Adversarial reinforcement of automatic responses to queries

Abstract

A system, computer program product, and method for managing an examination having adversarial reinforcement of queries for automated responses are presented. The method includes electronically receiving an original query (502). A response to the original query is to be submitted electronically by a human. The method also includes modifying the original query, thereby generating a modified query (504). The modified query is configured to be human understandable and cannot be properly responded to by an electronic device without human support.

Description

BACKGROUND

[0001] The present disclosure relates to the administration of examinations, and more particularly to improving adversarial reinforcement of queries for automated responses.

[0002] Many known examination events occur through online means, where the examinee is located remotely and it is impractical for the examination officer to be physically present to administer the examination. With the continued pervasive presence of the Internet in our society, access to information has become more readily available, and the Internet serves as the primary destination for answers to queries. For example, answers to queries can be obtained quite conveniently through well-known and nearly ubiquitous artificial intelligence (AI) agents, such as digital assistants and chatbots, which are readily accessible through all forms of computing devices, including mobile devices. Thus, the possibility of administering examinations that utilize such technology to generate automated responses to examination queries requires some form of remote examination proctoring to maintain the integrity of the examination process. At least some known methods of administration and processing of remote, online examinations include intrusive monitoring, such as video-based monitoring of the examinee and associated screen and screen sharing. Both methods require the full, undivided, and uninterrupted attention of a human monitoring personnel. Moreover, both methods use additional computing resources that can impact the performance of the examination hardware and software, which can be burdensome for timed examinations. Furthermore, both methods increase the cost of the examination by using human monitoring personnel and additional hardware and software. SUMMARY

[0003] A system, computer program product, and method for administering examinations having adversarial reinforcement of queries for automated responses are provided.

[0004] In one aspect, a computer system for administering examinations having adversarial reinforcement of queries for automated responses is provided. The system includes one or more processing devices and at least one memory device operably coupled to the one or more processing devices. The one or more processing devices are configured to electronically receive an original query. A response to the original query is to be submitted electronically by a human. The one or more processing devices are further configured to modify the original query, thereby generating a modified query. The modified query is configured to be human understandable and not properly responded to by an electronic device without human support.

[0005] In another aspect, a computer program product for managing examinations having adversarial reinforcement of queries for automated responses is provided. The computer program product includes one or more computer-readable storage media, and program instructions collectively stored on the one or more computer-readable storage media. The product further includes program instructions for electronically receiving an original query. A response to the original query is to be electronically submitted by a human. The product further includes program instructions for modifying the original query, thereby generating a modified query. The modified query is configured to be human understandable and cannot be properly responded to by an electronic device without human support.

[0006] In yet another aspect, a computer-implemented method for managing examinations having adversarial reinforcement of queries for automated responses is provided. The method includes electronically receiving an original query. A response to the original query is to be electronically submitted by a human. The method further includes modifying the original query, thereby generating a modified query. The modified query is configured to be human understandable and cannot be properly responded to by an electronic device without human support.

[0007] This summary of the invention is not intended to identify key or essential aspects of the disclosure, nor is it intended to limit the scope of the disclosure. These and other features and advantages will become more apparent from the following detailed description of the embodiments, taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0008] The accompanying drawings, which are incorporated in and form a part of the specification, illustrate embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure. The drawings illustrate certain embodiments and are not limiting of the present disclosure.

[0009] Figure 1 is a schematic diagram illustrating a cloud computer environment, in accordance with some embodiments of the present disclosure.

[0010] Figure 2 is a block diagram illustrating a set of functional abstraction model layers provided by a cloud computing environment, in accordance with some embodiments of the present disclosure.

[0011] Figure 3 is a block diagram illustrating a computer system / server that can be used as a cloud-based support system to implement the processes described herein, in accordance with some embodiments of the present disclosure.

[0012] Figure 4 is a block diagram illustrating a computer system configured for managing examinations having adversarial reinforcement of queries for automated responses, in accordance with some embodiments of the present disclosure.

[0013] Figure 5Ais a flowchart illustrating a process for managing a check with adversarial augmentation for an automated response according to some embodiments of the present disclosure.

[0014] Figure 5B is a continuation of the flowchart from Figure 5A according to some embodiments of the present disclosure.

[0015] Figure 5C is a continuation of the flowchart from Figure 5B according to some embodiments of the present disclosure.

[0016] Figure 6A is a graphical image illustrating an example image associated with a check query according to some embodiments of the present disclosure.

[0017] Figure 6B is a graphical image of Figure 6A illustrating an example at least partially modified image associated with an at least partially modified check query according to some embodiments of the present disclosure.

[0018] Figure 6C is a graphical image of Figure 6A illustrating an example further modified image associated with a further modified check query according to some embodiments of the present disclosure.

[0019] Figure 6D is a graphical image of Figure 6A illustrating an example modified image associated with a modified check query according to some embodiments of the present disclosure.

[0020] Figure 7A is a textual representation illustrating an example textual check query according to some embodiments of the present disclosure.

[0021] Figure 7B is a block diagram illustrating an example audio check query based on a textual check query from Figure 7A according to some embodiments of the present disclosure.

[0022] Figure 7C is a block diagram illustrating an example audio check query based on an at least partially modified textual check query from Figure 7A according to some embodiments of the present disclosure.

[0023] Figure 7D is a block diagram illustrating an example audio check query based on a textual check query from Figure 7A with an added audio noise signal according to some embodiments of the present disclosure.

[0024] While the disclosure is amenable to various modifications and alternative forms, specifics thereof have been shown by way of example in the drawings and will be described in detail. It should be understood, however, that the intention is not to limit the application to the particular embodiments described. On the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the scope of the disclosure. DETAILED DESCRIPTION

[0025] It will be readily understood that the components of the embodiments, as generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of devices, systems, methods and computer program products of the present embodiments as presented is not intended to limit the scope of the claimed embodiments, but is merely representative of selected embodiments. Additionally, it should be appreciated that, although specific embodiments have been described herein for illustrative purposes, various modifications and changes in light thereof will be obvious to those skilled in the art.

