CN110421574B

CN110421574B - Robot creative action evaluation generation system

Info

Publication number: CN110421574B
Application number: CN201910663756.7A
Authority: CN
Inventors: 丁刚毅; 梅澎; 金乾坤
Original assignee: Beijing Institute of Technology BIT
Current assignee: Beijing Institute of Technology BIT
Priority date: 2019-07-22
Filing date: 2019-07-22
Publication date: 2020-07-03
Anticipated expiration: 2039-07-22
Also published as: CN110421574A

Abstract

The invention provides a robot creative action evaluation generation system which is characterized by comprising a creative generation unit and a creative degree evaluation unit, the creative generation unit is constructed as a complex system including subsystems and includes a machine learning unit such as a deep learning unit, the machine learning unit is configured to train by means of a training set comprising a plurality of first robot actions and to generate a plurality of second robot actions, the creative evaluation unit is configured to evaluate creative degrees of the plurality of second robot actions and to compare creative generation coefficients of the plurality of second robot actions with a predetermined evaluation criterion to select at least part of the second robot actions, wherein a quantum system is defined in a subsystem of the complex system, and the creative generation coefficient is obtained by measuring and/or calculating an entanglement entropy of the quantum system.

Description

Robot creative action evaluation generation system

Technical Field

The invention relates to the field of robots and creative generation, in particular to a system and a method for evaluating and generating creative actions of a robot.

Background

A Humanoid Robot (also called Humanoid Robot), which is a special Robot with human-like appearance, similar motion ability to human and certain intelligence. Since the first humanoid robot Wabot-1 was born, the humanoid robot has become a comprehensive and difficult research direction combining multiple fields of machinery, electronics, computer programming, sensing technology, materials, control technology, artificial intelligence technology and the like, and is an important mark of high-tech strength and development level in China.

The dancing robot is one of a plurality of robots, the existing dancing robot is only like a human body in appearance and generally only has a fixed dancing posture instead of simulating the dancing action of the human body in real time according to the action of the human body.

Therefore, dancing or movements with strong ornamentation, high flexibility or complexity are desired in dancing robots or humanoid robots.

CN105904460B provides a real-time humanoid dance robot control system. The system collects human body dance action information in real time, controls the dance robot steering engine controller to act through data processing of the processor, and makes human body dance actions in real time.

Inventive ideas, concepts, etc. Currently, evaluation of originality of actions or designs of various systems is subjective, and creation of creative actions or designs depends on creativity and ideas of manpower. This results in an inability to evaluate the originality of the system's actions or designs in a relatively objective, automated manner, nor to create creative actions or designs in such a manner. For example, in the real-time humanoid dancing robot control system disclosed in the above-mentioned CN105904460B patent, it is necessary to collect information of dancing movements of the human body in real time and then control the dancing robot to reproduce the dancing movements of the human body in real time.

Disclosure of Invention

It is an object of some embodiments of the present invention herein to provide a robot creative action assessment generation system and method that is capable of evaluating the creativity of generated robot actions and generating robot actions that are more likely to be creative in a relatively objective, automated-suitable manner.

It is an object of some embodiments of the invention herein to provide a system and method for assessing the action originality of a complex system or a system and method for generating creative actions in a complex system that is capable of assessing the action originality of a complex system or generating creative actions in a complex system in a relatively objective, automated suitable manner.

In one aspect, a robot creative action evaluation generation system is provided, including a creative generation unit configured as a complex system including subsystems and including a machine learning unit, such as a deep learning unit, configured to train by means of a training set including a plurality of first robot actions and generate a plurality of second robot actions, and a creative evaluation unit configured to evaluate a creative degree of the plurality of second robot actions and compare a creative degree coefficient of the plurality of second robot actions with a predetermined evaluation criterion to select at least a part of the second robot actions, wherein a quantum system is defined in a subsystem of the complex system, the creative generation coefficient being obtained by measuring and/or calculating an entanglement entropy of the quantum system.

In one embodiment, the evaluation generation system is further configured to intervene in the complex system to cause the quantum system to decoherence.

In one embodiment, the creative generation unit is a computer program implemented creative generation engine, wherein the intervening complex system comprises changing, increasing or decreasing inputs and/or outputs and/or parameters of the computer program implemented creative generation engine.

