for answering. The learning / task presentation unit 34 may extract tasks with priority over learning problems. The learning / task presentation unit 34 encourages learning of an animal by giving an incentive to the animal when the animal gives a desired answer (for example, when the answer is correct). An example of an incentive when using a skinner box is to feed. In addition, the learning / task presentation unit 34 transmits the processing results of the animal-related information processing device 3 (answer data, animal information (animal type, current state, learning result, etc.)) to the subsequent processing device 4. .

  The learning / task buffer unit 35 stores information such as learning data (including correct / incorrect information on learning problems), task data, learning problem types, task types, learning problem difficulty levels, and task difficulty levels.

  FIG. 5 is a functional block diagram of the post-processing device 4 according to an embodiment of the present invention. The post-processing device 4 includes a control unit 41, a processing result check unit 42, a score calculation unit 43, a task result notification unit 44, a statistical processing unit 45, a transmission destination management unit 46, and a task request creation unit 47. Various operations of the post-processing device 4 are controlled by the control unit 41.

  The processing result check unit 42 receives the processing result of the animal-related information processing device 3 from the animal-related information processing device 3 and checks whether there is clearly any unnatural data. The processing result check unit 42 does not use unnatural data.

  The score calculation unit 43 calculates the likelihood (score) of the processing result using an algorithm or the like.

  The task result notification unit 44 shapes the processing result and score data of the animal-related information processing device 3 into a form that can be received by the client device 1 and transmits the data to the client device 1.

  The statistical processing unit 45 performs statistical processing using a majority vote or the like using processing results of the plurality of animal-related information processing devices 3 and animal information. The statistical processing unit 45 combines the results of statistical processing into one. This statistical processing improves the accuracy of task results.

  The transmission destination management unit 46 determines the transmission destination of the processing result as either the client device 1 or the pre-processing device 2 based on the transmission condition. As a transmission condition parameter, a score of a processing result, an engine history indicating which animal-related information processing device 3 (or a processing device described in a third embodiment described later) is used, an elapsed time since a task request is made ( Timeout time). When the transmission destination management unit 46 determines that the processing result is to be transmitted to the pre-processing device 2, the processing result is passed to the task request creation unit 47. The task request creation unit 47 adds the engine history of the animal-related information processing apparatus 3 (or the processing apparatus described in the third embodiment) that is shaped and used for the task request and transmits the processing result data. On the other hand, when the transmission destination management unit 46 determines to transmit the processing result to the client device 1, the processing result is passed to the task result notification unit 44. The task result notifying unit 44 shapes the processing result and score data into a form that can be received by the client device 1 and transmits it as a task result.

[Embodiment 1]
Hereinafter, Embodiment 1 of the present invention will be described. In the first embodiment, the number of animal-related information processing devices 3 is one.

  In the first embodiment, the processing performed by the animal-related information processing system has three phases. The three phases are a state monitoring phase, a learning phase, and a task phase. The state monitoring phase is performed in parallel with the learning phase and the task phase in an asynchronous manner. When the client device 1 does not make a task request, the animal-related information processing system is in the learning phase. When the client device 1 makes a task request, the animal-related information processing system shifts to the task phase.

(Status monitoring phase)
The state monitoring phase is a phase for monitoring the state of the animal. The animals are for example in a skinner box (or operant box). When the animal-related information processing device 3 is requested by the pre-processing device 2 to transmit the current state of the animal, the animal state, whether or not the animal type and task can be performed as the animal state, and whether or not learning is possible The local time of the place where the animal is located, the position information of the animal (GPS information: Global Positioning System information), etc. are acquired and transmitted to the pre-processing device 2. The state is stored in the state DB 26a. Since the state of the animal changes, the animal-related information processing device 3 may perform acquisition and transmission a plurality of times. The animal related information processing apparatus 3 can adjust the frequency of acquisition and transmission.