[0026] Reference throughout this specification to "one implementation", "at least one implementation", "one embodiment", "another embodiment", "other embodiments", or "embodiments" and similar language means that a particular feature, structure, or characteristic described in connection with the implementation is included in at least one implementation. Thus, appearances of the phrase "one implementation", "at least one implementation", "in one implementation", "in another implementation", "in other implementations", or "in embodiments" in various places throughout this specification are not necessarily referring to the same implementation.

[0027] The embodiments shown will be best understood by reference to the drawings, wherein like parts are designated with like numerals throughout. The following description is merely intended to exemplify certain selected embodiments in conformance with the principles of the embodiments claimed herein.

[0028] It should be appreciated that, although the present disclosure includes a detailed description of cloud computing, implementation of the teachings recited herein are not limited to a cloud computing environment. Rather, embodiments of the present disclosure are capable of implementation in conjunction with any other type of computing environment now known or later developed.

[0029] Cloud computing is a model of service delivery for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, network bandwidth, servers, processing, memory, storage, applications, virtual machines, and services) that can be rapidly provisioned and released with minimal management effort or interaction with a provider of the service. This cloud model can include at least five characteristics, at least three service models, and at least four deployment models.

[0030] The characteristics are as follows.

[0031] On-demand self-service: cloud consumers can unilaterally provision computing capabilities, such as server time and network storage, as needed automatically without requiring human interaction with the service's provider.

[0032] Broad network access: capabilities are available over a network and accessed through standard mechanisms that promote use by heterogeneous thin or thick client platforms (e.g., mobile phones, laptops, and PDAs).

[0033] Resource pooling: the provider's computing resources are pooled to serve multiple consumers using a multi-tenant model, with different physical and virtual resources dynamically assigned and reassigned according to demand. There is a sense of location independence in that the consumer generally has no control or knowledge over the exact location of the provided resources but can be able to specify location at a higher level of abstraction (e.g., country, state, or datacenter).

[0034] Rapid elasticity: capabilities can be rapidly and elastically provisioned, in some cases automatically, frequently with little or no management effort or service provider interaction. For consumers, the capacity available for provisioning is generally unbounded.

[0035] Measured service: cloud systems automatically control and optimize resource use by leveraging a metering capability at some level of abstraction appropriate to the type of service (e.g., storage, processing, bandwidth, and active user accounts). Resource usage can be monitored, controlled, and reported providing transparency for both the provider and consumer of the service.

[0036] Service models are as follows.

[0037] Software as a Service (SaaS): the capability provided to the consumer is to use the provider's applications running on a cloud infrastructure. The applications are accessible from various client devices through a thin client interface such as a web browser (e.g., web-based e-mail). The consumer does not manage or control the underlying cloud infrastructure including network, servers, operating systems, storage, or even individual application capabilities, with the possible exception of limited user-specific application configuration settings.

[0038] Platform as a Service (PaaS): the capability provided to the consumer is to deploy onto the cloud infrastructure consumer-created or acquired applications created using programming languages and tools supported by the provider. The consumer does not manage or control the underlying cloud infrastructure including networks, servers, operating systems, or storage, but has control over the deployed applications and possibly application hosting environment configurations.

[0039] Infrastructure as a Service (laaS): the capability provided to the consumer is to provision processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software, which can include operating systems and applications. The consumer does not manage or control the underlying cloud infrastructure but has control over operating systems, storage, deployed applications, and possibly limited control of select networking components (e.g., host firewalls).

[0040] Deployment models are as follows.

[0041] Private cloud: the cloud infrastructure is operated solely for an organization. It can be managed by the organization or a third party and can exist on-premises or off-premises.

[0042] Community cloud: the cloud infrastructure is shared by several organizations and supports mission-oriented

[0043] Public cloud: the cloud infrastructure is made available to the general public or a large industry group and is owned by an organization selling cloud services.

[0044] Hybrid cloud: the cloud infrastructure is a composition of two or more clouds (private, community, or public) that remain unique entities but are bound together, creating the hybrid cloud.

[0045] A cloud computing environment is service-oriented, centrally focused on stateless, loosely-coupled, modular, and semantic interoperability. At the core of cloud computing is an infrastructure comprising a network of interconnected nodes.

[0046] Referring now to the drawing, in which is depicted an illustrative cloud computing Figure 1 environment 50. As shown, cloud computing environment 50 includes one or more cloud computing nodes 10 with which local computing devices used by cloud consumers, such as, for example, personal digital assistant (PDA) or cellular telephone 54A, desktop computer 54B, laptop computer 54C, and / or automobile computer system 54N can communicate. Nodes 10 can communicate with one another. They can be grouped (not shown) physically or virtually, in one or more networks, such as Private, Community, Public, or Hybrid clouds as described hereinabove, or a combination thereof. This allows cloud computing environment 50 to offer infrastructure, platforms and / or software as services with Figure 1 The types of computing devices 54A-N shown in FIG. 54 are intended to be illustrative only and computing nodes 10 and cloud computing environment 50 can communicate with any type of computerized devices over any type of network and / or network addressable connection (e.g., using a web browser).

[0047] Referring now to FIG. 1 A, Figure 2 , a set of functional abstraction layers are provided by cloud computing environment 50 Figure 1 It should be understood that any number of components, layers and functions can be provided. Figure 2 The components, layers and functions shown in

[0048] As described, the following layers and corresponding functions are provided:

[0049] Hardware and software layer 60 includes hardware and software components. Examples of hardware components include: mainframes 61; RISC (Reduced Instruction Set Computer) architecture based servers 62; servers 63; blade servers 64; storage devices 65; and networks and networking components 66. In some embodiments, software components include network application server software 67 and database software 68.

[0050] Virtualization layer 70 provides an abstraction layer from which the following examples of virtual entities can be provided: virtual servers 71; virtual storage 72; virtual networks 73, including virtual private networks; virtual applications and operating systems 74; and virtual clients 75.

[0051] In one example, management layer 80 can provide the functions described below. Resource provisioning 81 provides dynamic procurement of computing resources and other resources that are utilized to perform tasks within the cloud computing environment. Metering and Pricing 82 provide cost tracking as resources are utilized within the cloud computing environment, and billing or invoicing for consumption of these resources. In one example, these resources can include application software licenses. Security provides identity verification for cloud consumers and tasks, as well as protection for data and other resources. User portal 83 provides access to the cloud computing environment for consumers and system administrators. Service level management 84 provides cloud computing resource allocation and management such that required service levels are met. Service Level Agreement (SLA) planning and fulfillment 85 provide pre-arrangement for, and procurement of, cloud computing resources for which a future requirement is anticipated in accordance with an SLA.