In one embodiment, the deep learning unit includes a neural network having a plurality of neurons, wherein at least one of the neurons is defined as a quantum of the quantum system.

In one embodiment, the robotic creative action assessment generation system includes or is implemented in a quantum computer having a quantum processor in which the quantum system is defined, such as a boltzmann machine.

In another aspect, a method of evaluating a generate robot creative action is provided, comprising the steps of:

s1: providing a creative generation unit configured as a complex system including subsystems, the creative generation unit including a machine learning unit;

s2: defining a quantum system in a subsystem of the complex system;

s3: training the machine learning unit with a training set comprising a plurality of first robot actions;

s4: generating a plurality of second robot actions by means of a machine learning unit;

s5: evaluating creativity of the plurality of second robot actions and comparing a creative generation coefficient of the plurality of second robot actions with a predetermined evaluation criterion to select at least part of the second robot actions, wherein the creative generation coefficient is obtained by measuring and/or calculating an entanglement entropy of the quantum system.

In one embodiment, the step S3 further includes: intervening the complex system to cause decoherence of the quantum system.

In one embodiment, the deep learning unit comprises a neural network having a plurality of neurons, wherein step S2 comprises: defining at least one of said neurons as said quantum system.

In one embodiment, the robotic creative action assessment generation system comprises a quantum computer having a quantum processor, such as a boltzmann machine or implemented therein, wherein step S2 comprises: defining the quantum system in the quantum processor.

In some embodiments of the invention, a robot learning framework is provided, an intention generation thinking mode is formulated, brain-like behaviors are simulated, and the robot learning framework has certain self-learning, self-evaluation and self-evolution capabilities, so that man-machine cooperative interactive learning can be performed on the basis of self-learning, a data set is expanded and perfected, a system emerging process is realized, and self-creation and creative evaluation of self-creation are completed.

Drawings

One or more embodiments of the invention are described below in conjunction with the following drawings, wherein:

FIG. 1 shows a schematic diagram of a robotic creative action evaluation generation system, according to one embodiment of the present invention;

FIG. 2 shows a schematic diagram of a neural network of a robotic creative action assessment generation system, according to one embodiment of the invention;

FIG. 3 shows a schematic diagram of a robotic creative action evaluation generation system, according to one embodiment of the present invention;

FIG. 4 shows a schematic diagram of a robotic creative action evaluation generation system, according to one embodiment of the present invention;

fig. 5A to 5D show examples of robots that perform actions generated by an evaluation generation system and/or an evaluation generation method according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following detailed description and accompanying drawings. The exemplary embodiments and descriptions of the present invention are provided to explain the present invention, but not to limit the present invention.

In this context, a "complex system" has its basic meaning in the art, e.g. referring to a system with a moderate number of intelligent, adaptive subjects that act on the basis of local information. Therefore, the overall behavior of the complex system has the characteristics of uncontrollable and innovative performance. The concept of complexity (entropy) has been proposed to measure the complexity of complex systems. In this context, a complex system may comprise or may be implemented in a computer program (product).

In this context, "quantum" and "quantum system" include the definition of its standard, so that the standard quantum entanglement phenomenon can occur in the standard quantum system, and the entanglement entropy can be calculated according to the standard physical formula. In reference to the context, "quantum system" in the description herein may also cover quantum-like systems that, although not containing physically meaningful quanta, present a minimal indivisible basic unit that can be seen as a quantum in such quantum systems, and the "(approximate) entanglement entropy" of such quantum systems may be calculated following the formula for calculating entanglement entropy. For simplicity, a standard quantum system or quantum-like system may be indiscriminately described herein as a quantum system, but its meaning may be clear by reference to the context. In this context, a quantum in a quantum-like system may include a neuron in a neural network engine/model.

Referring to FIG. 1, a schematic diagram of a robotic creative action assessment generation system 1 is shown, according to one embodiment of the present invention. The robotic creative action evaluation generation system 1 may include a creative generation unit 11 and a creative degree evaluation unit 12. In some embodiments of the invention, the robotic creative action assessment generation system 1 may be or include a computer-implemented system. For example, the creative generation unit 11 may be a creative generation engine 110 implemented by a computer program. In the illustrated embodiment, the creative generation unit 11, such as the creative generation engine 110, may be configured as a complex system including subsystems.