  The animal-related information processing device 3 may determine the “animal type” using information registered in advance, or may determine by performing object recognition in an image obtained by imaging an animal. The animal-related information processing device 3 determines “whether or not the task can be performed” and “whether or not learning is possible” based on “whether or not the animal is awake” or “whether or not the animal is full” May be.

(Learning phase)
The learning phase is a phase in which animals learn. Because the correct answer rate of animals can vary greatly depending on learning, keep the animals always learning along with their tasks. Learning is performed using a skinner box (or an operant box) that feeds the animal when it selects the correct answer.

  The pre-processing device 2 extracts a learning problem from the learning problem DB 26b based on a predetermined rule and transmits it to the animal related information processing device 3. The learning problem DB 26b stores a learning problem, a type of learning problem, a difficulty level of the learning problem, and the like.

  For example, the “predetermined rule” is that the pre-processing device 2 extracts and transmits a learning problem having a high difficulty level when it is determined that an animal is waking up with reference to the state DB 26a.

  The animal-related information processing device 3 determines whether the data received from the pre-processing device 2 is a learning problem or a task, and stores them in the learning / task buffer unit 35. At the reference time, the animal-related information processing device 3 extracts the learning problem from the learning / task buffer unit 35 and presents it to the animal. The animal-related information processing device 3 determines whether or not the animal's answer is correct and transmits it to the pre-processing device 2 as a learning result. The pre-stage processing apparatus 2 stores the learning results in the results DB 26c as numerical values such as correct answer rates and deviation values. In the second and subsequent learning phases, the pre-processing device 2 may determine a learning problem to be transmitted to the animal related information processing device 3 with reference to the learning results stored in the results DB 26c. As a determination method, there is a method of giving a difficult problem to an animal with good learning results and giving an easy problem to an animal with poor learning results. This makes it possible to grasp the degree of learning in real time and provide feedback.

  An example of the learning phase will be specifically described with reference to FIG. For example, a terminal (for example, a tablet terminal) is directed to a chicken in a poultry farm, and the display unit of the terminal presents two images divided into left and right. The image displayed on the right changes as the reference time elapses. When an object in the image displayed on the right side matches an object in the image displayed on the left side, the animal performs a predetermined action (such as picking an image displayed on the display unit and pulling a lever in the skinner box). The animals are fed. In this way, the animal learns to recognize the identity of the object in the image.

(Task phase)
The task phase is a phase in which an animal performs a task. The client device 1 makes a task request to the upstream processing device 2. In response to receiving the task request, the pre-processing device 2 creates a task and stores it in the task DB 26d. The pre-processing device 2 refers to the state DB 26a and the results DB 26c, and determines which task is to be transmitted to the animal-related information processing device 3 based on the rules. As an example of a method for determining which task is to be transmitted when, for an animal having a good learning result, there is a method of transmitting a problem that is highly difficult during the time the animal is awake. The pre-processing device 2 transmits the determined task to the animal related information processing device 3.

  The animal-related information processing device 3 determines whether the received data is a learning problem or a task, and stores it in the learning / task buffer unit 35. At the reference time, the animal-related information processing device 3 extracts a task from the learning / task buffer unit 35, presents it to the animal, and makes it answer. The animal-related information processing device 3 may extract the task from the learning / task buffer unit 35 with priority over the learning problem.

  When the animal answers, the animal-related information processing device 3 transmits the response data and the animal information (animal type, current state of the animal, learning result at the time of learning, etc.) to the subsequent processing device 4 as a processing result. The post-processing device 4 receives the processing result and checks whether or not the answer data is clearly strange (whether it is common sense or whether the correct answer rate is too high compared to the learning result at the time of learning, etc.). . The post-processing device 4 calculates the probability (score) of the processing result using animal information or the like. The post-processing device 4 shapes the processing result and the score data into a form that can be received by the client device 1. The post-processing device 4 transmits the shaped data to the client device 1 as a task result.

