Time-varying psycho-perceptual attention modeling method and system

Abstract

Human interaction with products and services can be difficult to measure or quantify. As an example solution, a computer-implemented method is provided. The method includes receiving human perception data in response to an external stimulus over a period of time. The human perception data includes a plurality of time periods. The method further includes determining an emotional response based on the human perception data. The method further includes assigning a numerical value to the emotional response. The method further includes calculating a change in the emotional response over the plurality of time periods over the period of time.

Description

[0001] Cross-references to related applications This application claims priority to U.S. Patent Application No. 63 / 747,199, filed January 20, 2025, entitled “Method and System for Modeling Time-Varying Mental Perception Attention,” the entire contents of which are incorporated herein by reference. Technical Field

[0002] The embodiments shown relate to methods and systems for modeling time-varying psychological attention to determine successful human interactions. Background Technology

[0003] Many industries rely on understanding how consumers or users will interact with their products or services to determine their viability. For this, industries may depend on multiple product and user data streams. In some cases, businesses may modify their products and services, or market them differently, based on anticipated perceptions, to improve them. Similar processes may occur in workplace management or administration. Processing data to draw definitive conclusions about the merits or demerits of products, services, or dynamics can be quite challenging.

[0004] Human interactions with products and services can be difficult to measure or quantify for a variety of reasons. For example, different products and services may be designed to offer different value propositions or target specific demographics. Furthermore, an individual's interaction with a product or service may vary based on their specific individual characteristics and may depend on their expectations of the product or service and its form. Additionally, data sources are often dispersed and difficult to process and manage in silos. Consequently, many existing processes can be time-consuming, costly, and fail to yield definitive conclusions.

[0005] Therefore, an improved approach and system is needed to quantify human interaction with products or services and determine their success potential. Summary of the Invention

[0006] According to a first aspect, a computer-implemented method is provided, comprising: receiving human perception data in response to external stimuli over a period of time, the human perception data including multiple time periods; determining an emotional response based on the human perception data; assigning a numerical value to the emotional response; and calculating the change in the emotional response within each of the multiple time periods over the period.

[0007] In some implementations, the external stimulus can be a media form, including at least one of advertising, video, film, text, image, or audio. The method may further include: predicting the financial success of the media based on at least one of emotional response or changes in emotional response.

[0008] In some implementations, human perception data may include data describing at least one of vision, hearing, touch, taste, or smell.

[0009] In some implementations, human perception data can be collected through biometric sensors or manual observation.

[0010] In some implementations, the desired emotional response can be an S-shaped curve, wherein the emotional response is low at the beginning of the time period and rises to a higher value at the end of the time period.

[0011] In some implementations, human perception data can be generated by an artificial intelligence agent.

[0012] In some implementations, an artificial intelligence agent can be trained to simulate a user with specific preferences.

[0013] In some implementations, the method may further include: calculating the derivative of the change in emotional response over the time period; and determining a success indicator related to external stimuli by comparing the derivative of the change in emotional response over the time period and the change in emotional response over the time period with the expected emotional response.

[0014] In some implementations, the method may further include generating suggestions for adjusting the external stimulus based on success metrics associated with the external stimulus.

[0015] According to another aspect, a system is provided, comprising a user interface component and a server component, the server component being configured to: receive human perception data in response to external stimuli over a period of time, the human perception data including multiple time periods; determine an emotional response based on the human perception data; assign a numerical value to the emotional response; and calculate the change in the emotional response within each of the multiple time periods over the period.

[0016] In some implementations, the external stimulus may be a media form, including at least one of advertising, video, film, text, image, or audio, and the system is further configured to predict the financial success of the media based on at least one of emotional response or change in emotional response.

[0017] In some implementations, human perception data may include data describing at least one of vision, hearing, touch, taste, or smell.

[0018] In some implementations, human perception data can be collected through biometric sensors or manual observation.

[0019] In some implementations, the desired emotional response can be an S-shaped curve, wherein the emotional response is low at the beginning of the time period and rises to a higher value at the end of the time period.

[0020] In some implementations, human perception data can be generated by an artificial intelligence agent.

[0021] In some implementations, an artificial intelligence agent can be trained to simulate a user with specific preferences.

[0022] In some implementations, the server component may be further configured to: calculate the derivative of the change in emotional response over the time period; and, based on the derivative of the change in emotional response over the time period and the change in emotional response over that time period, determine a success metric related to the external stimulus by comparing it with the expected emotional response.

[0023] In some implementations, the server may be further configured to generate recommendations for external stimuli based on success metrics associated with those stimuli.

[0024] According to another aspect, one or more non-volatile computer-readable media are provided, having stored executable instructions that, when run by at least one computer, cause the at least one computer to perform a method comprising: receiving human perception data in response to external stimuli over a period of time, the human perception data including multiple time periods; determining an emotional response based on the human perception data; assigning a numerical value to the emotional response; and calculating changes in the emotional response within each of the multiple time periods over the period.

[0025] In some implementations, the external stimulus may be a form of media, including at least one of advertising, video, film, text, image, or audio, and the method further includes: predicting the financial success of the media based on at least one of emotional response or changes in emotional response.

[0026] According to another aspect, a computer-implemented method for determining successful human interaction is provided, the method comprising: receiving human perception data in response to external stimuli over a period of time, said period of time including more than one time period; determining an emotional response based on the human perception data; assigning a numerical value to the emotional response; and calculating the change in the emotional response within each said time period.

[0027] In some implementations, the method may further include: determining a success metric based on changes in emotional response over a period of time, by comparing it with the expected emotional response.

[0028] In some implementations, the method may further include generating suggestions for adjusting external stimuli based on success metrics.

[0029] In some implementations, the method may further include: adjusting the external stimulus based on the suggestion.

[0030] In some implementations, human perception data may include multiple datasets relating to at least one of vision, hearing, touch, taste, or smell.

[0031] In some implementations, the various method steps can be performed for each of the plurality of datasets.

[0032] In some implementations, the method may further include: calculating the derivative of the change in emotional response over the time period; and determining a success index based on the derivative of the change in emotional response over the time period and the change in emotional response over the time period, by comparing it with the expected emotional response.

[0033] In some implementations, human perception data can be collected through manual observation.

[0034] In some implementations, human perception data can be collected using biometric sensors.

[0035] In some implementations, human perception data can be collected by an artificial intelligence agent.

[0036] In some implementations, human perception data can be generated using an artificial intelligence engine.

[0037] In some implementations, the method may further include: predicting emotional responses to external stimuli using an artificial intelligence agent.

[0038] In some implementations, an artificial intelligence agent can be trained to simulate a user with specific preferences.

[0039] In some implementations, human perception data can be generated based on users with specific preferences.

[0040] In some implementations, the prediction of emotional responses may further include: an artificial intelligence agent simulating a user with a specific tendency.

[0041] In some implementations, a particular preference may be at least one of age or gender.

[0042] In some implementations, external stimuli can be a form of mediation.

[0043] In some implementations, the medium may be advertising.

[0044] In some implementations, the medium may be at least one of video or film.

[0045] In some implementations, human perception data can be generated by an artificial intelligence engine through user engagement analysis, video content analysis, or natural language processing.

[0046] In some implementations, the desired emotional response can be an S-shaped curve, wherein the emotional response is low at the beginning of the time period and rises to a higher value at the end of the time period.

[0047] In some implementations, artificial intelligence engines can be used to predict the financial success of a medium.

[0048] In some implementations, the external stimulus can be an experience.

[0049] In some implementations, the experience can be the process of driving a vehicle.

[0050] In another aspect, a system for determining successful human interaction is provided, the system including a user interface component and a server component, the server component being configured to: receive human perception data in response to external stimuli over a period of time, said period of time including more than one time interval; determine an emotional response based on the human perception data; assign a numerical value to the emotional response; and calculate the change in the emotional response within each time interval of the period.

[0051] In some implementations, the server may be further configured to determine a success metric by comparing changes in emotional response over a period of time with the desired emotional response.

[0052] In some implementations, the server can be further configured to generate recommendations for external stimuli based on success metrics.

[0053] In some implementations, the server may be further configured to transmit suggestions to the user interface components.

[0054] In some implementations, human perception data may include multiple datasets relating to at least one of vision, hearing, touch, taste, or smell.

[0055] In some implementations, the server may be further configured to: calculate the derivative of the change in emotional response over the time period; and, based on the derivative of the change in emotional response over the time period and the change in emotional response over this time period, determine a success metric by comparing it with the desired emotional response.

[0056] In some implementations, the system may further include biometric sensors for collecting human sensory data.

[0057] In some implementations, the system may further include an artificial intelligence agent.

[0058] In some implementations, the AI ​​agent can be further configured to collect human perception data.

[0059] In some implementations, the system may further include an artificial intelligence engine.

[0060] In some implementations, the artificial intelligence engine can be further configured to generate human-perceived data.

[0061] In some implementations, the AI ​​agent can be further configured to predict emotional responses to external stimuli.

[0062] In some implementations, the AI ​​agent can be further configured to simulate a user with specific preferences.

[0063] In some implementations, the artificial intelligence engine can be further configured to generate human-perceived data based on users with specific preferences.