With continued reference to FIG. 1, the creative generation unit 11, such as creative generation engine 110, may include a machine learning unit 111, such as a deep learning unit. The machine learning unit is configured to train by means of a training set comprising a plurality of first robot actions and to generate a plurality of second robot actions. In one embodiment, the machine learning unit 111, such as a deep learning unit, may comprise a neural network 2 having a plurality of neurons 20, as shown in fig. 2.

In some embodiments, the neural network may include various neural networks including, but not limited to, an RCNN network (regional convolutional neural network), an hed (probabilistic neural Edge Detection network) network, and a GAN network (generative countermeasure network).

In some embodiments, the training and generating of the machine learning unit may include a variety of possible implementations, and may further include a plurality of pre-or post-sub-modules, such as an indexing sub-module for the data set before training and a feedback sub-module after generating the second action. In some embodiments, for example, the RCNN network indexing is used to identify skeletal information of a person or robot in a training set, the HED network is used to generate a limb template, and the GAN network is used to generate continuous robot motion.

With continued reference to fig. 1, the creative evaluation unit 12 is configured to evaluate the creative of the plurality of second robot actions and to compare the creative generation coefficients of the plurality of second robot actions with predetermined evaluation criteria to select at least some of the second robot actions.

In the robotic creative action evaluation generation system 1 of the illustrated embodiment, a quantum system may be defined in a subsystem of the creative generation unit 11, such as the creative generation engine 110, configured as a complex system. In some embodiments, for example, neurons in a neural network are defined as quanta in a "class" quantum system. Thus, in some embodiments, the creative generation coefficients may be obtained by measuring and/or calculating the entanglement entropy of the quantum system.

In some embodiments, the evaluation generation system 1 may be further configured to intervene in the complex system to cause the quantum system to decoherence. For example, in the embodiment shown in FIG. 1, the creative generation unit 11, such as creative generation engine 110, may include a decoherence contribution unit 112 or a subsystem environment contribution unit that contributes to the quantum system to contribute to quantum system decoherence. Although not being bound by theory, the subsystem environment can act on the microscopic quantum system to remove the internal entangled state and generate the phenomenon of decoherence, and the quantum decoherence promotes the quantum behavior of the system to be changed into the classical behavior, and the process is called 'quantum to classical transition'; the process is that the quantum coherence of an open quantum system is gradually lost over time due to quantum entanglement with the external environment; and then in the decoherence process of the quantum system, the interior of the quantum system is actively enabled to be more ordered, the information is completely stored, the entropy value is reduced, and the complexity of the subsystem of the complex system is increased. Therefore, by decoherence of the quantum system, the emergence phenomenon of the complex system can be accelerated, and energy can be released, so that creative actions with creative degrees (probabilities) can be generated with higher efficiency.

In some embodiments, the facilitating quantum decoherence includes, but is not limited to, any means that can facilitate quantum system decoherence, such as automatically, semi-automatically changing, increasing, or decreasing inputs and/or outputs and/or parameters of a complex system or subsystem thereof, such as changing, increasing, or decreasing inputs and/or outputs and/or parameters of the creative generation engine 110. In an exemplary embodiment, the decoherence facilitation unit 112 or subsystem environment action unit may be or include a computer program module that periodically or time-randomly changes a plurality of inputs or parameters of the creative generation engine 110, such as its machine learning unit, such as a neural network, for example, in a random or ordered manner.

In some embodiments, the creative generation unit 11 may also include a graphical user display unit 114 that may, for example, selectively display the generated plurality of second actions and the selected plurality of second actions.

In some embodiments, the creative generation unit 11 may also include one or more functional modules/units. For example, in the embodiment shown in fig. 1, the creative generation unit 11 may further include a logic/planning unit (not shown) that may, for example, determine whether the selected second action corresponds to the action logic of the robot, and/or plan a plurality of second actions to form a coherent third action, such as an overall performance action of the robot 50 (coherent actions are shown in fig. 5A-5D). In some embodiments, the creative generation unit 11 may additionally or alternatively include a degree of freedom constraint unit (not shown) that may provide degree of freedom constraints when the creative generation unit 11, such as the creative generation engine 110, such as its machine learning unit, such as a neural network, generates the second action such that the generated second action may be realizable for the degree of freedom that the robot 50 (shown in fig. 5A-5D) itself may have.