  FIG. 7 is a diagram showing an operation flow in the state monitoring phase of the animal-related information processing system according to the first embodiment of the present invention. The pre-processing device 2 sends to the animal-related information processing device 3 the current animal status information (type of animal, whether or not the task can be performed, whether or not learning is possible, the local time where the animal exists, (Position information etc. of the place where there is) is requested (step S1). The animal-related information processing device 3 acquires the current animal state information (if the animal state has been acquired in the past, the changed state information) (step S2), and transmits it to the pre-processing device 2 ( Step S3). The pre-stage processing apparatus 2 stores the acquired animal state information in the state DB 26a (step S4).

  FIG. 8 is a diagram showing an operation flow in the learning phase of the animal-related information processing system according to the first embodiment of the present invention. The pre-processing device 2 extracts a learning problem from the learning problem DB 26b based on the rule (step S11). The pre-stage processing apparatus 2 refers to the state DB 26a, confirms the state of the animal, and transmits a learning problem (step S12). The animal-related information processing device 3 determines whether the received data is a learning problem or a task, and stores it in the learning / task buffer unit 35 (step S13). The animal related information processing device 3 extracts the learning problem from the learning problem DB 26b and presents it to the animal (step S14). The animal-related information processing device 3 determines whether or not the animal has correctly answered the learning problem (step S15), and transmits the learning result to the pre-processing device 2 (step S16). The pre-stage processing apparatus 2 stores the learning result in the result DB 26c (step S17). In the second and subsequent learnings, the pre-processing device 2 may extract learning problems based on the rules with reference to the learning results in the results DB 26c (step S18). The pre-processing device 2 refers to the state DB 26a to confirm the state of the animal, and transmits the extracted learning problem to the pre-processing device 2 (step S19).

  FIG. 9 is a diagram showing an operation flow in the task phase of the animal-related information processing system according to the first embodiment of the present invention. The client device 1 requests a task from the upstream processing device 2 (step S21). In response to receiving the task request, the pre-processing device 2 creates a task and stores it in the task DB 26d (step S22). The pre-processing device 2 refers to the result DB 26c and the state DB 26a and determines which task is to be transmitted to the animal-related information processing device 3 based on the rule (step S23). The pre-stage processing apparatus 2 transmits the determined task to the animal related information processing apparatus 3 (step S24). The animal-related information processing device 3 determines whether the received data is a learning problem or a task (in this case, it is determined as a task), and stores it in the learning / task buffer unit 35 (step S25). The animal-related information processing device 3 extracts the task from the learning / task buffer unit 35 and presents it to the animal (step S26). The animal performs the presented task.

  The animal-related information processing device 3 transmits the processing result (answer data, animal information (animal type, current animal state, learning result during learning, etc.)) to the post-processing device 4 (step S27). The post-processing device 4 receives the processing result and checks whether or not there is clearly unnatural data (step S28). The post-processing device 4 calculates the probability (score) of the processing result from the animal information or the like (step S29), and shapes the processing result and the score data into a form that can be received by the client device 1 (step S30). The post-processing device 4 transmits the task result to the client device 1 (step S31).

[Embodiment 2]
The second embodiment of the present invention will be described below. In the second embodiment, the number of animal-related information processing devices 3 is plural. Description of matters common to the first embodiment is omitted.

  Also in the second embodiment, the processing performed by the animal-related information processing system has a state monitoring phase, a learning phase, and a task phase as three phases.

(Status monitoring phase)
In the state monitoring phase of the second embodiment, the pre-processing device 2 performs the same processing as that performed on the animal-related information processing device 3 of the first embodiment for each of the two animal-related information processing devices 3. Description is omitted.

(Learning phase)
In the learning phase of the second embodiment, the pre-processing device 2 performs the same processing as that performed on the animal-related information processing device 3 of the first embodiment for each of the two animal-related information processing devices 3. Description is omitted.

(Task phase)
The task phase is a phase in which an animal performs a task. In the task phase, the client device 1 makes a task request to the upstream processing device 2. In response to receiving the task request, the pre-processing device 2 creates a task and stores it in the task DB 26d. The pre-processing device 2 transmits the task type, the required score, the current score, and the like as input information to the multiple animal grouping device 2a.