[0064] In some implementations, the AI ​​agent can be further configured to predict emotional responses to external stimuli by simulating users with specific tendencies.

[0065] In some implementations, external stimuli can be a form of mediation.

[0066] In some implementations, the external stimulus can be advertising.

[0067] In some implementations, the medium may be at least one of video or film.

[0068] In some implementations, the artificial intelligence engine can be configured to generate human-perceived data through user engagement analysis, video content analysis, or natural language processing.

[0069] In some implementations, external stimuli may include experiences.

[0070] In some implementations, the experience can be the process of driving a vehicle. Attached Figure Description

[0071] Several embodiments will now be illustrated by way of example with reference to the accompanying drawings, wherein: Figure 1 This is a block diagram of a system integrating an analytics engine according to some embodiments; Figure 2 This is a block diagram of an exemplary artificial intelligence agent; Figure 3 It is based on some implementation methods Figure 1 A schematic diagram of the analysis engine is shown; Figures 4A-4B It is based on other implementation methods. Figure 1 Another schematic diagram of the analysis engine shown; Figure 5This illustrates an exemplary integrated AI agent employing an analytics system; Figure 6 This illustrates an exemplary hierarchical structure of an AI system; Figure 7 This shows various types of dopamine levels; Figure 8 This section shows an example of the classification of expected violation factors; Figure 9 It is an integration based on some embodiments. Figure 1 The diagram shows the block diagram of the analysis system of the analysis engine. Figure 10 yes Figure 9 A block diagram of external stimuli in the system shown. Figure 11 yes Figure 9 A block diagram of the output of the system shown; Figures 12A-12C Examples of perception, energy levels, and effectors are shown; Figure 13 yes Figure 9 A block diagram of the analysis engine in the system shown. Figure 14 yes Figure 9 A block diagram of the sensing data in the system shown; Figures 15A-15D Examples of dopamine and serotonin concentrations are shown; Figures 16A-16C An example of human perception bandwidth is shown; Figure 17 yes Figure 14 The diagram shows the process of acquiring sensor data. Figure 18 An example of sensory data collected by a biometric sensor is shown; Figure 19 Collected from real audiences Figure 14 The diagram shows the sensing data. Figure 20 Examples of measured sensory data are shown, categorized according to various external stimuli and energy levels. Figure 21 It is a simulation Figure 14 The diagram shows the sensing data. Figure 22 Based on external stimulus simulation Figure 14 The diagram shows the sensing data. Figure 23 From Figure 14 The diagram shown illustrates the process of deriving emotional responses from perceived data. Figure 24 It includes multiple time periods. Figure 14 The diagram shows the sensing data. Figure 25 Based on Figure 23 The diagram shown illustrates how emotional responses can predict financial success. Figure 26A -B shows examples of evaluating the financial success of various films; Figure 27 Examples of expected emotional responses are shown; Figure 28 Examples of unintended emotional reactions are shown; Figure 29 Through integration Figure 9 The diagram shown illustrates how the analytics engine predicts financial success. Figure 30 Utilization according to some embodiments is shown Figure 9 The system shown is used to determine changes in emotional responses. Figure 31 Utilization according to some embodiments is shown Figure 25 The method shown indicates that the system has successfully determined its operation. Figure 32A -B indicates the use of Figure 9 An example of a movie analysis engine for the analysis system shown; Figure 33 This is a block diagram of a computing device according to one embodiment. Detailed Implementation

[0072] Analysis Engine Figure 1 The system 100 is shown, which includes user equipment 102, analysis engine 104, and environment 106.

[0073] User equipment 102 may be a computing device, such as a mobile device, personal computer, server, embedded system, or other device with computing capabilities. User equipment 102 may receive input from a user (e.g., a human) or other computing devices (e.g., one or more sensors, instruments, and / or information systems).

[0074] User device 102 communicates with analytics engine 104, for example, via a network (not shown). User device 102 and analytics engine 104 can exchange information with each other, allowing user device 102 to both send and receive information from analytics engine 104. The network may include the Internet, intranet, WiFi network, Bluetooth network, iBeacon network, or other communication protocols that allow user device 102 and analytics engine 104 to exchange information.

[0075] In some implementations, the analytics engine 104 may run, be hosted, and / or stored on one or more servers or other computing devices. In these implementations, cloud computing may be used to allow user device 102 to communicate with the analytics engine 104.

[0076] In some implementations, analytics engine 104, or a portion thereof, may run, be hosted, and / or stored on user device 102. In these implementations, edge computing or cloud-edge collaborative computing may be used to allow user device 102 to communicate with analytics engine 104. In implementations where only a portion of analytics engine 104 runs, is hosted, and / or stored on user device 102, the analytics engine 104 contained on user device 102 may communicate with that portion of analytics engine 104 that runs, is hosted, and / or stored on a server or other external computing device.

[0077] The analytics engine 104 can also communicate with the environment 106, for example, via a network (not shown). Examples of networks may include the Internet, an intranet, a WiFi network, a Bluetooth network, an iBeacon network, or other communication protocols. The analytics engine 104 can query the environment 106, send data to the environment 106, retrieve data from the environment 106, and respond to queries from the environment 106. The analytics engine 104 and the environment 106 can communicate bidirectionally, similar to the relationship between the user device 102 and the analytics engine 104. As will be further detailed below, the environment 106 may include databases, the Internet (such as websites), application-specific interfaces (APIs), computing devices, sensors, intranets, internal systems, etc.

[0078] The analytics engine 104 can be used to process data, generate analytical insights based on the data, respond to queries received from the input device 102, automate tasks and processes, and / or solve problems.

[0079] In some implementations, analytics engine 104 can provide a unified application that embodies artificial intelligence in the narrow sense (ANI), artificial intelligence general (AGI), and artificial intelligence superintelligence (ASI). Analytics engine 104 enables organizations or users of user devices 102 to design artificial intelligence (AI) solutions and automate tasks.

[0080] ANI can also be called "weak AI." ANI can refer to AI that focuses on a specific task or a narrow range of tasks. The key characteristics of ANI are that it is task-specific, limited in scope, and lacks true understanding. However, in some examples, ANI can be more efficient than humans.

[0081] AGI can understand, learn, and perform any intelligent task that humans can accomplish.

[0082] ASI surpasses human capabilities across various domains, thus going a step further than AGI. ASI can create self-driven purposes and / or goals.

[0083] AGI and ASI are capable of solving a wide variety of problems across multiple domains, demonstrating versatility and adaptability. Some of the problems that AGI and ASI can solve include complex decision-making, cross-domain multitasking, autonomous innovation, improving efficiency and productivity, enhancing personalization, and addressing global challenges.

[0084] Currently, AGI and ASI may still be theoretical concepts that are difficult to achieve. While ANI exists, its development is often customized, cumbersome, and time-consuming. System 100 and Analytics Engine 104 offer a streamlined solution that can efficiently address and resolve any problem across the entire AI capability spectrum.

[0085] For example, existing technologies can include very different ANI solutions, such as demand prediction models, image classification models, etc. These solutions may often be limited to specific, narrow applications. In addition, generative AI solutions (such as GPT-4, Gemini 1.5, and Llama 3) are available, but cannot be integrated into a unified system that encompasses ANI, AGI, and ASI.

[0086] In some implementations, the analysis engine 104 may also include a group of AI agents. In this document, an AI agent may include at least one AI model, such as a large language model (LLM) or a multimodal large language model (MLLM). An AI agent may also include one or more other models or tools to allow the AI ​​agent to perform one or more tasks. Additionally, the terms "AI agent" and "agent" may be used herein to refer to a single AI agent or to one or more AI agents, such as a group of AI agents that can collaboratively solve a problem.

[0087] Figure 2 An exemplary AI agent is illustrated, which may include LLM or MLLM, knowledge and memory, and tools. The AI ​​agent can receive input. In some examples, the AI ​​agent may also include system prompts. The AI ​​agent can generate output actions based on the input, LLM / MLLM, knowledge and memory, and / or system prompts.

[0088] In some examples, an AI agent can be a computational entity designed to perform tasks by perceiving input, processing information, and executing actions to achieve a specific purpose. At its core, the AI ​​agent may include an LLM or MLLM, which acts as the brain of the AI ​​agent, enabling understanding of input data and its processing, contextual reasoning, and decision-making. The AI ​​agent can be equipped with tools, such as task-specific APIs, plugins, or computational modules, that extend its capabilities beyond language processing to include data retrieval, numerical analysis, and / or automated workflows. The AI ​​agent can receive input through various connections, including natural language commands, structured data (such as tables or databases), sensed data (such as audio, video, or environmental metrics), and external APIs for real-time information. These inputs can be preprocessed in the processing layer to ensure context-aware decision-making. The AI ​​agent can produce output actions ranging from generating natural language responses to performing tasks via APIs, controlling physical devices, and / or providing data insights and visualizations. To achieve continuous improvement, the AI ​​agent can integrate feedback loops and learning mechanisms, leveraging user feedback, recorded interactions, and / or reinforcement learning to optimize its performance over time.

[0089] As briefly described above, the implementation of system 100 can include cloud computing, edge computing, and cloud-edge collaborative computing. For example, in some implementations of cloud computing, edge computing, and cloud-edge collaborative computing, predictive AI, generative AI, and agent frameworks consisting of a group of AI agents can be employed. These implementations can be used to implement ANI / AGI / ASI.