In some embodiments, the robotic creative action evaluation generation system may include creative generation units 11 (and optional evaluation units) arranged in series/parallel such that second actions generated (or selected) by some of the creative generation units 11 may be first actions of additional creative generation units and/or second actions generated and selected by multiple creative generation units as a selected second action group.

Although in the illustrated embodiment the defined quantum system is in a machine learning unit, such as a deep learning unit, such as a neural network, it is conceivable to define the quantum system in other (related) subsystems, modules, units in the creative generation unit 11 as a complex system, such as other functional modules or units described above.

Although the robotic creative action evaluation generation system 1 is described in the form of functional modules in the embodiment shown in FIG. 1, the robotic creative action evaluation generation system 1 may be implemented by means of a computer program as previously described.

Thus, for example, with reference to fig. 3, there is also proposed in some embodiments of the invention a computer device or system 3 comprising one or more processors 31; one or more memories 32; and one or more programs capable of implementing a creative generation function and a creative evaluation function, wherein the one or more programs are stored in the memory and, when executed by the processor, cause the computer device or system to implement the effects of the creative generation unit and the creative evaluation and optionally the other functional modules described above. And, a complex system with subsystems is constructed in the one or more programs and one or more (class) quantum systems can be defined.

In some embodiments, one or more training sets may be stored in the one or more memories, or may be stored in cloud memory, and the computer device or system 3 according to embodiments of the present invention may invoke a training set in cloud memory for training.

In the embodiment illustrated in FIG. 3, the computer device or system 3 may be implemented in a variety of forms including, but not limited to, a personal computer, a laptop computer, a vehicle-mounted human-computer interaction device, a cellular telephone, a personal digital assistant, a media player, a navigation device, an email device, a gaming console, a tablet computer, a graphics processor cluster for artificial intelligence, a supercomputer, and the like.

The processor 31 may comprise one or more processing units. For example, the processor 31 may include an Application Processor (AP), a modem processor, a Graphics Processor (GPU), an Image Signal Processor (ISP), a controller, a video codec, a Digital Signal Processor (DSP), a baseband processor, and/or a neural-Network Processing Unit (NPU), etc. The different processing units may be separate devices or may be integrated into one or more processors.

The memory 32 may include an internal memory, an external memory, or a cloud memory. Memory 32, such as internal memory, may be used to store computer-executable program code, including instructions. The processor 31 executes various functional applications of the computer device or system 3 and data processing by executing instructions stored in the memory 32. The memory may include a program storage area and a data storage area. Wherein, the storage program area can store the application program (such as creative generation engine) required by at least one function, etc. The stored data area may store data (such as training sets, etc.) created during use of the computer device or system 3, and the like. In addition, the internal memory may include a high-speed random access memory, and may further include a nonvolatile memory, such as at least one magnetic disk storage device, a flash memory device, a universal flash memory (UFS), and the like.

Optionally, the computer device or system 3 may comprise a user interface 33.

Referring to FIG. 4, a schematic diagram of a robotic creative action assessment generation system is shown, according to one embodiment of the present invention. In the embodiment shown in FIG. 4, the robotic creative action assessment generation system includes or is implemented in a quantum computer having quantum processor 40. The quantum computer is, for example, a boltzmann machine or is implemented in the quantum computer. Thus, the quantum system may be a "true" quantum system and may be defined in the quantum processor. In a preferred embodiment, a trigger or a controller of the quantum computer may also be provided to actively cause decoherence in the quantum processor or quantum system. In the illustrated embodiment, the quantum computer may be a hybrid computing system including a digital computer 42 coupled to an analog computer 41. In some embodiments, analog computer 41 is a quantum computing unit. The exemplary digital computer 42 includes a digital processor (CPU)420 operable to perform classical digital processing tasks. The digital computer 42 may include at least one digital processor, such as a central processor unit 420 having one or more cores, at least one system memory 422, and at least one system bus 424 that couples various system components, including the system memory 422, to the central processor unit 420.