  Animals vary greatly in ability depending on individuals and species. Therefore, the multi-animal grouping device 2a grasps the ability of animals and performs dynamic grouping. The multiple animal grouping device 2a refers to the input information, information in the results DB 26c (current learning results, learning curves, etc.), and information in the state DB 26a (type of animals, current state of animals, etc.) Based on at least one of the predetermined algorithms, a plurality of IDs of the animal related information processing devices 3 to be used in the current task phase are output as output information. The multiple animal grouping device 2 a transmits a plurality of IDs to the pre-processing device 2. The pre-processing device 2 duplicates and transmits the same created task to the animal-related information processing device 3 corresponding to the received plurality of IDs.

  Each of the plurality of animal related information processing devices 3 that has received the task performs the same processing as the animal related information processing device 3 of the first embodiment.

  The subsequent processing device 4 that has received a plurality of processing results from the plurality of animal related information processing devices 3 checks whether or not the answer data is clearly unnatural. The post-processing device 4 performs statistical processing using a majority vote using a plurality of processing results, animal information, and the like. The post-processing device 4 combines the processing results into one by statistical processing. By performing such statistical processing, the accuracy of the task result is improved. The post-processing device 4 calculates the probability (score) of the processing result using animal information or the like. The post-processing device 4 shapes the processing result and the score data into a form that can be received by the client device 1. The post-processing device 4 transmits the shaped data to the client device 1 as a task result.

  The operation flow of the state monitoring phase and the learning phase of the second embodiment is the same as the operation flow of the state monitoring phase and the learning phase of the first embodiment except that there are a plurality of animal-related information processing devices 3 that perform processing. . Therefore, the description of the operation flow of the state monitoring phase and the learning phase of Embodiment 2 is omitted.

  FIG. 10 is a diagram showing an operation flow in the task phase of the animal-related information processing system according to the second embodiment of the present invention. In the operation flow of the second embodiment, it is assumed that there are two animal-related information processing devices 3.

  The operation flow of the task phase will be described. The client device 1 makes a task request to the upstream processing device 2 (step S41). The pre-processing device 2 creates a task in response to receiving the task request and stores it in the task DB 26d (step S42). The pre-processing device 2 transmits the input information to the multiple animal grouping device 2a (step S43). The multiple animal grouping device 2a refers to the input information, information in the results DB 26c (current learning results, learning curve, etc.) and information in the status DB 26a (animal type, current status, etc.) Two IDs (A and B) of the animal-related information processing device 3 are output as output information based on at least one (step S44). The multi-animal grouping device 2a transmits the output information to the pre-processing device 2 (step S45). The upstream processing device 2 duplicates and transmits the same task to the animal related information processing devices 3 (A and B) corresponding to the received IDs of the two animal related information processing devices (step S46).

  Each of the two animal-related information processing devices 3 performs the same processing as in steps S25 to S27. Further, the post-processing device 4 performs the same process as step S28. Description is omitted.

  Next, the post-processing device 4 performs statistical processing using a majority vote using the two processing results and animal information, etc., and collects the results into one processing result (step S47). This improves the accuracy of task results. The post-processing device 4 calculates the probability (score) of the processing result from the animal information or the like (step S48). The post-processing device 4 shapes the processing result and score data into a form that can be received by the client device 1 (step S49), and transmits it as a task result (step S50).

[Embodiment 3]
Hereinafter, Embodiment 3 of the present invention will be described. In the third embodiment, the number of animal related information processing devices 3 is plural, which is the same as the number of animal related information processing devices 3 in the second embodiment. Description of matters common to the second embodiment is omitted.

  Also in the third embodiment, the processing performed by the animal-related information processing system has a state monitoring phase, a learning phase, and a task phase as three phases.

(Status monitoring phase and learning phase)
Since the state monitoring phase and the learning phase of the third embodiment are the same as the state monitoring phase and the learning phase of the second embodiment, description thereof is omitted.