[0090] Figure 3 A schematic diagram of an analysis engine 104 according to some implementations is shown.

[0091] The analysis engine 104 includes input 110, which can be received from user device 102 and / or environment 106. Input 110 may include information, data, queries, prompts, unresolved problems, and / or other inputs. Input 110 can exist in the form of text, images, video, streams, documents, and / or other data formats.

[0092] In some implementations, input 110 (which may include data and / or one or more prompts) may be passed to each step of the analytics engine 104, as discussed below. Passing input 110 to each step allows each step in the analytics engine 104 to determine whether a portion of input 110 is relevant to the current step, and this can reduce the response time of the analytics engine 104 to input 110. The steps in the analytics engine 104 may include AI agents and / or LLMs within the analytics engine 104, which will be discussed in further detail below.

[0093] The analysis engine 104 may also include an additional information loop 112. At the additional information loop 112, the analysis engine 104 can determine whether the input 110 is sufficient for the analysis engine 104 to provide a result, solution, or answer, such as a response to a query or a solution to a problem. If the input 110 is insufficient, for example, if more information or data is needed, the additional information loop 112 can request more information, data, and / or other input at the input 110.

[0094] The analytics engine 104 may also include an AI agent 114. In some implementations, the AI ​​agent 114 can distinguish between the organization's, enterprise's, or other entity's internal pressing needs (i.e., hunger) and long-term goals (i.e., vision). For example, the AI ​​agent 114 can identify attributes of senior management within the organization / company and attributes of the company itself.

[0095] The analysis engine 104 may include a decision point 116 that determines whether the input 110 involves future predictions or historical queries. If the input 110 involves future predictions, the decision point 116 may connect to the future prediction module 118. Conversely, if the input 110 involves historical queries, the decision point 116 may connect to the historical query module 120. In some examples, such as when the input 110 involves both future predictions and historical queries, the decision point 116 may connect to both the future prediction module 118 and the historical query module 120.

[0096] The future prediction module 118 can evaluate whether the problem explicitly specified in input 110 is a new problem or a known problem. Based on the problem type determination, a pre-trained model can be employed, which can be one or more generative AI models. Specifically, if the problem is a known problem, the problem type can be identified and one or more historical models, pre-trained models, or custom-trained models can be employed. If the problem is a new problem, predictive and / or generative AI models can be used to generate or select one or more pre-trained models or historical models. In some examples, input 110 may contain multiple problems, and / or a single problem may require multiple models; therefore, the outputs of multiple models can be integrated into a solution list at the future prediction module 118.

[0097] The historical query module 120 may include a Retrieval Enhanced Generation (RAG) model and / or a data vault. The historical query module 120 may also include resources for Enterprise Resource Planning (ERP), including Customer Relationship Management (CRM), Material Requirements Planning (MRP), and financial resources.

[0098] The analysis engine 104 may also include a comparator 122. At comparator 122, one or more output solutions from the future prediction module 118 and / or the past history query module 120 can be evaluated to determine whether this or these solutions provide an acceptable or complete answer to the question or query contained in input 110. If this or these solutions are unacceptable or incomplete, comparator 122 can loop back to an earlier stage within the analysis engine 104 to repeat or optimize the solution generation process, for example by requesting more data or information at input 110. If this or these solutions are acceptable or complete, comparator 122 can continue.

[0099] As a prerequisite for execution, the analysis engine 104 may also include a planner 124, which may include a solution planner or a project manager. The planner 124 may break down the solution into multiple smaller steps as needed before execution. The planner 124 may include solutions from the comparator 122.

[0100] The analysis engine 104 may also include an executor 126, which can perform actions based on the planner 124. Actions may include computer actions, such as sending emails, performing or coordinating sales, robotic process automation (RPA), etc.

[0101] As described above, analytics engine 104 can communicate with environment 106. In some examples, environment 106 may include company infrastructure, systems, computing devices, resellers, websites, etc. Actuator 126 can perform actions on environment 106. Analytics engine 104 can also receive feedback from external feedback mechanism 128, which may be connected to company or organizational infrastructure, such as within environment 106. Feedback may include response feedback, new requests from customers (e.g., customers of the company or organization or the organization itself), or other forms of feedback. This feedback can be fed back to input 110, which can be used in another process loop of analytics engine 104 or considered for further processing.

[0102] In addition, motion feedback 130 can be generated by actuator 126 for analysis engine 104 so that it can be fed into input 110 in subsequent process loops or considered for further processing.

[0103] It should be understood that other implementations and examples of the analysis engine 104 are also possible. Some or all of the modules or stages discussed above in the analysis engine 104 may be rearranged, removed or replaced, and new modules or other modules not yet discussed may also be incorporated into the analysis engine 104.

[0104] The analytics engine 104 can integrate predictive AI, generative AI, and agent-based AI workflows. As discussed further below, the analytics engine 104 can employ a group of AI agents in conjunction with a telephone application (e.g., user device 102) for data input. The computations can then be performed in the cloud, at edge devices, and / or a combination of both.

[0105] Data Integration and Sources: Analytics engine 104 can connect AI agents to various data sources, such as ERP, CRM, and financial systems (e.g., within environment 106). Analytics engine 104 facilitates seamless data flow and AI configuration. Data can also be collected from noise, vibration, harshness (NVH), GPS, voice, and visual sensors embedded in telephones (e.g., providing input to analytics engine 104, such as user device 102). This data can be used to train custom models or for real-time inference to predict future outcomes.

[0106] Overall application functionality: The analytics engine 104 can perform comprehensive analysis by examining historical data to answer questions about past events. Predictions can be generated using pre-trained models and / or custom-trained models, which can be deployed in the cloud or on edge devices.

[0107] Feedback loops and continuous improvement: Feedback loops can be integrated into the analytics engine 104 to address unresolved or partially resolved issues. Even fully resolved issues can remain open until the corresponding response or result is recorded, ensuring continuous improvement and accuracy.

[0108] Figures 4A-4B Analysis engine 104' is shown according to some other embodiments of analysis engine 104. It should be understood that analysis engine 104 and analysis engine 104' are interchangeable within system 100, and all references to analysis engine 104 herein may also refer to analysis engine 104'.

[0109] The analysis engine 104' can receive external input. In some examples, external input may include sensor information and new information. Sensor information may include noise / sound, vibration, severity, and visual (NVH-V) information.

[0110] In a further example, external input may supplementally or alternatively include information describing the company, such as name and industry information, company revenue, number of company employees, company competitors, company customers, and company suppliers.

[0111] As a supplement or alternative, external input may include computational data and / or cue data. Cue data can be parsed by a large language model, such as via an API like the ChatGPT™ API.

[0112] The analytics engine 104' may include external inputs to the decryptor. The decryptor may generate business overviews, internal analyses, external landscapes, AI suggestions, and / or AI opportunities identified by the analytics engine 104'. The decryptor may also store its inputs and outputs in the memory of the analytics engine 104'.

[0113] It should be understood that memory and cognition can be important determinants in decision-making. Memory can be analogous to weights and biases in a pre-trained AI model. Cognition can be analogous to a processor. Memory and cognition can be derived from Human Feedback Reinforcement Learning (RLHF), which can incorporate human cognition and benefit from well-trained models.

[0114] The analytics engine 104' may also include a business creator / generator, which can generate AI workflows, AI value, and / or AI roadmaps. The business creator / generator may also store its inputs and outputs in the memory of the analytics engine 104'.

[0115] The analytics engine 104' may also include an analytical answer generator, which can receive and / or generate CRM, ERP, and documents. The analytical answer generator may also store its inputs and outputs in the memory of the analytics engine 104'.

[0116] The analytics engine 104' may also include an AI / machine learning (ML) predictor, which can receive and / or generate data, models, and software applications. The AI / ML predictor may also store its inputs and outputs in the memory of the analytics engine 104'. In this document, the term "artificial intelligence (AI)" also includes machine learning.

[0117] The analytics engine 104' can also generate AI insights, which may include predictions, analyses, and / or recommendations. These AI insights can also be stored in the memory of the analytics engine 104'.

[0118] The analysis engine 104' may also include an AI trainer and / or be able to perform actions on the output. In this paper, the AI ​​engine may receive and / or generate data, models, and / or other content.

[0119] The effector and external output can be passed in a feedback loop, acting in conjunction with human actions on the output. The feedback loop may include a comparator that compares the output with past memories (such as the memories of analysis engine 104'). The feedback loop can return to the input and be fed into analysis engine 104' as external input. In some examples, the output may be discarded by analysis engine 104'.

[0120] It should be understood that other implementations and examples of the analysis engine 104' are also possible. Some or all of the modules or stages discussed above in the analysis engine 104' can be rearranged, removed or replaced, and new modules or other modules not yet discussed can also be incorporated into the analysis engine 104'.

[0121] Additionally, system 100 may include, or interact with, one or more modules and / or AI agents, for example, as part of another system. Figure 5 As shown, system 100 can be used or communicated with an integrated AI agent, which can communicate with other AI agents. For example, the integrated AI agent can communicate with a strategic planning AI agent, a strategic alignment AI agent, an execution AI agent, a teaching AI agent, and / or a sales model AI agent.