The digital processor may be any logic processing unit, such as one or more central processing units ("CPUs"), graphics processing units ("GPUs"), digital signal processors ("DSPs"), application specific integrated circuits ("ASICs"), programmable gate arrays ("FPGAs"), Programmable Logic Controllers (PLCs), and the like.

The digital computer 42 may include a user input/output subsystem 426. In some embodiments, the user input/output subsystem includes one or more user input/output components, such as a display, a mouse, and/or a keyboard.

The hybrid computing system may include an appropriate interface or controller 44 coupled to the system bus. The controller is coupled to the digital computer 42 and the quantum computation unit 41.

In some embodiments, the processes of the robotic creative action assessment generation system may be performed in a digital or analog computer as desired, but other processes may be performed in either or both, except for the quantum system defined in the quantum processor. In some embodiments, the robotic creative action evaluation generation system may include or be implemented by a single quantum computer.

In some embodiments of the invention, a method of evaluating a generate robotic creative action is provided, comprising the steps of:

s1: providing a creative generation unit configured as a complex system including subsystems, the creative generation unit including a machine learning unit, such as a deep learning unit;

s2: defining a quantum system in a subsystem of the complex system;

s5: evaluating the creative degree of the plurality of second robot actions and comparing the creative generation coefficients of the plurality of second robot actions with a predetermined evaluation criterion to select at least part of the second robot actions,

wherein the creative generation coefficient is obtained by measuring and/or calculating the entanglement entropy of the quantum system.

In some embodiments, the step S3 further includes: intervening the complex system to cause decoherence of the quantum system.

In some embodiments, the creative generation unit is a computer program implemented creative generation engine, wherein the intervening complex system comprises changing, increasing or decreasing inputs and/or outputs and/or parameters of the computer program implemented creative generation engine.

In some embodiments, the deep learning unit comprises a neural network having a plurality of neurons, wherein step S2 comprises: defining at least one of said neurons as said quantum system.

In some embodiments, the robotic creative action assessment generation system comprises or is implemented in a quantum computer having a quantum processor, such as a boltzmann machine, wherein step S2 comprises: defining the quantum system in the quantum processor.

The information about the entanglement and decoherence of the quantum system is described below by way of explanation and not limitation.

Entangled state and disassociation

The representation of the quantum entanglement state can be represented by a dirac mark as

Wherein | ↓ > | |) indicates the spin of the example is upward spin or downward spin, respectively.

Quantum entanglement can also be expressed mathematically, assuming a composite system is composed of two subsystems A, B, the Hilbert spaces of the two subsystems A, B are respectively H_A，H_BHilbert space H of the composite system_ABIs tensor product

Setting the quantum state of the subsystem A, B to | α>_A′|β>_BIf the quantum state | ψ of the composite system>_ABCan not be written as tensor product

This composite system is referred to as an entangled system of subsystems A, B, and the two subsystems A, B are entangled with each other.

The quantum state of the pure state can be represented by a state vector | ψ>Described, the density operator ρ ═ ψ may also be used><Psi | description statistical mixture states can only be described by density operators, i.e. p ∑_kP_k|ψ_k><ψ_kL, wherein P_kIs a pure state | ψ in a statistical mixed state ρ_k>The probability of occurrence satisfies the normalization condition

In a complex system (cavity field + environment), the initial state is

|ψ(0)>∝(|α>+|-α>)|ε>

Composite system state change at time t

|ψ(t)>∝(|β(t)>|ε₁>+|-β(t))|ε₂>)

Wherein β (t) α e^-γt/2After the cavity field is entangled with the environment, the density operator of the composite system is

ρ＝|ψ(t)><ψ(t)|

Tracing the environment variable, and then the reduction density operator of the coelenterazine is:

although the cavity field is initially in a pure state, i.e. coherent superposition state of coherent state, after the cavity field is acted with the environment, the cavity field will be in a mixed state, i.e. incoherent superposition state of coherent state.

For linear entropy, the trace of the density operator ρ has the following properties:

trp＝1

the definition of entropy is then:

S(ρ)＝-Tr(plnρ)

unfolding to obtain:

S(ρ)＝Tr(ρ-ρ²)

the rate of change of entropy with time is thus:

entanglement and complexity representation

In a stand-alone quantum system, the entanglement entropy of quantum entanglement can be calculated and measured using the principles of dynamics.