(Task phase)
The task phase is a phase in which an animal performs a task. In the task phase of the third embodiment, the animal-related information processing system may acquire a task result by causing not only an animal but also a computer or a human to perform a task. Specifically, the animal-related information processing system includes not only the animal-related information processing device (animal engine) 3 but also other processing devices (engines using computers (image recognition engines, etc.), engines using humans (crowd sourcing). Etc.))).

  As an example of using not only the animal-related information processing device 3 but another processing device, in the task of selecting an image determined to be identical to a specified image from a plurality of images, the computer first performs image recognition. Do. Then, there is a method in which the computer outputs a plurality of correct answer candidates, and then the animal selects one of the candidates determined to be correct.

  With reference to FIG. 11, an example of a task phase in a case where an image recognition engine as a computer performs a task will be schematically described. (1) By a user operation, the terminal (corresponding to the client device 1) images a specific target using the imaging function. (2) The terminal sends a captured image to the pre-processing device 2 to make a task request. When receiving the task request, the pre-processing device 2 creates and determines a task. The pre-processing device 2 transmits the determined task to the image recognition engine. When receiving a task including a captured image, the image recognition engine performs image recognition on the captured image. (3) The image recognition engine outputs a plurality of correct answer candidates from the result of image recognition. (4) When a terminal (corresponding to the animal-related information processing device 3) directed to a plurality of chickens in the poultry farm receives the correct candidate, each of the correct candidate images and the captured image are displayed separately on the left and right. Present to the department. (5) The plurality of chickens select an image that is determined to include the same object as the captured image from among the correct candidate images. The terminal accepts the selection. (6) The post-processing device 4 receives a plurality of selection results from a plurality of chickens from a terminal directed to the chicken and performs statistical processing. (7) The post-processing device 4 derives the most probable result (the highest identity with the target imaged by the user at the terminal). The post-processing device 4 transmits the result as a task result to the user's terminal.

  The animal-related information processing device 3 and the other processing device may perform tasks in parallel, or may perform the tasks in order as described above.

  For animals, even if the correct answer is obvious, the recognition accuracy may not be 100% in image recognition by a computer. However, there is a high probability that a correct answer is included in a plurality of recognition result candidates output by the computer as a result of image recognition. Therefore, this method is effective.

  The task phase will be described specifically. Until the post-processing device 4 receives processing results from a plurality of animal-related information processing devices 3 and calculates the certainty (score) of the processing results from animal information or the like, other processing devices may perform tasks. Since it is the same as that of Embodiment 2 except for a good point, description is abbreviate | omitted.

  The post-processing device 4 determines which of the pre-processing device 2 and the client device 1 is to transmit the processing result based on the transmission condition.

  The transmission condition parameter is a score of the processing result, an engine history indicating which animal-related information processing device 3 (or other processing device) is used, a time elapsed since the task request was made by the client device 1, and the like.

  For example, when the post-processing device 4 determines that it is difficult to specify one correct answer from the score of the processing result, the post-processing device 4 requests the pre-processing device 2 for the same task. The pre-processing device 2 requests the animal-related information processing device 3 or another processing device that has not been used yet for the same task. This improves the accuracy of task results. Further, from the engine history, for example, when the post-processing device 4 determines that the task is not yet requested to the engine using a human, the post-processing device 4 transmits the task to the engine using the human so as to transmit the task. Request device 2. On the other hand, when the post-processing device 4 determines from the engine history that the processing results have already been output using all the engines, the post-processing device 4 transmits the task result to the client device 1.

  When the post-processing device 4 transmits a task request to the pre-processing device 2, the post-processing device 4 shapes the processing result data for the task request and uses any animal-related information processing device 3 or other processing device. Send it with the history added. On the other hand, when the post-processing device 4 transmits a task result to the client device 1, the post-processing device 4 transmits the processing result and the score data in a form that can be received by the client device 1.