[0052] Workloads layer 90 provides examples of functionality for which the cloud computing environment can be utilized. Examples of workloads and functions which can be provided from this layer include: mapping and navigation 91; software development and lifecycle management 92; virtual classroom education delivery 93; data analytics processing 94; transaction processing 95; and adversarial reinforcement for queries for automated responses 96.

[0053] Referring to Figure 3, a block diagram of an example data processing system, referred to herein as computer system 100, is provided. System 100 can be embodied in a computer system / server in a single location, or in at least one embodiment, can be configured in a cloud-based system sharing computing resources. For example, and without limitation, computer system 100 can be used as a cloud computing node 10.

[0054] Aspects of computer system 100 can be embodied in a computer system / server in a single location, or in at least one embodiment, can be configured in a cloud-based system sharing computing resources as a cloud-based support system to implement the systems, tools and processes described herein. Computer system 100 is operable with many other general purpose or special purpose computer system environments or configurations. Examples of well-known computer systems, environments, and / or configurations that can be suitable for use with computer system 100 include, but are not limited to, personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputer systems, mainframe computer systems, and file systems including any of the above systems, devices, and their equivalents (e.g., distributed storage environments and distributed cloud computing environments).

[0055] Computer system 100 can be described in the general context of computer system-executable instructions, such as program modules, being executed by computer system 100. Generally, program modules can include routines, programs, objects, components, logic, data structures, and so on that perform particular tasks or implement particular abstract data types. Computer system 100 can be practiced in a distributed cloud computing environment where task are performed by a remote processing device that is linked through a communications network. In a distributed cloud computing environment, program modules can be located in both local and remote computer system storage media including memory storage devices.

[0056] As Figure 3As shown, computer system 100 is in the form of a general-purpose computing device. The components of computer system 100 can include, without limitation, one or more processors or processing devices 104 (sometimes referred to as processors and processing units), such as a hardware processor, system memory 106 (sometimes referred to as a memory device), and a communication bus 102 that couples various system components including system memory 106 to processing device 104. Communication bus 102 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, and without limitation, such architectures include Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA (EISA) bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus. Computer system 100 typically includes a variety of computer system readable media. Such media can be any available media that is accessible by computer system 100 and includes both volatile and non-volatile media, removable and non-removable media. Further, computer system 100 can include one or more persistent storage devices 108, communication units 110, input / output (I / O) units 112, and displays 114.

[0057] Processing device 104 is configured to execute instructions that can be loaded into system memory 106. Depending upon the particular implementation, processing device 104 can be a number of processors, a multi-core processor, or some other type of processor. As used herein with reference to items, a plurality means one or more items. Further, processing device 104 can be implemented using a multiple heterogeneous processor system, where a primary processor exists on a single chip with a secondary processor. As another illustrative example, processing device 104 can be a symmetric multi-processor system containing a plurality of the same type of processors.

[0058] System memory 106 and persistent storage 108 are examples of storage devices 116. A storage device can be any piece of hardware that is capable of storing information (e.g., data, program code in functional form, and / or other appropriate information) either temporarily or permanently. In these examples, system memory 106 can be, for example, a random access memory or any other suitable volatile or non-volatile storage device. System memory 106 can include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory.

[0059] The persistent storage 108 can take different forms depending on the particular implementation. For example, the persistent storage 108 can include one or more components or devices. For example, but not limited to, the persistent storage 108 can be provided for reading from and writing to non-removable, non- volatile magnetic media (not shown and typically called a "hard drive"). Although not shown, a magnetic disk drive for reading from and writing to a removable, non- volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading from or writing to a removable, non- volatile optical disk (e.g., a CD-ROM, DVD-ROM or other optical media) can be provided. In such cases, each can be connected to the communication bus 102 by one or more data media interfaces.

[0060] In these examples, the communication unit 110 can provide communication with other computer systems or devices. In these examples, the communication unit 110 is a network interface card. The communication unit 110 can provide communication using either or both physical and wireless communication links.

[0061] The input / output unit 112 can allow input and output of data to and from other devices that can be connected to the computer system 100. For example, the input / output unit 112 can provide a connection for user input through a keyboard, mouse, and / or some other suitable input device. Further, the input / output unit 112 can send output to a printer. The display 114 can provide a mechanism to display information to a user. Examples of input / output units 112 that facilitate the establishment of communications between the various devices within the computer system 100 include, but are not limited to, network cards, modems, and input / output interface cards. Further, the computer system 100 can communicate with one or more networks, such as a local area network (LAN), a general wide area network (WAN), and / or a public network, such as the Internet, via a network adapter (not shown in FIG. 1). Figure 3 It should be appreciated that, although not shown, other hardware and / or software components could be used in conjunction with the computer system 100. Examples of such components include, but are not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data archival storage systems.

[0062] Instructions for the operating system, applications, and / or programs can be located in storage device 116, which is in communication with processing device 104 via communication bus 102. In these illustrative examples, the instructions are in a functional form on permanent memory 108. These instructions can be loaded into system memory 106 for execution by processing device 104. The processes of the different embodiments can be performed by processing device 104 using computer- implemented instructions, which can be located in a memory, such as system memory 106. These instructions are referred to as program code, computer-usable program code, or computer-readable program code that can be read and executed by a processor in processing device 104. The program code in the different embodiments can be embodied on different physical or tangible computer-readable media, such as system memory 106 or permanent memory 108.

[0063] Program code 118 can be in a functional form on a removable computer- readable media 120 and can be loaded or transferred to computer system 100 for execution by processing device 104. In these examples, program code 118 and computer- readable media 120 can form a computer program product 122. In one example, computer-readable media 120 can be a computer-readable storage media 124 or a computer-readable signal media 126. Computer-readable storage media 124 can include, for example, an optical or magnetic disk that is inserted into or placed into a drive or other device that is part of permanent memory 108 for transfer onto a storage device (such as a hard drive) that is part of permanent memory 108. Computer-readable storage media 124 can also take the form of permanent memory that is connected to computer system 100, such as a hard drive, thumb drive, or flash memory. In some instances, computer-readable storage media 124 can not be removable from computer system 100.