Under the chaos state of a macroscopic system, the complex system has emergent characteristics. The whole system is composed of a plurality of subsystems, and the emerging performance is the attribute of the integrated system, and each subsystem does not have the attribute. The subsystem environment acts on the microscopic quantum system to remove the internal entanglement state and generate the phenomenon of decoherence. The new entanglement is established between the quantum system and the subsystem environment of the complex system, so that the internal information of the quantum system is stored more independently, the internal part of the quantum system is more ordered, the information is stored completely, the entropy value is reduced, and the complexity of the subsystem of the complex system is increased. In the process, the inventor uses a system dynamics theory to traverse the entanglement entropy and calculates the quantum system information entropy after decoherence. Similarly, the invention uses entropy to measure the change of the complexity of the complex system, and realizes the information transfer from the micro system to the macro system.

Describing the quantum system is the density matrix ρ in the Hilbert space H, and the subspace H₁Is defined as the entanglement entropy

Namely H₁All other parts are removed, and the von Neumann entropy of the remaining parts is calculated. Therefore, the entanglement entropy is a property of the subspace of the Hilbert space, denoted as S (H)₁). Because of the special nature of quantum systems, the entropy of entanglement is not additive. Can thus define

I(A，B)＝S(A∪B)-S(A)-S(B)+S(A∩B)

I.e. common information, embodying the entanglement between the two subspaces. Common information is a useful information measure in information theory, which can be seen as the amount of information contained in a random variable about another random variable, or the unsuitability of a random variable to be reduced by knowing another random variable.

In some embodiments herein, the emphasis is placed on dynamics, and the complexity of the complex system is evaluated according to the principles of statistics and thermodynamics, thereby leading to the information entropy theory of the complex system.

From the relationship between the microscopic system and the macroscopic system, the increase of the complexity of the macroscopic complex system can be mapped in parallel by the change of the entanglement of the microscopic quantum system. The complexity of the complex system is continuously observed and measured from a pure state along with the change process of time. The complexity of a complex system is a complex paradigm that varies with time t, since the system always evolves towards an unordered state. On the other hand, since the quantum system has a certain degree of entanglement, which is determined by the inherent properties of the microparticle system, the complexity of the complex system does not increase infinitely, and after reaching a certain critical point, the system will show a phenomenon of flooding, releasing energy. This is to receive the creative time and to filter the most effective creative method. Before reaching the critical point, the system has strong and weak emergent performance, and the complexity and the information entropy are used as evaluation criteria of the system by the invention. For the change process of a complex system, the inventor formulates a coefficient which is called a system creative generation coefficient. The coefficient always generates an intention generation process when approaching a certain threshold value, namely, a surge phenomenon occurs. Obviously, the closer the time is to the critical point, the more the system is emerging. However, the inventor does not use the time t as a criterion, but uses the entropy value of the system as a criterion, because the rate v of entropy increase of the system is uncertain, and is limited by the external environment and the property of the system. The determination of the creative generation coefficients is linear with the entropy of the system. In summary, an occurrence is considered to be imminent when the entropy value reaches a certain value, the closer to this critical value the stronger the occurrence of the system itself and vice versa.

Traversal of general entropy

By way of explanation and not limitation, the inventors have realized that entropy acts as a bridge between complex systems and quantum systems, enabling mapping of microscopic systems to macroscopic systems. Through the calculation of the quantum system entanglement entropy, the information quantity of the subsystem is quantitatively judged, so that the complexity of the complex system is estimated by means of the change of the density matrix dimension, and the strength of the system emergence and the creation power can be preliminarily determined.

If the entanglement degree of the quantum system is low, the composite system formed by the subsystems of the quantum system is relatively small in entanglement amount, at the moment, the entropy value of a quantum state is low, the interior of the subsystem tends to an ordered state, the quantum information holding amount is relatively complete, the quantum information loss is small, and the density matrix dimension for describing the system is low. And vice versa.

If the aim of enhancing the system emergence is to a certain extent, the overall characteristics of the complex system can be controlled by reducing the entanglement degree of the quantum system. And the increase of the appearance of a complex system is realized by utilizing the quantum decoherence process.