  The operation flow in the state monitoring phase and the learning phase in the third embodiment is the same as the operation flow in the state monitoring phase and the learning phase in the second embodiment. Therefore, the description of the operation flow of the state monitoring phase and the learning phase of Embodiment 3 is omitted. The task phase process of the third embodiment is the same as that of the second embodiment in which the client apparatus 1 makes a task request to the pre-stage processing apparatus 2, and the plurality of animal related information processing apparatuses 3 learn the task / task buffer unit 35. The process is the same as the process from extraction to presentation to the animal (except that other processing devices may perform tasks). Therefore, description of the operation flow of the processing in this part of the task phase is omitted.

  FIG. 12 is a diagram illustrating an operation flow in the task phase of the animal-related information processing system according to the third embodiment of the present invention.

  Each of the two animal-related information processing devices 3 transmits a processing result (response data, animal information (animal type, current animal state, learning result during learning, etc.)) to the subsequent processing device 4 (step S51). ). The post-processing device 4 receives the processing result and checks whether or not there is clearly unnatural data (step S52). In step S52, the post-processing device 4 may obtain a task result from another processing device. The post-processing device 4 performs statistical processing by majority vote using a plurality of processing results and animal information (step S53). The processing results are combined into one. The post-processing device 4 calculates the probability (score) of the processing result from the animal information or the like (step S54), and determines whether to transmit the processing result to the pre-processing device 2 or the client device 1 (step S55). . If it is determined that the processing result is to be transmitted to the client device 1, the subsequent processing device 4 shapes the processing result and score data into a form that can be received by the client device 1, and transmits the result to the client device 1 as a task result (step S56). . On the other hand, if the post-processing device 4 determines that the processing result is to be transmitted to the pre-processing device 2, the processing result data is formatted for task request, and the animal-related information processing device 3 or the engine history of another processing device used is used. Is transmitted to the pre-processing device 2 as a task request (step S57).

  As described above in Embodiments 1 to 3, it is possible to suppress the occurrence of problems such as time consuming, high cost, depending on human motivation, and attention to personal data handling problems. It becomes.

[Modification]
In the state monitoring phase of the first to third embodiments, the pre-processing device 2 requests the animal-related information processing device 3 for current animal state information. Next, the animal-related information processing device 3 acquires current animal state information and transmits it to the pre-processing device 2. On the other hand, in the modification, the pre-processing device 2 does not request the animal-related information processing device 3 for current animal state information. When the predetermined condition is satisfied, the animal related information processing device 3 acquires current animal state information and transmits it to the pre-processing device 2. The predetermined condition is, for example, that the animal-related information processing device 3 has detected that the state of the animal has changed, or that a reference time has elapsed.

  FIG. 13 is a diagram showing an operation flow in the state monitoring phase of the animal-related information processing system according to the modified example of the present invention. The animal related information processing apparatus 3 determines whether or not a predetermined condition is satisfied (step S61). When it is detected that the state of the animal has changed, or when the reference time has elapsed (Yes in step S61), the animal-related information processing device 3 acquires the current animal state information (step S62), and the pre-processing It transmits to the apparatus 2 (step S63). The pre-stage processing apparatus 2 stores the acquired animal state information in the state DB 26a (step S64). If the animal state has not changed and the reference time has not elapsed (No in step S61), the animal-related information processing device 3 repeats step S61.