[0122] In addition, such as Figure 6 As shown, system 100 can be used by or communicate with an AI-driven organization. An AI-driven organization may include one or more AI work units, each of which may include one or more AI agents. Each AI agent may include one or more AI models and / or tools (e.g., retrieval-enhanced generative agents, web crawlers, etc.). The AI ​​models may be pre-trained (e.g., LLM, DocAI™, Route-AI™, etc.), custom-trained (e.g., image classification), and / or deployed in the cloud for use with API handshakes. The output of the AI ​​work unit can be fine-tuned based on the target audience.

[0123] Other examples of analytics engines are also described in U.S. Patent Application No. 19 / 028,962, filed January 17, 2025, entitled "System and Method for Planning Using Artificial Intelligence," which claims priority to U.S. Provisional Patent Application No. 63 / 667,639, filed July 3, 2024, and U.S. Patent Application No. 19 / 029,167, filed January 17, 2025, entitled "System and Method for Monitoring Using Artificial Intelligence," which itself claims priority to U.S. Provisional Patent Application No. 63 / 667,639, filed July 3, 2024, the entire contents of which are incorporated herein by reference.

[0124] Technical background and considerations In existing solutions, product or service providers may attempt to quantify the success of user interactions with their products or services by requesting users to complete surveys and / or submit ratings (such as 4 / 5 stars). In response to such feedback, providers can adjust their products or services and monitor subsequent results. Providers may also use sales data alone as a proxy for user interaction success. However, these approaches may have some limitations.

[0125] For example, providers may not be able to directly measure users' emotional responses to a product or service, may lack personalized improvement suggestions, and may only receive feedback (such as ratings or sales data) after the interaction has occurred. This can lead to providers being unable to predict user reactions and / or feedback to a product or service, which may be undesirable.

[0126] Furthermore, over-reliance on the test user group may impair the ability to obtain feedback from the intended target population. The number of variations that can be tried may also be limited, and the data collected may be insufficient to support meaningful conclusions.

[0127] The embodiments described herein can address one or more of the limitations mentioned above by allowing the measurement or quantification of human interaction behaviors (such as interactions with products, services, and experiences) based on emotional responses. The embodiments described herein can support the time-varying processing of multimodal perceived data to compute emotional responses and can improve human-computer interaction and / or emotion modeling.

[0128] In some implementations, real-time feedback loops can be used to adjust products, services, or experiences based on interactions and measured emotional responses, thereby providing technical improvements over the existing methods described above. The implementations described herein can also provide methods for predicting emotional responses from different groups by simulating users with specific tendencies.

[0129] Interaction and emotional response Interactions can be difficult to quantify, and the information exchange process can constitute a technological black box: explicitly sending information is easy, but determining what is actually perceived or received is extremely difficult. Determining emotional responses to stimuli (such as products, events, experiences, and / or services) can depend on multiple factors. For example, an individual's interaction with a product, event, experience, and / or service may depend on their specific individual characteristics and their expectations, as well as the form of the product or service they are interacting with. Interactions and emotional responses can differ between entities, such as AI to humans, humans to AI, and AIs themselves. Other types of entity interactions are also possible.

[0130] Furthermore, in this article, the terms “product,” “event,” “experience,” and “service” can refer to media (such as videos, images, social media posts), means of transportation, travel experiences, places, social interactions (such as restaurant service, customer service calls, etc.), and any other stimuli that evoke an emotional response in an individual.

[0131] In some implementations, identifying the context can be crucial for understanding interactions with a product, service, or experience. This context can be set by residual chemicals such as dopamine and cortisol, whose concentrations decay slowly over time. For example, when someone is humiliated, their emotional state worsens that day, which constitutes a context for understanding whether a product, service, or experience was successful or not.

[0132] In some implementations, dopamine levels can be measured for users experiencing or interacting with a product or service. Users may experience dopamine spikes, drops, deficits, and / or a return to their baseline dopamine levels over time in response to stimuli. For example, a comedic experience or event (such as a comedy film / show or performance) may trigger anticipatory violation, resulting in high levels of dopamine secretion. Similarly, anticipation of positive events, such as listening to music or shopping, may also lead to dopamine secretion for some users. Anticipated dopamine can be based on the user's expectations, such as... Figure 7 As shown. The expected violation factor (referred to as the "S factor" in this paper) can be based on expected dopamine and actual event dopamine.

[0133] In some cases, such as narrative films or books, the climax of the story or the resolution of the conflict can satisfy the audience and / or soothe their emotions. For example, excellent stories often follow Freitag's pyramid, where the story begins in the setup phase, rising to a climax as tension, conflict, and challenge are presented. After the climax, the story may fall back as the conflict is resolved, ultimately leading to the ending, allowing the audience to reach a new narrative balance or state of reflection. Successful films often also feature protagonists with charismatic personalities who can attract the audience.

[0134] Successful products and services often share similar characteristics with successful movies. For example, they both act on different parts of the brain, framing expectations and meeting or exceeding those expectations, and following a trajectory of starting from a low point and climbing to a high point.

[0135] Furthermore, successful films evoke emotional responses in viewers, characterized by one or more S-shaped curves throughout the viewing experience. These S-shaped emotional response curves may start at a low point during interaction with external stimuli (such as watching a film) and then climb to higher values. The S-shaped curve can be continuous in nature, representing an ideal emotional response. For example, setting multiple points of expectation deviance throughout the film can elicit a more successful response.

[0136] In some implementations, as described herein, a user's dopamine levels can be used as a proxy indicator of emotional response. For example, emotional response can be based on a user's S-factor dopamine release levels. When a user has high S-factor dopamine levels exhibiting an S-shaped curve during a given time period, emotional response will be significantly enhanced.

[0137] In addition to dopamine secretion, a user's S factor can also be based on other sensory measurements. For example, a user's serotonin levels can be measured and / or estimated and used to determine the S factor. In some cases, serotonin levels can be predicted or estimated using an AI agent. Other senses (including vision, hearing, touch, taste, and smell) can be measured and used to calculate the S factor. Other measurements may also be used. There are individual differences in users' sensory measurements and emotional responses to different products or services.

[0138] In some implementations, the S factor can be personalized for users. For example, in determining the context of emotional response to a romantic movie, users with a higher degree of human romantic inclination may have a higher S factor compared to users with a lower degree of human romantic inclination.

[0139] Emotional responses using the S-factor as a proxy indicator can be assigned numerical values. In some implementations, emotional responses can be quantified using vector embeddings or numbers.

[0140] A user's S-factors can be determined for specific emotional energy levels. For example, energy levels may include: safety (EL1), pleasure (EL2), focus and competence (EL3), courage and fearlessness (EL4), communication and self (EL5), and goal orientation (e.g., a goal-driven vision with a holistic perspective) (EL6); however, other variations of energy levels are also possible and can be assigned different EL labels. Determining S-factors for specific energy levels can provide the context necessary to determine emotional responses and / or the success of a product or service. For example, a product provider may only be concerned with the product's success at a particular energy level. In some implementations, energy levels may also be referred to as emotional profiles. Energy levels can also be mapped to Maslow's hierarchy of needs.

[0141] In some implementations, the S-factor, which measures how a user’s response to a product, event, experience, or service changes over time, can be used to determine the success of that product, event, experience, or service. Figure 8 Examples are shown of segmenting the data for different levels of perception and energy to calculate the S-factor. As described herein, in some implementations, the derivative of the S-factor with respect to time can also be calculated and used to determine the success of a product, event, experience, or service, including financial success, changes in engagement, or loyalty.

[0142] When determining the success of a product, event, experience, or service, a user's emotional response or change in emotional response can be compared to the expected emotional response or change in expected emotional response. The expected emotional response may exhibit an S-shaped curve, starting at a low point during interaction with the external stimulus and then rising to a higher value. However, it is possible that the expected emotional response may possess other curvilinear characteristics.

[0143] Expected emotional responses can be specific to energy levels, user profiles, or other factors. For example, expected emotional responses used for comparison might differ for comedy and action movies. Similarly, expected emotional responses can depend on user demographics and other goals. Expected emotional responses can also include or describe power, ego, and / or global goal-oriented responses.

[0144] Exemplary analytics engine for emotional responses Figure 9 A system 200 is shown that can be used to determine emotional responses. System 200 includes perceptual data 204, an analysis engine 206, and output 208. Optionally, system 200 may include external stimuli 202. System 200 can be used by a company or enterprise. Analysis engine 206 may be the same as or similar to analysis engine 104 and / or analysis engine 104' described above.

[0145] Figure 10 An external stimulus 202 is shown, which may be or include a media form, including an advertisement 210, a video 212, and / or a film 214. The external stimulus 202 may also be text, an image, audio, and / or other media. In some embodiments, the external stimulus 202 may be adjusted based on an analysis engine 206. For example, the analysis engine may modify, correct, delete, add, or regenerate the external stimulus 202. Alternatively or complementary, the external stimulus 202 may include other media forms.

[0146] In some examples, external stimulus 202 may be or include experience 216, such as driving a vehicle, riding a roller coaster, or other stimuli that lead to the generation of perceptual data. As a supplement or alternative, external stimulus 202 may include other types of stimuli.