[0064] Alternatively, program code 118 can be transferred to computer system 100 using computer-readable signal media 126. Computer-readable signal media 126 can be, for example, a propagated data signal containing program code 118. For example, computer-readable signal media 126 can be an electromagnetic signal, an optical signal, and / or any other suitable type of signal. These signals can be transmitted over communication links, such as wireless communication links, optical fiber cables, coaxial cables, wire, and / or any other suitable type of communications links. In other words, in illustrative examples, communication links and / or connections can be physical or wireless.

[0065] In certain illustrative embodiments, program code 118 can be downloaded to persistent storage 108 from another device or computer system through a computer readable signal medium 126 for use within computer system 100. For instance, program code stored in the computer readable storage medium in a server computer system can be downloaded over a network to computer system 100. The computer system providing program code 118 can be a server computer, a client computer, or some other device capable of storing and transmitting program code 118.

[0066] Program code 118 can include one or more programs modules (not shown in FIG. 1) and be implemented in software and / or firmware, including one or more programs, applications, applets, app, or other modules, which can be stored on computer readable storage medium 124, such as computer readable storage media 104, 106, and 108. The one or more programs modules can include, by way of example and not limitation, an operating system, one or more applications, other program modules, and program data. Figure 3 The different components exhibited for computer system 100 are not meant to provide architectural limitations to the manner in which different embodiments can be implemented. The different illustrative embodiments can be implemented in a computer system including components in addition to and / or in place of those illustrated for computer system 100. Some of the different illustrative embodiments can be implemented by computer system 100.

[0067] The different components exhibited for computer system 100 are not meant to provide architectural limitations to the manner in which different embodiments can be implemented. The different illustrative embodiments can be implemented in a computer system including components in addition to and / or in place of those illustrated for computer system 100. Some of the different illustrative embodiments can be implemented by computer system 100.

[0068] The present disclosure can be a system, a method, and / or a computer program product at any possible technical detail level of integration. The computer program product can include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present disclosure.

[0069] A computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, semiconductor, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer readable storage media include the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device such as punch-cards or raised structures in a groove having instructions recorded thereon, and any suitable combination of the foregoing. A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.

[0070] Computer readable program instructions described herein can be downloaded to respective computing / processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and / or a wireless network. The network can comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing / processing device.

[0071] Computer readable program instructions described herein can be downloaded to respective computing / processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and / or a wireless network. The network can comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing / processing device.

[0072] Computer readable program instructions for carrying out operations of the present disclosure can be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, configuration data for an integrated circuit, or source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like, and a procedural programming language such as the "C" programming language or the like. The computer readable program instructions can execute entirely on a user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate array (FPGA), or programmable logic array (PLA) can execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present disclosure.

[0073] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0074] These computer readable program instructions can be provided to a processor of a computer, or other programmable data processing apparatus, to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer readable program instructions can also be stored in a computer readable storage medium that can include, without limitation, random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other data storage technology. When the computer readable program instructions are executed by the computer, other programmable data processing apparatus, or other devices, a series of operational steps are performed.

[0075] These computer readable program instructions can be provided to a computer, other programmable data processing apparatus, or other device to produce a machine, such that the instructions, which execute via the computer's, other programmable data processing apparatus, or other device's processors, implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0076] The flow diagrams and the block diagrams in the drawings are meant only to illustrate ways in which systems, methods, and computer program products according to the present disclosure can be implemented. Other ways to implement the present disclosure are possible. There can be many other items that are not depicted in the drawings. The present disclosure is not limited to the implementation depicted in the drawings. The intent is to include all changes that equally effect the spirit of the underlying technological principles of the present disclosure.

[0077] Many known examination events occur through online means, where the examinee is located remotely and it is impractical for the examining officer to be physically present. With the continued ubiquitous presence of the Internet in our society, access to information has become more readily available, and the Internet serves as the primary destination for answers to queries. For example, answers to queries can be obtained quite conveniently through well-known and nearly ubiquitous artificial intelligence (AI) agents (e.g., digital assistants and chatbots, which are readily accessible through all forms of computing devices, including mobile devices). One example is the use of reverse image search engines to respond to image-based examination queries. Thus, the possibility of examinations that utilize such technology to generate automated responses to examination queries requires some form of remote examination process to maintain the integrity of the examination process. At least some known methods of remote management and processing of online examinations include intrusive monitoring, such as video-based monitoring of the examinee and associated screen and screen sharing. Both methods require the full, undivided, and uninterrupted attention of a human monitoring personnel. Further, both methods use additional computing resources that can impact the performance of the examination hardware and software, which can be burdensome for timed examinations. Further, both methods increase the cost of the examination by using human monitoring personnel and additional hardware and software. At least some known methods of content manipulation that address, at least in part, the use of AI agents include watermarking of text and picture content, and adding classic adversarial noise to audio content, however, these methods are weighted to preserve semantic content, where the modifications are constrained to be imperceptible, and the recognition of AI agents is at most slightly suppressed. Further, when designing queries to examinations, the potential use of AI agents by the examinee can limit or exclude many forms of questions, e.g., the recognition of music, images, or text passages.

[0078] Disclosed and described herein is a system, computer program product, and method directed to leveraging adversarial machine learning to facilitate the design of examination queries that cannot be easily solved by artificial intelligence (AI) agents, thereby augmenting parameters of test sequences to exclusively test actual human examinees. In at least some embodiments, the delivery of queries to examinees is reconfigured, including manipulating the content of the queries, to make it substantially more difficult, if not outright prohibitive, to use any AI agent to generate automated examination query responses. The degree to which queries can be manipulated is constrained to preserve semantic content, i.e., to allow the examinee to fully understand the query. However, unlike pure watermarking and classic adversarial noise insertion, where modifications must be constrained to be imperceptible to the examinee, the present disclosure describes clear and perceptible modifications to the query content. Specifically, as opposed to pure steganography that attempts to completely hide information while ensuring that the content is recoverable, the present disclosure describes stripping at least some non-essential information from the query, however, without allowing complete obfuscation. In some embodiments, the content manipulation is multi-modal, i.e., the modifications to the query are not limited to maintaining the same modality. For example, a textual description can be converted to audio, followed by the addition of adversarial noise.