Shannon proposed the definition of information entropy as "event X in set X_iAmount of information I (x) provided at the time of occurrence_i) Mathematical expectation of ":

however, although quantum chaos can lead to an increase in the degree of entanglement, i.e., an increase in quantum entanglement entropy, the increasing tendency is suppressed by the quantum effect. That is, the entanglement entropy cannot be infinite. In other words, the complex system cannot lose the original emergent characteristic by increasing the entanglement of the subsystems. Meanwhile, if the system has strong emergence and an emergence process occurs, the system is gradually changed from an unordered state to an ordered state, the entropy of the system is reduced, and the system emergence amount can be represented by the difference of the entropy from the beginning of the system to the end of the system emergence:

ΔH＝H_start-H_end

basis of kinetic traversal in isolated quantum system

Wherein J_yAnd J_zIs the angular momentum operator. Entanglement can be characterized by an entanglement entropy S

S＝-Trρ_sqlog₂(ρ_sq)

Where ρ is_sqIs a density matrix of single quantum bits.

The dynamic traversal process of the isolated quantum system is to determine the entanglement state condition of the microscopic system, evaluate the entanglement amount and the information amount of the system, and use the estimated entanglement amount and the information amount as a first reference value for expressing the surging evaluation of the macroscopic complex system.

By way of explanation and not limitation, the inventors have established a relationship between entropy values and complex system salience. The calculation of the entropy value is divided into a complexity entropy of a complex system and an entanglement entropy of a quantum system, and the two parts are in a negative correlation relationship, wherein the former is used for comparison and evaluation, and the latter is used for calculation and measurement. The results of the two reflect the creative capability of the system, namely the strength of the occurrence of the system.

It is a probabilistic problem for the amount of information. The probability is that only one bit needs to be transmitted, i.e. the smaller the probability the more bits need to be transmitted. If the information quantity is measured by the number of bits, the lower the probability is, the larger the information quantity is. Based on which a formula is given for the amount of information

I(x)＝-log₂p(x)

That is, if the probability of the information source is larger, the amount of information carried by the information source is smaller, and the transmission and storage cost is smaller; the smaller the probability of the information source is, the larger the information carried by the information source is, and the larger the transmission and storage cost is.

From the above section, the expectation of the information amount in the information entropy is needed, so that when the information amount of the complex system is researched, the complexity initial state value of the pure state of the complex system needs to be determined, the information entropy of the system is calculated on the basis of the complexity initial state value, the information amount of the system is obtained, and the occurrence of the complex system is judged.

By way of explanation and not limitation, quantum decoherence promotes the transition of the quantum behavior of the system to classical behavior, a process known as "quantum-to-classical transition". The process is that the quantum coherence of an open quantum system is gradually lost over time due to quantum entanglement with the external environment. The quantum decoherence process occurs during the action of the subsystems of a complex system on an isolated quantum system. The information of the isolated system cannot be lost, but once the information is acted by the subsystem environment and coupled with the subsystem environment, the original entanglement relation is broken, and new entanglement is generated. The entanglement entropy inside the quantum system is reduced, the complexity of the complex system is increased, and the entropy is increased. The information entropy is reflected on the information entropy, the information entropy is increased, and the currency is enhanced. Thus, in some instances herein, the currency and creative evaluation coefficients may be used instead.

In some embodiments of the invention, dance performance segments are classified and integrated through deep learning, classification features are extracted, a feature recognition model is constructed, and logic interaction is performed in a simulation engine and iteration is performed continuously. On the basis of large-scale learning, primary reproduction is realized, secondary iteration is performed, the process of entropy increase is completed, system phase change is realized, the effect of creative generation is achieved, and performance data depicting is completed by combining scene migration and scene generation. The data is used as a rehearsal guide, and the robot can be driven to complete dance demonstration and the like.

While a robotic creative action evaluation generation system is described in some embodiments, it is contemplated that some embodiments of the invention also present a generation, evaluation, and/or guidance system that may be used for dance or other learned actions of a person, such as sports actions, that falls within the scope of the invention.

As will be appreciated by one skilled in the art, embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects.