[0147] like Figure 11 As shown, the output 208 of the analysis engine 206 can be an emotional response 220 and / or a change in emotional response 222. The emotional response 220 can be determined by the analysis engine 206 based on perceptual data 204. Numerical values ​​can be assigned to the emotional response 220. For example, the S-factor of the user's response to external stimuli 202 over time, based on perceptual data 204, can represent the emotional response 220 or be used as a proxy indicator.

[0148] In some implementations, output 208 may also include the derivative of the change in emotional response and / or the predicted financial success of the medium (such as advertisement 210, video 212 or film 214 and / or text, images, audio, etc.) or experience 216. In some implementations, output 208 may provide feedback to analytics engine 206 to determine the derivative of the change in emotional response. For example, the derivative of the S-factor with respect to time may be calculated to represent the derivative of the change in emotional response 222.

[0149] In some implementations, output 208 can provide feedback to analysis engine 206 to determine the second derivative of emotional response change 222.

[0150] In another implementation, the change in emotional response 222 may refer to the difference relative to a predetermined threshold.

[0151] Determining the emotional response 220 may also include mapping or classifying the external stimulus 202 to a specific energy level. Classification of the external stimulus 202 can provide context for determining the emotional response 220 and / or the success of the external stimulus 202 in providing the desired emotional response 220. Classification can be performed by the analysis engine 206 using one or more AI agents. As mentioned above, the AI ​​agent may include one or more LLM and / or other software modules, tools, models, etc. In other embodiments, classification may also be performed using one or more propensity profiles, each representing an exemplary person (e.g., a young man, an elderly woman, a doctor, a child) with specific characteristics or propensities, allowing the success of the external stimulus 202 to be evaluated from the perspective of that exemplary person. Classification using propensity profiles can be performed using one or more AI agents. In other embodiments, mapping may also be performed manually.

[0152] Figure 12A -C illustrates examples of various input sensing, energy levels, and effectors. For example, such as... Figure 12A As shown, auditory, olfactory, visual, gustatory, and / or tactile senses can provide input perception, which can generate measured perceptual data 204 and / or corresponding emotion curves in response to triggering conditions. Additionally, Figure 12B Five possible energy levels are shown, such as self, love / fear, power / focus, pleasure / energy / DNA evolution, and / or security, which can act as invisible driving forces within a black box. Each of these energy levels can reside in the human brain and can provide a corresponding emotional curve in response to triggering conditions such as external stimuli.

[0153] Figure 12CVarious output effectors, such as speech, legs, hands, genitals, and / or anus, are illustrated, each capable of providing a corresponding emotional curve. For example, an output effector may only be triggered when a human takes action in response to a stimulus, regardless of whether the stimulus is external or originates from internal bodily functions. In some cases where the human is a passive observer, the effector may not be triggered. In other cases, the effector may be triggered when there is a response to a threat, opportunity, and / or fight-or-flight response. At any given moment, a human's attention and / or cognition may be associated with or correlated with one or more input perceptions, energy levels, and / or effectors. In some implementations, genitals, anus, and / or other effectors may not be externally relevant or may not provide a useful emotional curve in response to triggering conditions.

[0154] In some implementations, output 208 may also include a success metric 224. Success metric 224 may be a prediction of financial success, such as predicted revenue, gross profit, or net income. For example, external stimulus 202 may be a form of media, and success metric 224 may be determined by analytics engine 206, which may describe the predicted success (e.g., financial success) of the media. In other implementations, success metric 224 may describe other aspects of external stimulus 202, such as ranking or rating, including comparisons to other external stimuli.

[0155] Success metric 224 could also be a change in engagement or loyalty. In another example, external stimulus 202 could be experience 216, such as driving a new car model, flying an airplane, or dining at a restaurant. Success metric 224 could be determined by analytics engine 206. In the exemplary experience 216 of driving a new car model, success metric 224 could represent a predicted increase in user loyalty after purchasing the new car. In this example, external stimulus 202 could be a specific experience associated with the new car model, such as sitting in the driver's seat, opening the trunk, or reversing out of the driveway.

[0156] Output 208 may further include suggestion 226. Analysis engine 206 may generate suggestion 226 based on success metrics 224 associated with external stimulus 202, which can be used to adjust external stimulus 202. For example, suggestion 226 may provide instructions to analysis engine 206 to adjust one or more aspects of external stimulus 202, thereby generating new perceptual data 204 that will produce the desired success metrics 224. Analysis engine 206 may output a new success metric based on the adjusted external stimulus 202 and the generated perceptual data 204.

[0157] Alternatively, suggestion 226 may provide instructions to analysis engine 206 to adjust one or more aspects of perceptual data 204, which will produce a desired success metric. For example, adjusting perceptual data 204 may include generating external stimuli that produce the perceptual data. Analysis engine 206 may perform continuous analysis of external stimuli 202 and / or perceptual data 204 any number of times and / or at any frequency, based on suggestion 226 and / or success metric 224.

[0158] In other implementations, output 208 may also include new content generated by analytics engine 206, for example, based on an S-curve. The new content may include text, images, videos, and / or other media. The expected sentiment response of the new content may, as described above, start at a low value at the beginning of a period and rise to a higher value at the end of that period. This expected sentiment response can be used to create new content with higher audience engagement.

[0159] Figure 13 An analysis engine 206 according to some embodiments is shown. In some embodiments, the analysis engine 206 may be the same as or substantially similar to the analysis engine 104 and / or the analysis engine 104'.

[0160] Analysis engine 206 may include one or more AI agents (referred to herein individually or collectively as AI agent 230). For example, analysis engine 206 may contain only a single AI agent 230 for calculating emotional response changes 222 in response to the reception of external stimulus 202. In other examples, analysis engine 206 may include multiple AI agents 230 that can work collaboratively to perform the analysis of analysis engine 206. AI agents 230 may interact with... Figure 2 and / or Figure 3 The AI ​​agent 114 shown is the same as or similar to the AI ​​agent shown.

[0161] AI agent 230 can be trained to simulate users with specific preferences. For example, analytics engine 206 may optionally include a preference profile 232. The preference profile 232 may include one or more datasets for training one or more AI agents 230 to simulate one or more users with specific preferences.

[0162] The propensity profile 232 may include a dataset for defining at least one of age, gender, knowledge, lifestyle, and / or preferences. The propensity profile 232 may also include users with specific emotional responses and / or perceptual data. In other embodiments, the propensity profile 232 may include an AI model trained to simulate a user with specific propensities (e.g., specific age, gender, knowledge, lifestyle, and / or preferences).

[0163] The analysis engine 206 can calculate a user's emotional response 220 and / or S-factor based on specific tendencies. For example, for a user profile that shows a weak response to human love, the analysis engine 206 can calculate a lower S-factor for romantic movie scenes.

[0164] AI agent 230 can communicate with these models in the propensity profile to simulate users with specific propensities, such as generating perceptual data 204 based on external stimuli 202. Other examples are also possible, such as propensity profiles being integrated into the AI ​​agent (e.g., the AI ​​agent has been trained to simulate one or more users with specific propensities).

[0165] Optionally, the analytics engine 206 can generate a prediction of financial success 234. The analytics engine 206 can generate a prediction of financial success 234 related to external stimuli 202 based on at least one of emotional response 220 or emotional response change 222. The analytics engine 206 can utilize an AI agent 230 trained to simulate users with specific tendencies to generate the prediction of financial success 234.

[0166] It should be understood that analysis engine 206 may contain fewer or more modules or components than analysis engine 104 and / or analysis engine 104'. Analysis engine 206 may also include or instead include other modules or components as a supplement to or replacement of existing modules or components in analysis engine 104 and / or analysis engine 104'.

[0167] In this document, an AI agent can refer to a single AI agent, which may include an LLM or a multimodal large language model (MLLM), equipped with a model or tool for performing one or more tasks of the analysis engine 206. An AI agent can also refer to a group of AI agents that can be used to perform one or more tasks of the analysis engine 206. Therefore, the terms "AI agent" and "a group of AI agents" are used interchangeably in this document.

[0168] Additionally, an AI agent may include at least one AI model, such as an LLM or MLLM. An AI agent may also include one or more other models or tools to enable it to perform one or more tasks, such as those used in analytics engine 206. It should be understood that the terms “LLM” and “MLLM” are used interchangeably in this document.

[0169] like Figure 14As shown, perceptual data 204 may include human perceptual data, such as data describing vision 240, hearing 242, touch 244, taste 246, and smell 248. Data describing other perceptual data may also be included in perceptual data 204. Perceptual data 204 may also include, describe, or be used to determine dopamine and / or serotonin concentrations, including as a function of time, such as Figures 15A-15D As shown in the example. Perceptual data 204 may include human perceptual data, and these terms may be used interchangeably in this document.

[0170] Sensing data 204 can be measured, generated by AI (e.g., an AI agent), and / or manually input by a person. Different types of sensing data 204 (e.g., human sensing data) can have associated bandwidths, such as... Figures 16A-16C As shown. For example, bandwidth can depict the hardware data required to measure and / or stream perceptual data 204 (or certain aspects of perceptual data 204, such as vision 240), or the relative amount of perceptual data 204 (e.g., vision 240 accounts for or constitutes 90% of the measurement data, while touch 224 accounts for or constitutes only 9% of the measurement data). The analysis engine 206 can utilize the bandwidth of perceptual data 204 when calculating emotional response 220.