[0079] Reference is made to Figure 4 A block diagram illustrating a computer system, namely, an examination query automation response adversarial augmentation system 400 (referred to herein as system 400) configured to manage examinations with adversarial augmentation of queries for automated responses is presented. The system 400 includes one or more processing devices 404 (only one shown) communicatively and operatively coupled to one or more memory devices 406 (only one shown). The system 400 also includes a data storage system 408 communicatively coupled to the processing device 404 and the memory device 406 by a communications bus 402. In one or more implementations, the communications bus 402, the processing device 404, the memory device 406, and the data storage system 408 are similar to their counterparts shown in FIG. 1 (i.e., the communications bus 102, the processing device 104, the system memory 106, and the persistent storage device 108), respectively. The system 400 further includes one or more input devices 410 and one or more output devices 412 communicatively coupled to the communications bus 402. In addition, the system 400 includes one or more internet connections 414 (only one shown) having one or more artificial intelligence (AI) agents 416 (only one shown). Figure 3

[0080] In one or more embodiments, an adversarial automated response augmentation engine 420 (referred to herein as engine 420) resides within the memory device 406. The engine 420 includes an examination module 422, a filter / edit module 424, a semantic service module 426, and an AI agent communication module 428. Reference will be made to Figure 4 ​These modules are discussed further with respect to FIG. 7. The stored data 430 is maintained in the data storage system 408 for access by the memory device 406.

[0081] In one or more embodiments, the actual implementation of the system 400 (including the engine 420) is a cloud service as described herein with respect to Figure 1 and Figure 2 As a cloud service, the engine 420 can reside within any computing device within a cloud-based infrastructure to deliver the services described herein. Generally, no part of the engine 420 resides on the examined device, i.e., the examined device is configured to receive the final version of the modified original examination query (as discussed further herein) and provide non-automated responses from the examination. Thus, in such embodiments, the cloud service receives the original query and one or more acceptable responses from its author through the input device 410. The cloud-based service uses the engine 420 to modify the original query. The modified query can be returned to the author through the output device 412 for its manual examination, where if the author is acceptable, the modified query can be stored on the author’s computing device or within the data storage system 408 as part of the stored data 430. In some embodiments, the modified query can be transmitted directly to one or more examinees through the output device 412. In some embodiments, the engine 420 is a static algorithm, where customization is at most minimal. In some embodiments, the engine 420 can be more flexibly customized by the author. Thus, only the bandwidth required to deliver the modified query and vetted responses is needed.

[0082] In at least one embodiment, the actual implementation of the system 400 (including the engine 420) is a standalone system embedded within a computing system directly accessible to the author, e.g., the author’s personal or employer-provided computing system, including but not limited to the desktop computer 54B and the laptop computer 54C (shown in Figure 1 and the server 63 (shown in Figure 2 ). In such embodiments, as with the cloud-based implementation described above, the examined device should not receive or store anything other than the final form modified query.

[0083] Referring to Figure 5A , a flowchart is provided that illustrates a process 500 for managing an examination by adversarial reinforcement of a query against automated responses. Further, reference is made to Figure 4In embodiments, the engine 420 electronically receives 502 an original query, where the examinee is expected to electronically submit a response as part of a remote examination process. The original examination query is authored by an administrator of the examination and is configured in the form of one or more of an image, one or more text passages, and an audio clip. The author of the original query can be concerned that the examinee can use an AI agent to generate an automated response. For example, a commercially available AI agent, such as an image search engine or a digital assistant, can respond to such a query that provides, for example, an image caption or a sentence that describes information sought in the examination query. Accordingly, in embodiments, the author can utilize the engine 420 to modify the original query by transmitting the original query and a response to the engine 420 via the input device 410 prior to delivery to one or more examinees.

[0084] The original query is transmitted to the examination module 422 in the engine 420. As part of an iterative cycle, as described further herein, the examination module 422 is configured to perform the receiving operation 502 of the original query and subsequently modified queries. The examination module 422 is further configured to, as part of the receiving operation, examine the incoming query for any adversarial features provided as described herein and then initiate the iterative cycle when it does not find any. Additionally, if at least some adversarial mechanisms are determined with respect to the current examination query, the examination module 422 also determines whether the adversarial features in the original query are fully satisfied or not fully satisfied. Further, when performing the iterative cycle that converts the original query into a final modified query, the examination module 422 facilitates continuation of the iterative cycle until a criterion for delivering the modified query to an examinee is satisfied.

[0085] As such, in embodiments, the original query is modified 504, generating an at least partially modified query. To transmit the modified query to the examinee, the at least partially modified query will be configured to satisfy two requirements, namely, the final modified query must be understood by a human examinee and the final query must not be able to be responded to by an electronic device without direct human support, i.e., the examinee must directly respond to the modified query without the aid of an AI agent. As such, the original query is converted 506 into an at least partially modified query. More specifically, the original query is transmitted to one or more filtering components, namely, the filtering / editing module 424 that is at least part of the original query to final modified query iterative cycle that includes the query conversion operation 506.

[0086] In one or more embodiments, the filtering / editing module 424 electronically receives the original query, where the original query includes one or a combination of modalities of an original textual query, an original image query, and an original audio query. The original query can be transformed 506 by one or more operations having predetermined parameters. For example, an image can be blurred, where the degree of blurring is governed according to established criteria and parameters. Further, one or more aspects of the image can undergo one or more perturbations or manipulations. Further, for example, an image can be manipulated by adding adversarial noise to at least partially obscure one or more features of the image. Reference is made to Figure 6A to Figure 6D Further discussion of image modification. Audio modification can include perturbations or manipulations with effects including, but not limited to, playing back a corresponding audio file in reverse, adding echo, removing certain frequency ranges, adding background noise, breaking into small pieces and rearranging, and reversing lyrics. Reference is made to Figure 7A to Figure 7D Further discussion of audio modification. A textual query can undergo a text-to-audio conversion (also shown in Figure 7A-7D ) or a text-to-image conversion, thereby transforming the query by transforming the modality of the query while maintaining the essence of the original query. For text-to-audio conversion, the audio perturbations and manipulations described previously can be used. Further, text can be converted into one or more images, where the text can be converted into images that are non-colorable and non-occludable. In addition to scrambling the text, in some embodiments, random words can be added to or removed from every other sentence. As such, the transformation operation 506 is performed at least in part as a function of the modality of the original inspection query.