Unless specifically stated otherwise, the actions or steps of a method, program or process described in accordance with an embodiment of the present invention need not be performed in a particular order and still achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

While various embodiments of the invention have been described herein, the description of the various embodiments is not intended to be exhaustive or to limit the invention to the precise forms disclosed, and features and components that are the same or similar to one another may be omitted for clarity and conciseness. As used herein, "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" are intended to apply to at least one embodiment or example, but not to all embodiments, in accordance with the present invention. And the above terms are not necessarily meant to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics of the various embodiments may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.

As used herein, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exhaustive, such that a process, method, article, or apparatus that comprises a list of elements may include those elements but do not exclude the presence of other elements not expressly listed. For purposes of this disclosure and unless specifically stated otherwise, "a" means "one or more". To the extent that the term "includes" or "including" is used in this specification and the claims, it is intended to be inclusive in a manner similar to the term "comprising" as that term is interpreted when employed as a transitional word. Furthermore, to the extent that the term "or" is used (e.g., a or B), it will mean "a or B or both". When applicants intend to indicate "only a or B but not both," only a or B but not both will be used. Thus, use of the term "or" is inclusive and not exclusive. See bryan.a. garner's "dictionary of modern law terminology" page 624 (2 d.ed.1995).

Exemplary systems and methods of the present invention have been particularly shown and described with reference to the foregoing embodiments, which are merely illustrative of the best modes for carrying out the systems and methods. It will be appreciated by those skilled in the art that various changes in the embodiments of the systems and methods described herein may be made in practicing the systems and/or methods without departing from the spirit and scope of the invention as defined in the appended claims. It is intended that the following claims define the scope of the system and method and that the system and method within the scope of these claims and their equivalents be covered thereby. The above description of the present system and method should be understood to include all new and non-obvious combinations of elements described herein, and claims may be presented in this or a later application to any new and non-obvious combination of elements. Moreover, the foregoing embodiments are illustrative, and no single feature or element is essential to all possible combinations that may be claimed in this or a later application.

Claims

1. A dance robot creative action evaluation generation system, characterized by comprising a creative generation unit and a creative evaluation unit, wherein the creative generation unit is configured as a complex system comprising subsystems and comprises a deep learning unit, the deep learning unit is configured to train by means of a training set comprising a plurality of first robot dance actions and generate a plurality of second robot dance actions based on the training set, the creative evaluation unit is configured to evaluate the creative of the plurality of second robot dance actions and compare creative generation coefficients based on the plurality of second robot dance actions with a predetermined evaluation standard to select at least part of the second robot dance actions, wherein a quantum system is defined in a subsystem of the complex system, the creative generation coefficients are obtained by measuring and/or calculating an entanglement entropy of the quantum system, wherein the deep learning unit includes a neural network having a plurality of neurons defined as quanta of the quantum system.

2. The dance robot creative action assessment generation system of claim 1, wherein the assessment generation system is further configured to intervene in the complex system to facilitate the quantum system decoherence.

3. The dance robot creative action assessment generation system of claim 2, wherein the creative generation unit is a computer program implemented creative generation engine, wherein the intervention of the complex system includes changing, increasing, or decreasing inputs and/or outputs and/or parameters of the computer program implemented creative generation engine.

4. A method for evaluating an action of generating a dance robot creative, comprising the steps of:

s1: providing a creative generation unit configured as a complex system including subsystems, the creative generation unit including a deep learning unit;

s2: defining a quantum system in a subsystem of the complex system;

s3: training the deep learning unit with a training set comprising a plurality of first robot dance motions;

s4: generating a plurality of second robotic dance movements by means of the trained deep learning unit;

s5: evaluating the creativity of the plurality of second robot dance actions and comparing the creativity generation coefficients of the plurality of second robot dance actions with a predetermined evaluation criterion to select at least part of the second robot dance actions,

wherein the creative generation coefficients are obtained by measuring and/or calculating an entanglement entropy of the quantum system,

wherein the deep learning unit includes a neural network having a plurality of neurons,

wherein, step S2 includes: defining the neuron as a quantum of the quantum system.

5. The method of evaluating a dancing robot creative action of claim 4, wherein the step S3 further comprises: intervening the complex system to cause decoherence of the quantum system.

6. The method of evaluating a generate dance robot creative action of claim 5, wherein the creative generation unit is a computer program implemented creative generation engine, wherein the intervening the complex system includes changing, increasing, or decreasing inputs and/or outputs and/or parameters of the computer program implemented creative generation engine.