[0171] In one implementation, the sensing data 204 can be collected via biometric sensor 252 or manual observation 254, such as Figure 17 As shown. Biometric sensor 252 may include measuring heart rate, skin conductance, measuring inflammation via fingerprint dilation, piloerection (e.g., goosebumps), and / or eye dilation.

[0172] For example, a skin conductance sensor can be applied to one or more fingers or other areas of the skin to measure the surface conductivity of the skin, which may change due to sweat gland secretion and may change in response to external stimuli. Figure 18 Examples of skin conductance and heart rate measured during movie watching are shown.

[0173] As another example, goosebumps can be measured and may respond to electrical signals from the hippocampus to the skin (e.g., nerve endings in the hands and sometimes the legs). Goosebumps can also extend to a piloerection response, as a reaction to hearing fingernails scratching a blackboard, listening to inspiring music, or feeling and / or recalling strong or positive emotions.

[0174] Other biometric sensors 252 and measurement methods may be used to acquire human-sensory data 204. For example, modules, multi-sensors, and recording sensors may be used in any combination to acquire human-sensory data 204. Modules and / or multi-sensors may include WiFi communication modules, Panda multi-sensors, graphic display modules, USB modules, battery modules, RF communication modules, and / or digital display modules. Recording sensors may include voltage, current, temperature, light, oxygen, pH, relative humidity, heart rate and pulse, photogate, pressure, force, sound, motion, magnetic field, conductivity, spirometer, electrocardiogram, colorimeter, CO2, barometer, blood pressure, dropper, flow rate, force plate, rotational motion, acceleration, salinity, soil moisture, UVB, turbidity, UVA, surface temperature, wide-range temperature measurement, infrared thermometer, respiratory monitoring band, hand grip strength meter, calcium, chloride, ammonium, nitrate, anemometer, GPS, dew point, charge, Geiger counter, milliampere current, and / or resistance recording sensors.

[0175] In some implementations, mapping functions, algorithms, or AI models (such as one or more AI agents 230) may be used to transform data acquired from biometric sensors 252 to obtain perceptual data 204.

[0176] Perceptual data 204 can also be obtained from a human audience 250, for example, using biometric sensors 252 and / or human observation 254, such as... Figure 19 As shown. The perceptual data 204 obtained from the human audience 250 can vary based on external stimuli 202 and can be used as training and / or input data for the AI ​​agent 230.

[0177] Figure 20 Examples of human perception data 204 measured based on different external stimuli 202 are shown, and in particular, different plot features throughout the film and their respective measured energy levels are shown.

[0178] In another embodiment, the perceived data 204 can be generated by the AI ​​agent 230, such as... Figure 21 As shown. In some implementations, the AI ​​agent 230 can be trained to simulate a user with specific preferences. In this document, the term "generated" can also refer to simulation, generation, artificial generation, and / or construction. Perceptual data 204 (such as human perceptual data) can also be collected by the AI ​​agent 230.

[0179] The analysis engine 206 can generate external stimuli 202, which can be used to generate perceptual data 204. For example, external stimuli 202 can be generated randomly, based on a saved template, based on some other input value, and / or using any number of other methods.

[0180] In another implementation, the AI ​​agent 230 can interpret the external stimulus 202 and simulate the responses that a specific user with a particular predisposition might make, such as... Figure 22 As shown. For example, AI agent 230 can generate perception data 204 for users with specific preferences, such as by using preference profiles 232. In a specific example, AI agent 230 can generate perception data 204 in response to driving a new car model. Analysis engine 206 can use simulated perception data 204 to predict success metrics 224 (such as satisfaction, changes in loyalty, etc.), generate predicted financial success 234, and / or generate recommendations 226.

[0181] In another example, the perceptual data 204 can be generated by the analytics engine 206 through user engagement analysis, video content analysis, natural language processing, image content analysis, audio content analysis, or video content analysis. The simulated perceptual data 204 can be provided as feedback to the analytics engine 206.

[0182] Analysis engine 206 can selectively utilize suggestions 226 and / or success metrics 224 to adjust external stimulus 202 (such as changing the appearance of a new car model) and / or propose adjustments to external stimulus 202. For example, analysis engine 206 can modify, correct, delete, add, or regenerate external stimulus 202 (such as suggesting changing the color of a new car model or modifying the length of the medium in external stimulus 202). Adjustments, regenerations, and / or adjustments to external stimulus 202 can be performed by analysis engine 206.

[0183] like Figure 23 As shown, perceptual data 204, including simulated / artificial perceptual data, can be used by analysis engine 206 to determine emotional response 220 for output. Emotional response change 222 can also be determined by analysis engine 206 based on emotional response 220. For example, emotional response change 222 can represent the change of emotional response 220 over time, such as the derivative of emotional response 220.

[0184] In some implementations, the change in emotional response 222 can be used as feedback provided to the analysis engine 206, for example, to determine the derivative of the change in emotional response.

[0185] In some implementations, the sensed data 204 may include multiple time periods 260, such as Figure 24 As shown. The changes 222 in emotional responses within each of the multiple time periods 260 can be calculated by the analysis engine 206. The intervals between the multiple time periods 260 can be equally spaced, monotonically varied, variable, and / or any other possible type.

[0186] The change 222 of emotional response within each time period 260 can be calculated for each type of sensory data 204 (such as visual 240, auditory 242, tactile 244, gustatory 246, and / or olfactory 248). Multiple time periods 260 can differ for each type of sensory data 204 and can be based on their respective bandwidths.

[0187] In some implementations, the prediction of financial success 234 by the external stimulus 202 can be determined based on at least one of emotional response 220 or change in emotional response 222, such as Figure 25 As shown. For example, the analytics engine 206 can measure predicted financial success 234 in relation to external stimulus 202 based on at least one of emotional response 220 or emotional response change 222 (e.g., utilizing AI agent 230). As described above, in some implementations, AI agent 230 can be trained to simulate a user with specific tendencies.

[0188] The calculation of the success metric 224 and / or predicted financial success 234 may depend on the value of the S-factor associated with energy level or emotional energy level classification. In other implementations, the success metric 224 and / or predicted financial success 234 may be calculated based on the total weighted value of the S-factor across any number of energy level categories. For example, a film that appeals to a wider audience may provide a higher weighted value for the S-factor across multiple energy levels.

[0189] Figures 26A-26B An example of predicting financial success for a film is shown. For instance, an S-factor can be calculated across one or more energy levels and weighted across other energy levels to predict financial success. In some implementations, analytics engine 206 can utilize regression models to evaluate predicted revenue, profit, or value. Alternatively, analytics engine 206 can utilize AI agent 230 and / or models trained to predict financial success based on external stimuli, perceptual data, and / or emotional responses.

[0190] In some implementations, multiple models can be stacked after an LLM. For example, these models may include expert roles, questioners, integrators, and regression models with historical sales data.

[0191] In some implementations, user engagement graphs (e.g., from a video player) can be used to train the AI ​​agent 230. For example, a user engagement graph can provide viewing time data, which may include skipped or replayed video portions, and can be represented by one or more peaks and / or troughs. Peaks may indicate popular or key moments, while troughs may indicate video portions that are less appealing to the viewer.

[0192] like Figure 27As shown, the expected emotional response can exhibit an S-shaped curve, where the emotional response is low at the beginning of a period and rises to a higher value at the end of that period. For example, the x-axis of the curve can represent the time period, and the y-axis can represent the emotional response. The expected emotional response can be determined based on human audience 250, external stimuli 202, AI agent 230, propensity profile 232, and / or other inputs (such as financial success goals).

[0193] like Figure 28 As shown, an unintended emotional response can present as a curve, where the emotional response does not rise to a higher value at the end of the period, or the emotional response declines during this period. Other examples may also exist, such as when the emotional response does not exhibit [a certain pattern / significance]. Figure 27 The S-curve is shown. Undesired emotional responses can be determined based on human audience 250, external stimuli 202, AI agent 230, propensity profile 232, and / or other inputs (such as minimum success metrics).

[0194] In some implementations, the emotional response 220 can be compared to the expected emotional response. For example, if the emotional response 220 is similar to the S-shaped curve of the expected emotional response, the system can output a higher success index 224. Alternatively, if the emotional response 220 does not simulate or is not similar to the S-shaped curve of the expected emotional response, the system can output a lower success index 224.

[0195] In some implementations, success metric 224 can be used to measure predicted financial success 234, such as Figure 29 As shown. The analysis engine 206 can determine success indicators 224 related to external stimuli 202 by comparing the derivative of the change in emotional response over a period of time and the change in emotional response 222 over that period with the expected emotional response.

[0196] Figure 30 A method 300 is shown for determining emotional responses to external stimuli (e.g., emotional response 220 and / or emotional response change 222). Method 300 can be executed by system 200, particularly using analysis engine 206.

[0197] In step S302, human perception data in response to external stimuli over a period of time is received, including multiple time periods.

[0198] Human sensory data may include vision 240, hearing 242, touch 244, taste 246 and / or smell 248. Sensory data 204 may also include dopamine concentration and / or serotonin concentration.