  When the client device 1, the pre-processing device 2, the multi-animal grouping device 2a, the animal-related information processing device 3 and / or the post-processing device 4 according to the present invention are configured by a computer, processing contents for realizing each function are described. This program can be realized by storing the program in a storage unit inside or outside the computer, and reading and executing the program by a central processing unit (CPU) of the computer. In addition, such a program can be distributed by selling, transferring, or lending a portable recording medium such as a DVD or a CD-ROM, and such a program is stored in a storage unit of a server on a network, for example. And the program can be distributed by transferring the program from the server to another computer via the network. In addition, a computer that executes such a program can temporarily store, for example, a program recorded on a portable recording medium or a program transferred from a server in its own storage unit. As another embodiment of the program, the computer may directly read the program from a portable recording medium and execute processing according to the program, and each time the program is transferred from the server to the computer. In addition, the processing according to the received program may be executed sequentially. Therefore, the present invention is not limited to the above-described embodiments, and various modifications can be made by combining or partially deleting the same without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1 Client apparatus 2 Pre-stage processing apparatus 2a Multiple animal grouping apparatus 3 Animal related information processing apparatus 4 Post-stage processing apparatus 11 Control part 12 Task transmission part 13 Task result reception part 21 Control part 22 State management part 23 Learning problem management part 24 Results management part 25 Task management unit 26 Storage unit 26a Status DB
26b Learning problem DB
26c Grade DB
26d Task DB
31 control unit 32 state acquisition unit 33 learning / task management unit 34 learning / task presentation unit 35 learning / task buffer unit 41 control unit 42 processing result check unit 43 score calculation unit 44 task result notification unit 45 statistical processing unit 46 transmission destination management Part 47 Task request creation part

Claims (8)

In an animal-related information processing system comprising a client device, a pre-processing device, and an animal-related information processing device,
The control unit of the pre-processing device is
In response to receiving a task request from the client device, a task is determined based on a rule with reference to a state DB storing an animal state,
An animal-related information processing system that transmits the task to the animal-related information processing apparatus and causes an animal to perform the task. In the animal related information processing system according to claim 1,
The control unit of the pre-processing device is
The learning problem is determined by referring to the learning problem DB,
Transmitting the learning problem to the animal-related information processing apparatus to cause the animal to answer the learning problem,
The result of the answer determined by the animal-related information processing apparatus is received and stored in the results DB as a learning result,
The control unit of the pre-processing device is
An animal-related information processing system, wherein the task is determined with further reference to the results DB storing the learning results of animals. In the animal related information processing system according to claim 1 or 2,
The control unit of the pre-processing device is
From the animal-related information processing device, at least one of the type of animal, whether the task can be performed, whether learning is possible, local time of the place where the animal exists, or position information of the place where the animal exists An animal-related information processing system characterized by receiving one of them and storing it in the state DB as state information of the animal. In the animal related information processing system according to any one of claims 1 to 3,
The animal-related information processing system is characterized in that the animal-related information processing apparatus causes the animal to perform tasks prior to learning. In the animal related information processing system according to any one of claims 1 to 4,
The animal-related information processing system further includes a post-processing device and a plurality of animal grouping units,
The control unit of the pre-processing device is
Receiving a plurality of IDs corresponding to a plurality of animal-related information processing devices from the plurality of animal grouping units;
Transmitting the task to the plurality of animal-related information processing devices corresponding to the plurality of IDs to cause the animal to perform the task;
The control unit of the post-processing device is
An animal-related information processing system that receives the processing result of the task from the plurality of animal-related information processing devices, performs statistical processing, and calculates the probability of the processing result. In the animal related information processing system according to any one of claims 1 to 5,
The animal-related information processing system further includes a post-processing device and a processing device that causes at least one of a computer or a human to perform a task and obtains a task result,
The control unit of the pre-processing device further includes
Sending the task to the processing device to cause at least one of a computer or a human to perform the task;
The control unit of the post-processing device is
Receiving the processing result of the task from the animal-related information processing device and the processing device, performing statistical processing, and calculating the probability of the processing result;
An animal-related information processing system characterized by this. The animal-related information processing system according to claim 6,
The control unit of the post-processing device is
The processing result using, as a transmission condition parameter, at least one of the certainty, a history of whether an animal-related information processing device or a processing device has been used, or an elapsed time since the task request was made The animal-related information processing system is characterized in that it determines which of the preceding processing apparatus and the client apparatus is to be transmitted. In an animal related information processing method by an animal related information processing system comprising a client device, a pre-processing device, and an animal related information processing device,
Determining a task based on a rule by referring to a state DB storing an animal state in response to receiving a task request from the client device by the control unit of the pre-stage processing device;
Transmitting the task to the animal-related information processing apparatus and causing the animal to perform the task by the control unit of the pre-processing device.

Images