[0087] In at least some embodiments, the at least partially modified query and the original query are transmitted 510 to the semantic services module 426. The semantic services module 426 is configured to perform 512 a similarity assessment between the original query and the at least partially modified query to determine a semantic similarity. Reference is made to Figure 5B , providing a continuation of the process 500 from Figure 5A Reference is also made to Figure 4And more specifically, the semantic services module 426 is configured to perform a determination operation 514 to decide whether the conversion thus far has preserved information sufficient for a human subject to understand. For example, the conversion operation 506 can modulate the audio query to remove at least a portion of the audio file; however, if the query includes voice recognition and assignment to a particular individual, then the addition of background noise can be implemented, and the semantic services module 426 will determine 514 whether the amplitude of the background noise frequency drowns out the portion of the audio query to be analyzed by the subject. In some embodiments, the stored data 430 can include examples of modified queries that are human understandable and human non-understandable. In some embodiments, the semantic services module 426 ensures that the information contained in the edited query remains sufficiently unchanged to facilitate human recognition. Such services can take various forms from human annotators to queries trained through large databases. A "yes" response to the determination operation 514 will be discussed further below. A "no" response to the determination operation 514 causes the transmission 516 of the at least partially modified query to the filtering / editing module 424, where the at least partially modified query is further converted 518. The process 500 returns to the similarity assessment performance operation 512, and the further converted query undergoes an iterative loop through operations 512-518 until a "yes" result is obtained for the determination operation 514. Further, feedback from the similarity assessment operation 512 for determining semantic similarity between the original query and the at least partially modified query, as well as the determination operation 514, are conveyed to the further conversion operation 518.

[0088] Further, in at least some embodiments, the at least partially modified query is transmitted 520 to a predetermined AI agent 416 via the AI agent communication module 418 and the internet connection 414. In some embodiments, the AI agent 416 is external to the engine 420. In some embodiments, one or more AI agents 416 are included within the engine 420 or reside within one or more of the memory device 406 and the data storage system 408. The engine 420 is agnostic to the nature of the AI agent 416. The AI agent 416 is configured to perform 522 a attempted labeling operation on the at least partially modified query via one or more machine learning models that are internal to the respective AI agent. The AI agent 416 returns the results of the attempted labeling operation 522 to the AI agent communication module 428, which is further configured to perform a determination operation 524 to determine whether the at least partially modified query can be responded to via electronic means without human support (i.e., a human response). If the at least partially modified query is not labeled, i.e., the AI agent 416 fails to assign a label to the at least partially modified query, the result of the determination operation 524 is “no” and a “no” response to the determination operation 524 will be discussed further below. If the at least partially modified query is labeled, the result of the determination operation 524 is “yes”. A “yes” response to the determination operation 524 results in the transmission 526 of the at least partially modified query to the filtering / editing module 424, where the at least partially modified query is further transformed 518.

[0089] Further, in some embodiments, the AI agent 416, or in some embodiments, the AI agent communication module 428, assigns a confidence score to the response from the AI agent 416. In some embodiments, the confidence score is a numerical value along a 0% to 100% scale, where the confidence score at least partially indicates the confidence that the AI agent 416 has correctly identified and labeled the at least partially modified query.

[0090] Further, in some embodiments, the AI agent 416 can be prompted to provide a response to the query. A determination operation 524 can also be performed to determine whether the response provided by the AI agent is sufficiently close to the check query answer (response) provided by the author in the receiving operation 502. If the AI agent 416 can provide a sufficiently accurate response, then the result from the determination operation 524 is “yes”. If the AI agent 416 cannot provide a sufficiently accurate response, including an incorrect response, then the result from the determination operation is “no”. The process 500 returns to the attempted labeling performing operation 522 and the further transformed query undergoes the operations 520-524-518 in an iterative loop until a “no” result from the determination operation 524 is obtained. Further, feedback from the attempted labeling operation 522 and the determination operation 524 is transmitted to the further transforming operation 518.

[0091] As discussed, there are two iterations or loops being processed. The first iteration loop includes operations 512-514-516 performed by the semantic service module 426 to generate a final modified examination query that is understandable by a human subject. The second iteration loop includes operations 522-524-526 performed by the AI agent communication module 428 and the AI agent 416 to generate a final modified examination query that cannot be responded to by the AI agent 416. In some embodiments, the two iteration loops can be performed in parallel. In some implementations, the two iteration loops can be performed in series in a loop including a first loop - a second loop - a first loop, etc. until both requirements of the determining operations 514 and 524 are achieved. In some embodiments, the two iteration loops can be performed in a manner where a satisfactory modified query is produced in one of the two loops (which in turn is provided to the other loop).

[0092] Reference is made to Figure 5C , a continuation of the process 500 from Figure 5B . With continued reference to Figure 4 , Figure 5A and Figure 5B , in those implementations where the result from the determining operation 514 is “yes” and the result from the determining operation 524 is “no”, the final modified query and result are transmitted to a merging operation 530 associated with the examination module 422. The result from the merging operation 530 results in the fully modified query being transmitted 532 to the human subject by the examination module 422 and the output device 412 and the process 500 ends. In other embodiments, where a “no” result from the determining operation 514 or a “yes” result from the determining operation 524 would facilitate the examination module 422 from preventing the modified query from being transmitted to the human subject and further modifications to the query are performed as described above.

[0093] An example of modifying an image-based examination query is presented. Reference is made to Figure 6A , a picture image is presented showing an example original image 600 of a turtle associated with an original examination query. The turtle image 600 is shown in black and white; however, a color image can also be presented as the original query. Reference is made to Figure 6B , the original image 600 of Figure 6A is provided showing an example of at least a partially modified image 610 associated with at least a partially modified examination query. Reference is also made to Figure 4 , Figure 5A , Figure 5B and Figure 5C, semantic services module 426 analyzes image 610 and determines that the result from determination operation 514 is “yes” because the edges of the features in image 610 are preserved and enough information in image 610 is preserved to allow a human to identify a turtle. However, the result from determination operation 524 is “yes” in which AI agent 416 correctly identifies (labels) image 610 as a turtle with a 99% confidence level. Thus, as described above, the “yes” result from determination operation 514 and the “yes” result from determination operation 524 will prohibit examination module 422 from delivering the modified query to the human subject and further modification of the query is performed as described above.