[0199] Perceptual data 204 can be acquired from a human audience 250. Biometric sensors 252 may include measuring heart rate, skin conductance, measuring inflammation via fingerprint dilation, piloerection (e.g., goosebumps), and / or eye dilation. However, other biometric sensors 252 and measurement methods may be used to acquire human perceptual data 204.

[0200] In some implementations, mapping functions, algorithms, or AI models can be used to transform the data acquired by biometric sensor 252 to obtain perceived data 204.

[0201] In another implementation, the perceptual data 204 can be generated by AI (e.g., using AI agent 230). In this document, "generated" can also mean simulated, generated, artificially generated, and / or constructed. For example, AI agent 230 can be trained to simulate a user with specific preferences. The perceptual data 204 (such as human perceptual data) can also be collected from AI agent 230.

[0202] The analysis engine 206 can also generate external stimuli 202, which can be used to generate perceptual data 204. For example, external stimuli 202 can be generated randomly, based on a saved template, in response to specific input data, and / or using one or more other methods.

[0203] In another implementation, AI agent 230 can interpret external stimuli 202 and simulate responses that a specific user with a particular preference might make. For example, AI agent 230 can use preference profiles 232 to generate perceptual data 204 for a user with a particular preference in response to driving a new car model.

[0204] In another embodiment, the perceptual data 204 can be generated by the analysis engine 206 through user engagement analysis, video content analysis, and / or natural language processing. The simulated perceptual data 204 can be provided to the analysis engine 206 as feedback.

[0205] In step S304, emotional response is determined based on human perception data.

[0206] Emotional response 220 can be determined by analysis engine 206 based on perceptual data 204.

[0207] Determining the emotional response 220 may also include mapping or classifying the external stimulus 202 to a corresponding energy level. Classification of the external stimulus 202 can provide the necessary context for determining the emotional response 220 and / or its success. Classification can be performed by the analysis engine 206 using LLM and / or AI agent 230 and / or propensity profile 232. In some implementations, mapping may also be performed manually.

[0208] The analysis engine 206 can calculate a user's emotional response 220 and / or S-factor based on specific tendencies. For example, for a user profile that shows a weak response to human love, the analysis engine 206 can calculate a lower S-factor for romantic movie scenes.

[0209] In step S306, a numerical value is assigned to the emotional response.

[0210] Numerical values ​​can be assigned to emotional response 220. For example, the S-factor, which is based on perceptual data 204 and represents the user's response to external stimuli 202 over time, can represent emotional response 220 or be used as a proxy indicator.

[0211] In step S308, the changes in emotional response during each of the multiple time periods within this period are calculated.

[0212] The emotional response changes 222 within each of the multiple time periods 260 can be calculated by the analysis engine 206. The intervals between the multiple time periods 260 can be equal, variable, or any other possible type.

[0213] The change 222 of emotional response within each time period 260 can be calculated for each type of sensory data 204 (e.g., visual 240, auditory 242, tactile 244, gustatory 246 and / or olfactory 248).

[0214] Figure 31 A method 400 for determining success based on emotional response is shown. Method 400 can be executed by system 200, particularly using analysis engine 206. Method 400 can be executed in combination with method 300.

[0215] In step S402, the derivative of the change in emotional response during this period is calculated.

[0216] In one implementation, output 208 can provide feedback to analysis engine 206 to determine the derivative of the change in emotional response. For example, the derivative of the "S" factor with respect to time can be calculated to represent the derivative of the change in emotional response 222.

[0217] In another implementation, output 208 can provide feedback to analysis engine 206 to determine the second derivative of the change in emotional response.

[0218] In step S404, based on the derivative of the change in emotional response over the time period and the change in emotional response over the time period, a success indicator related to external stimuli is determined by comparing it with the expected emotional response.

[0219] Success metric 224 can be a prediction of financial success, such as projected revenue, gross profit, or net income. For example, external stimulus 202 can be a form of media, and success metric 224 can be determined by analytics engine 206, which can describe the predicted success (e.g., financial success) of the media. In other implementations, success metric 224 can describe other aspects of external stimulus 202, such as ranking or rating, including comparisons with other external stimuli.

[0220] Alternatively or additionally, success metric 224 could be a change in engagement or loyalty. In another example, external stimulus 202 could be an experience 216, such as the process of driving a new car model. Success metric 224 can be determined by analytics engine 206, which could represent a predicted increase in user loyalty after purchasing a new car. In another example, external stimulus 202 could be a specific experience associated with the new car model, such as sitting in the driver's seat, opening the trunk, or reversing out of the driveway.

[0221] The calculation of success metric 224 and / or predicted financial success 234 may depend on the value of the S-factor associated with the energy level. In other implementations, success metric 224 and / or predicted financial success 234 may be calculated based on the total weighted value of the S-factor across any number of energy level categories. For example, a film that appeals to a wider audience may provide a higher weighted value for the S-factor across multiple energy levels.

[0222] In one implementation, the derivative of the change in emotional response over this period can be determined based on the S-factor. For example, the derivative of the S-factor with respect to time can be calculated and used to determine success.

[0223] Expected emotional responses can exhibit an S-shaped curve, where the emotional response is low at the beginning of the period and rises to a higher value at the end. Expected emotional responses can be determined based on human audience 250, external stimuli 202, AI agent 230, propensity profile 232, and / or other inputs (such as financial success goals).

[0224] Unexpected emotional responses can be represented by the following curve, where the emotional response does not rise to a higher value at the end of the period, or the emotional response declines during the period. Unexpected emotional responses can be determined based on human audience 250, external stimuli 202, AI agent 230, propensity profile 232, and / or other inputs (such as minimum success indicators).

[0225] The emotional response 220 can be compared to the expected emotional response. For example, if the emotional response 220 resembles the S-shaped curve of the expected emotional response, the system can output a higher success index 224. Alternatively, if the emotional response 220 does not mimic the S-shaped curve of the expected emotional response, the system can output a lower success index 224.

[0226] Expected emotional responses can be specific to energy levels, user profiles, or other factors. For example, the expected emotional responses used for comparison may differ for comedies versus action films. Similarly, expected emotional responses can depend on user demographics and goals.

[0227] Optionally, output 208 may further include suggestions 226. Analysis engine 206 may selectively utilize suggestions 226 and / or success metrics 224 to adjust external stimulus 202 (e.g., changing the appearance of a new car model). For example, analysis engine 206 may modify, correct, delete, add, or regenerate external stimulus 202 (e.g., changing the color of a new car model). Adjustment and / or regeneration of external stimulus 202 can be performed by analysis engine 206.

[0228] Alternatively, suggestion 226 may provide instructions to analysis engine 206 to adjust one or more aspects of the perceived data 204 (which will produce the desired success metric) and generate the external stimulus that generated the perceived data. Analysis engine 206 may then output a new success metric based on the adjusted external stimulus 202 and the generated perceived data 204.

[0229] The analysis engine 206 can generate suggestions 226 based on success metrics 224 associated with the external stimulus 202, which can be used to adjust the external stimulus 202. The analysis engine 206 can perform continuous analysis of the external stimulus 202 and / or perceived data 204 any number of times based on the suggestions 226 and / or success metrics 224.

[0230] Exemplary Movie Analysis System The following examples of film analysis systems may be included or executed by the aforementioned system 100 and / or system 200.

[0231] like Figure 32A As shown in -B, the film analysis system can be implemented using analysis system 200 to predict film value. For example, film scripts, recorded footage, or videos (external stimuli 202) can be received by the strategic planning AI agent, which can interact with other modules as previously described. The strategic planning AI agent may include or host other AI agents programmed to perform other specific tasks.

[0232] Recorded material or video can be converted into text using a speech-to-text translation AI model or agent. An emotion "S" curve (such as anticipated violation factors) can be extracted based on the text conversion. In some implementations, the emotion curve can be calculated using an AI agent 230 trained to simulate a specific tendency. The AI ​​agent can generate or produce perceptual data 204 for a specific tendency when calculating the emotion curve.

[0233] In addition, the strategic planning AI agent can calculate emotional responses 220 and / or changes in emotional responses based on external stimuli 202.

[0234] The combiner can receive emotional response 220 and / or emotional curves and use a value prediction model to calculate an estimated value for the film. For example, the predicted financial success 234 of external stimulus 202 can be determined based on at least one of emotional response 220 or emotional response change 222.

[0235] Success Indicator 224 and / or Predicted Financial Success 234 can be calculated based on the total weighted value of the S-factor across any number of energy level categories. For example, a film that appeals to a wider audience may provide a higher weighted value for the S-factor across multiple energy levels and thus receive a higher film value estimate.

[0236] Other variations and implementations may exist in the exemplary movie analysis engine. Furthermore, the above description is not intended to limit the variations and implementations described herein.

[0237] In other embodiments of the methods and systems described herein, the methods and systems can also be used to generate new content based on an S-curve. The new content may include text, images, videos, and / or other media. The expected emotional response to the new content may begin at a low value at the start of a period and rise to a higher value at the end of that period, as described above. This expected emotional response can be used to create new content with a high audience engagement factor.

[0238] Exemplary computing device Figure 33 This is a schematic diagram of a computing device 1100 according to some embodiments, configured as a component of system 100 and / or system 200. The computing device 1100 includes a memory 1102, a processor 1104, and a bus 1106. The computing device 1100 may also include a network interface 1108. The memory 1102, processor 1104, and network interface 1108 are communicatively connected via the bus 1106.

[0239] Processor 1106 and network interface 1108 are configured to perform the steps of method 300 and / or method 400 when a program or computer-executable instructions stored in memory 1102 are executed by processor 1104. Processor 1104 and network interface 1108 may also be configured to perform any other processes or modules discussed with respect to system 100, system 200, system 800, analysis engine 104, analysis engine 104', and / or analysis engine 206 when a program or computer-executable instructions stored in memory 1102 are executed by processor 1104.

[0240] Memory 1102 may be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). Memory 1102 may store programs or computer-executable instructions. Memory 1102 may be non-volatile memory.

[0241] The processor 1104 may be a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), a graphics processing unit (GPU), or one or more integrated circuits.

[0242] Additionally, processor 1104 can be an integrated circuit chip with signal processing capabilities. During implementation, the steps of method 300 and / or method 400 can be executed by integrated logic circuitry in hardware or by instructions in software within processor 1104. Furthermore, processor 1102 can be a general-purpose processor, digital signal processor (DSP), ASIC, field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component.

[0243] Processor 1102 can implement or execute the methods, steps, and logic diagrams disclosed in the embodiments. The general-purpose processor can be a microprocessor, or the processor can be any conventional processor or similar device. The method steps described herein can be executed directly by a hardware decoding processor, or by combining the hardware of the decoding processor with software modules. The software modules can reside in industry-standard storage media, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. The storage medium can reside within memory 1102. Processor 1104 can read information from memory 1104 and use the hardware of processor 1104 to complete the steps of method 300 and / or method 400.

[0244] Network interface 1108 enables communication between computing device 1100 and one or more other devices and / or computing devices via a communication network, such as using transceiver devices (e.g., including but not limited to transceivers). For example, components of system 100 and / or system 200 may be configured to communicate with each other via a communication network. In a particular example, user device 102 and analytics engine 104 may communicate with each other using their respective network interfaces.

[0245] Bus 1106 may include paths and / or communication channels for transmitting information between all components of computing device 1100.

[0246] It should be noted that, although Figure 33The computing devices shown in the figure only include memory, processor and communication interface. However, those skilled in the art will understand that in the specific implementation, system 100 and / or system 200 and analysis engine 104, analysis engine 104' and / or analysis engine 206 may further include other components required for implementation, such as one or more additional computing devices, servers, networks, memory, processors, etc.

[0247] Furthermore, based on specific needs, those skilled in the art should understand that the components of these systems may further include hardware components for implementing other additional functions. Additionally, those skilled in the art should understand that system 100 and / or system 200 may include only the components necessary for implementing embodiments of the present invention, without including... Figure 33 All the components shown.

[0248] In the method described above, boxes may represent events, steps, functions, processes, modules, state-based operations, etc. Although some of the examples above are described as occurring in a specific order, those skilled in the art will understand that some of these steps or processes may be performed in different orders, provided that a change in the order of steps does not prevent or hinder the occurrence of subsequent steps.

[0249] Furthermore, some of the aforementioned messages or steps may be deleted or merged in other implementations, and some messages or steps may be divided into several sub-messages or sub-steps in other implementations. Further, some or all steps may be repeated as needed. Elements described as methods or steps also apply to systems or sub-components, and vice versa. Terms such as "send" or "receive" may be used interchangeably depending on the perspective of a specific device, module, or logic element.

[0250] While some embodiments are described at least in part as methods, those skilled in the art will understand that some embodiments also refer to various components for performing at least some aspects and features of the process, whether implemented as hardware components, software or any combination of the two or in any other manner.

[0251] Additionally, some embodiments also point to pre-recorded storage devices or other similar computer-readable media having program instructions stored thereon for performing the processes described herein. Computer-readable media include any non-volatile storage media, such as RAM, ROM, flash memory, optical discs, USB drives, DVDs, HD-DVDs, or any other such computer-readable storage devices.

[0252] It should be understood that the device described herein includes one or more processors and associated memory. The memory may include one or more application programs, modules, or other programming structures containing computer-executable instructions, which, when executed by one or more processors, are used to implement the methods or processes described herein.

[0253] In this document, an AI agent may include at least one AI model, such as a large language model or a multimodal large language model. An AI agent may also include one or more other models or tools to allow the AI ​​agent to perform one or more tasks. Furthermore, in this document, the term "AI agent" may also refer to a single AI agent, or to one or more AI agents, such as a group of AI agents.

[0254] The various embodiments described above are merely examples and are not intended to limit the scope of the embodiments. Various variations of the innovations described herein will be apparent to those skilled in the art and fall within the intended scope of the embodiments. In particular, features can be selected from the above embodiments to form alternative embodiments that include combinations of features that may not be explicitly described.

[0255] Furthermore, features can be selected and combined from one or more embodiments to form alternative implementations that include combinations of features that may not be explicitly described. Features applicable to these combinations and sub-combinations will be apparent to those skilled in the art upon a comprehensive review of the described embodiments. The subject matter described herein is intended to cover all suitable technical variations.

[0256] Certain adaptive adjustments and modifications can be made to the described embodiments. Therefore, the above embodiments should be considered exemplary.

[0257] The computer may be a computing device, such as a mobile device, personal computer, server, embedded system, or other device with computing capabilities.

Claims

1. A computer-implemented method, the method comprising: Receive human perception data in response to external stimuli over a period of time, the human perception data including multiple time periods; Emotional responses are determined based on the aforementioned human perception data; Assign a numerical value to the emotional response; and Calculate the changes in emotional response within each of the multiple time periods during this period.

2. The computer implementation method according to claim 1, wherein, The external stimulus is a form of media, including at least one of advertising, video, film, text, image, or audio, and the method further includes: predicting the financial success of the media based on at least one of the emotional response or changes in the emotional response.

3. The computer implementation method according to claim 1, wherein, The human sensory data includes data describing at least one of vision, hearing, touch, taste, or smell.

4. The computer implementation method according to claim 1, wherein, The human perception data is collected through biometric sensors or manual observation.

5. The computer implementation method according to claim 2, wherein, The expected emotional response follows an S-shaped curve, wherein the emotional response is at a low value at the beginning of the time period and rises to a higher value at the end of the time period.

6. The computer-implemented method according to claim 1, wherein, The human perception data is generated by an artificial intelligence agent.

7. The computer-implemented method according to claim 6, wherein, The AI ​​agent is trained to simulate users with specific tendencies.

8. The method according to claim 1, further comprising: Calculate the derivative of the change in emotional response over the aforementioned period of time; and Based on the derivative of the change in emotional response over a certain period of time and the change in emotional response over that period of time, a success indicator related to the external stimulus is determined by comparing it with the expected emotional response.

9. The computer-implemented method according to claim 1, further comprising: Based on the success metrics associated with the external stimulus, suggestions for adjusting the external stimulus are generated.

10. A system comprising a user interface component and a server component, the server component being configured to: Receive human perception data in response to external stimuli over a period of time, the human perception data including multiple time periods; Emotional responses are determined based on the aforementioned human perception data; Assign a numerical value to the emotional response; and Calculate the changes in emotional response within each of the multiple time periods during this period.

11. The system according to claim 10, wherein, The external stimulus is a form of media, including at least one of advertising, video, film, text, image, or audio, and the system is further configured to predict the financial success of the media based on at least one of the emotional response or changes in the emotional response.

12. The system according to claim 10, wherein, The human sensory data includes data describing at least one of vision, hearing, touch, taste, or smell.

13. The system according to claim 10, wherein, The human perception data is collected through biometric sensors or manual observation.

14. The system according to claim 11, wherein, The expected emotional response follows an S-shaped curve, wherein the emotional response is at a low value at the beginning of the time period and rises to a higher value at the end of the time period.

15. The system according to claim 10, wherein, The human perception data is generated by an artificial intelligence agent.

16. The system according to claim 15, wherein, The AI ​​agent is trained to simulate users with specific tendencies.

17. The system of claim 10, wherein the server component is further configured as follows: Calculate the derivative of the change in emotional response over the aforementioned period of time; and Based on the derivative of the change in emotional response over a certain period of time and the change in emotional response over that period of time, a success indicator related to the external stimulus is determined by comparing it with the expected emotional response.

18. The system according to claim 11, wherein, The server is further configured to generate recommendations for the external stimulus based on the success metrics associated with the external stimulus.

19. One or more non-volatile computer-readable media having executable instructions stored thereon, the executable instructions causing the at least one computer to perform a method when executed by the at least one computer, the method comprising: Receive human perception data in response to external stimuli over a period of time, the human perception data including multiple time periods; Emotional responses are determined based on the aforementioned human perception data; Assign a numerical value to the emotional response; and Calculate the changes in emotional response within each of the multiple time periods during this period.

20. One or more non-volatile computer-readable media according to claim 19, wherein, The external stimulus is a form of media, including at least one of advertising, video, film, text, image, or audio, and the method further includes: predicting the financial success of the media based on at least one of the emotional response or changes in the emotional response.

Images