[0094] Referring to Figure 6C , a picture of a turtle in Figure 6A is presented that exhibits an example further modified image 620 associated with a further modified examination query. Referring also to Figure 4 , Figure 5A , Figure 5B and Figure 5C , semantic services module 426 analyzes image 620 and determines that the result from determination operation 514 is “no” because substantially none of the features in image 600 are present in image 620 and human identification of a turtle is all but impossible. However, the result from determination operation 524 is “no” in which AI agent 416 is unable to identify (label) image 620 as a turtle and thus does not provide a confidence level. Thus, as described above, the “no” result from determination operation 514 and the “no” result from determination operation 524 will prohibit examination module 422 from delivering the modified query to the human subject and further modification of the query is performed as described above.

[0095] Referring to Figure 6D , a picture of a turtle in Figure 6A is presented that exhibits an example of a final modified image 630 associated with a modified examination query. Semantic services module 426 analyzes image 630 and determines that the result from determination operation 514 is “yes” because the edges of the features in image 630 are preserved and enough information in image 630 is preserved to allow a human to identify a turtle. The result from determination operation 524 is “no” in which AI agent 416 is unable to identify (label) image 630 as a turtle and thus does not provide a confidence level. As described above, the “yes” result from determination operation 514 and the “no” result from determination operation 524 allow examination module 422 to transmit 532 the final modified query and result to merge operation 530 to the human subject and process 500 ends.

[0096] Therefore, the inspection module 422 checks a simple AND condition through the merge operation 530: whether the semantic service module 426 considers that the semantic features of the modified query are sufficiently preserved to remain human-understandable, and whether the AI ​​agent 416 gives an incorrect, substantially ambiguous, or no answer (completely disabled). If both conditions are met, the inspection module 422 interrupts the loop and transmits the fully modified query to the inspector 532; otherwise, the loop continues.

[0097] This presents an example of modifying a text-to-audio based inspection query. References Figure 7A This provides a text representation of the example text check query and the 700 answer / response. (Reference) Figure 7B Provided based on Figure 7A A block diagram of an example audio inspection query 710 for the text inspection query 700. The semantic service module 426 will determine the audio inspection query 710 in a way that a human subject can understand. Furthermore, the AI ​​agent 416 will process the audio inspection query 710 and provide a correct response with a 99% confidence value. Therefore, the two requirements for the inspection query to be modified and transmitted to the subject, as allowed by the inspection module 422, are not met. (Reference) Figure 7C It provides an illustration based on from Figure 7A The example text inspection query 700 is derived from at least partially modified text inspection query 722, and a block diagram of an example edited and converted audio query 720. The edited and converted audio query 720 is substantially incomprehensible to both humans and the AI ​​agent 416, thus rendering the edited and converted audio query 720 untransmittable to human subjects by the inspection module 422. Reference Figure 7D The document provides an illustration of an audio noise signal 732 with added specific frequencies and amplitudes. Figure 7A A block diagram of an example audio inspection query 730 for text inspection query 700. The semantic service module 426 is able to identify the converted audio through noise 732, and the AI ​​agent 416 cannot mark the noisy converted audio query 730, thus satisfying the two requirements of the inspection module 422, and the noisy converted audio query 730 can be transmitted 532 to the human subject for inspection.

[0098] In some embodiments, there can be opportunities where neither the semantic service module 426 nor the AI agent 416 results can be satisfied, i.e., only one of the two results from the determine operations 514 and 524 are satisfied to meet the requirements of the check module 422, regardless of the number of iterations. For such conditions, the engine 420 includes sufficient logic of an iteration counter and stopping the iteration operations at a predetermined number of iterations. Once the loop iteration is stopped, a notification can be provided to the author, who will make a decision on how to proceed, including but not limited to eliminating the check query from the check library, manually changing the query by trial and error, or using the latest iteration of the modified query, with the expectation that completely eliminating the potential use of the AI agent 416 for the respective query is not feasible.

[0099] Further, in some embodiments, the original query can be configured such that both requirements of the check module 422 are satisfied without any modification. In such embodiments, the filter / edit module 424 can be configured to perform the first iteration with a “do nothing” command, such that the unmodified check query will trigger a “yes” result from the determine operation 514 and a “no” result from the determine operation 524, terminating the iteration loop, and triggering the merge operation 530 to allow the unmodified check query to be transmitted 532 to the examinee.

[0100] The systems, computer program products, and methods as disclosed herein help overcome the shortcomings and limitations of known systems and methods for conducting remote checks on individuals requiring examinee access to the internet. In particular, the present disclosure describes an automated process and system for generating queries from regular queries that are obfuscated in a manner that tests the cognitive of human checks but avoids the current capabilities of AI agents. The testing mechanisms described herein facilitate reducing the need for extensive remote processes, thereby reducing the need for intrusive and resource-intensive computing environments, including establishing and maintaining stable, high-bandwidth communication links to be established. Further, many previous limitations on the modality of testing can be lifted, including the use of music, images, and text passages. Thus, a significant improvement in the behavior and integrity of known remote check systems is achieved by the present disclosure.

[0101] The description of the different embodiments of the present disclosure has been presented for purposes of illustration but is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope of the described embodiments. The terminology used herein was chosen to best explain the principles of the embodiments, the practical application or technical improvement over technologies found in the marketplace, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A computer system, comprising: One or more processing devices and at least one memory device operatively coupled to the one or more processing devices, the one or more processing devices being configured to: The original query is received electronically, wherein the response to the original query is submitted electronically by a human. The original query is modified to generate a modified query, wherein one or more features of the modified query are preserved, and the modified query is understandable to the human. The modified query is transmitted to one or more artificial intelligence (AI) agents; The modified query is attempted to be marked by one or more AI agents; as well as Based on the fact that the attempted tagging operation did not assign a label to the modified query, it was determined that the modified query could not be responded to by an electronic device without human support, and based on the fact that the attempted tagging operation assigned a label to the modified query, it was determined that the modified query could be responded to by an electronic device without human support.

2. The computer system according to claim 1, further comprising: One or more filtering components communicatively coupled to the one or more processing devices, wherein the one or more processing devices are further configured to: The original query is transmitted to one or more filtering components, thereby performing one or more transformations on the original query.

3. The computer system according to claim 2, further comprising: One or more semantic services communicatively coupled to the one or more processing devices, wherein the one or more processing devices are further configured to: Transmit the modified query to the one or more semantic services; Transmit the original query to one or more semantic services; and The similarity assessment between the original query and the modified query is performed using one or more semantic services.

4. The computer system according to claim 3, wherein, The one or more processing devices are further configured to: The modified query is determined to be human-understandable through one or more semantic services.

5. The computer system according to claim 1, wherein, The one or more processing devices are also configured to perform one or more of the following: The modified query is assigned a label by one or more AI agents; The confidence value is assigned to the modified query by the one or more AI agents; as well as The AI ​​agent generates a response to the modified query.

6. The computer system according to claim 1, wherein, The one or more processing devices are also configured to perform one or more of the following: Failed to assign a label to the modified query through the one or more AI agents; The one or more AI agents do not generate a response to the modified query; and Incorrect responses to the modified query are generated by one or more AI agents.

7. The computer system according to claim 1, wherein: The original query includes one or more of the following modalities: Raw text query; Original image query; and Original audio query; and The one or more processing devices are further configured to modify the original query, including modifying the modality of the original query, including converting the original text query into one or more of the following: The audio query is at least partially modified, and the audio query includes at least some adversarial noise; as well as Image queries that are at least partially modified.

8. The computer system according to claim 1, further comprising: One or more filter components communicatively coupled to the one or more processing devices; One or more semantic services communicatively coupled to the one or more filtering components; Access to the one or more AI agents communicatively coupled to the one or more filtering components, wherein the one or more processing devices are further configured to: The original query is transformed into a query that is at least partially modified through one or more filtering components; The original query and the at least partially modified query are iteratively transmitted to the one or more semantic services; The at least partially modified query is iteratively transmitted to the one or more AI agents; The at least partially modified query is iteratively transmitted to the one or more filtering components; Iteratively determine: Through the one or more semantic services, the human's ability to understand the at least partially modified query; and Through the one or more AI agents, it is impossible to respond to the at least partially modified query via the electronic device without the human support, thereby establishing that the at least partially modified query is a fully modified query; and The completely modified query is transmitted to the human.

9. A computer program product, the computer program product comprising: One or more computer-readable storage media; as well as Program instructions stored on the one or more computer-readable storage media, the program instructions comprising: The program instructions for receiving the original query electronically, wherein the response to the original query is to be submitted electronically by a human. The program instructions modify the original query to generate a modified query, wherein one or more features of the modified query are preserved, and the modified query can be understood by the human. Transmit the modified query to one or more AI agents as program instructions; Program instructions to perform a flagging operation on the modified query through the one or more AI agents; and Based on the failed tagging operation of the attempted tagging operation, it was determined that the modified query cannot be responded to by an electronic device without human support, and based on the failed tagging operation, a program instruction was given to assign a tag to the modified query, determining that the modified query can be responded to by an electronic device without human support.

10. The computer program product according to claim 9, further comprising: Program instructions for converting the original query into a query that is at least partially modified by the one or more filtering components; Program instructions that iteratively transmit the original query and the at least partially modified query to the one or more semantic services; Iteratively transmit at least partially modified queries to program instructions of the one or more AI agents; Program instructions that iteratively transmit the at least partially modified query to the one or more filtering components; Program instructions used to iteratively determine: The comprehensibility of the at least partially modified query by the person through the one or more semantic services; as well as The one or more AI agents cannot respond to the at least partially modified query via the electronic device without the human support, thus establishing that the at least partially modified query is a fully modified query; as well as The completely modified query is transmitted to the person's program instructions.

11. A computer-implemented method, comprising: The original query is received electronically, wherein the response to the original query is submitted electronically by a human. The original query is modified to generate a modified query, wherein one or more features of the modified query are preserved, and the modified query is understandable to the human. The modified query is transmitted to one or more artificial intelligence (AI) agents; The modified query is attempted to be marked by one or more AI agents; as well as Based on the fact that the attempted tagging operation did not assign a label to the modified query, it was determined that the modified query could not be responded to by an electronic device without human support, and based on the fact that the attempted tagging operation assigned a label to the modified query, it was determined that the modified query could be responded to by an electronic device without human support.

12. The method according to claim 11, wherein, The generated modified queries include: The original query is transmitted to one or more filtering components to perform one or more transformations on the original query.

13. The method of claim 12, further comprising: Transmit the modified query to one or more semantic services; The original query is transmitted to one or more semantic services; as well as The similarity assessment between the original query and the modified query is performed using one or more semantic services.

14. The method according to claim 13, wherein, Performing the similarity assessment includes: The modified query is determined to be human-understandable through one or more semantic services.

15. The method of claim 11, wherein the flagging operation performed on the modified query comprises one or more of the following: Assign labels to the modified query; Assign a confidence value to the modified query; and Generate a response to the modified query.

16. The method according to claim 11, wherein, The flagging operation performed on the modified query includes one or more of the following: Failed to assign a label to the modified query; No response is generated for the modified query; and An incorrect response is generated to the modified query.

17. The method of claim 11, wherein: Receiving the original query electronically includes receiving the original query in one or more of the following modalities: Raw text query; Original image query; and Original audio query; and Modifying the original query includes modifying the modality of the original query, including converting the original text query into one or more of the following: The audio query is at least partially modified, and the audio query includes at least some adversarial noise; as well as Image queries that are at least partially modified.

18. The method of claim 11, further comprising: The original query is transformed into a query that is at least partially modified through one or more filtering components; The original query and the at least partially modified query are iteratively transmitted to the one or more semantic services; The at least partially modified query is iteratively transmitted to the one or more AI agents; The at least partially modified query is iteratively transmitted to the one or more filtering components; Iteratively determine: The human's ability to understand the at least partially modified query through the one or more semantic services; as well as It is impossible to respond to at least partially modified queries via electronic devices without human support, thereby establishing that the at least partially modified query is a fully modified query. as well as The completely modified query is transmitted to the human.

Images