Maintaining data integrity in cognitive multi-agent systems

Abstract

The present disclosure relates to maintaining data integrity in a cognitive multi-agent system. A method is provided, wherein the method comprises applying a first axiom to a set of data points by a first node to generate a first set of outputs. The method comprises applying a second axiom to the set of data points by a second node to generate a second set of outputs. The first node and the second node are part of a computer network comprising a plurality of nodes. The method comprises calculating a first nuance based on a set of differences between the first set of outputs and the second set of outputs, and adjusting the reliability of the first node in the computer network based on the first nuance.

Description

Background Art

[0001] Advances in computing hardware technology have led to a rapid increase in the use of intelligent / cognitive computing devices for automation. Most of these systems rely on sensors to collect the data needed to automate decision-making processes. For example, manual monitoring of temperature in a nuclear power plant can be replaced by an array of thermal sensors and a programmable logic controller (PLC) for releasing coolant. Based on the temperature data collected from these sensors, the PLC can reduce the manual intervention required to maintain reactor temperature. This can make a huge difference in catastrophic events such as a reactor meltdown. Therefore, it is necessary to maintain data integrity between these devices to ensure high-quality automation of decision-making. In a situation where four out of ten thermal sensors fail, detecting the failed thermal sensor becomes critical.

[0002] Typically, these sensor / cognitive devices have their own cognitive processes to detect phenomena (e.g., temperature) and map them into numerical values. These devices are typically deployed as networks that allow data points such as temperature values ​​to be shared between various devices. This data sharing enables large-scale automation. For example, a network of 50 devices can be used to estimate weather readings for an area of ​​approximately 100 square kilometers. A group of N devices (e.g., N=5) can be responsible for every 10 square kilometers. By sharing weather data between N adjacent devices in a region, the aggregate weather readings for that region can be estimated. Each node in the region can rely on data sent by the adjacent N-1 nodes to determine its own accuracy. Summary of the Invention

[0003] According to one embodiment of the present disclosure, a method is provided, wherein the method comprises: applying a first axiom to a set of data points by a first node to generate a first output set; applying a second axiom to the set of data points by a second node to generate a second output set; the first node and the second node are part of a computer network comprising a plurality of nodes; calculating a first nuance based on a set of differences between the first output set and the second output set; and adjusting the reliability of the first node in the computer network based on the first nuance.

[0004] The foregoing is a summary and therefore necessarily contains simplifications, generalizations, and omissions of detail; therefore, those skilled in the art will appreciate that this summary is illustrative only and is not intended to be limiting in any way. Other aspects, inventive features, and advantages of the present invention, as defined solely by the claims, will become apparent in the non-limiting detailed description set forth below. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] The present disclosure may be better understood, and its numerous objects, features, and advantages made apparent to those skilled in the art by referencing the accompanying drawings, in which:

[0006] Figure 1 is a block diagram of a data processing system in which the methods described herein may be implemented;

[0007] Figure 2 Provided Figure 1 The illustrated information handling system environment is expanded to illustrate that the methods described herein can be executed on various information handling systems operating in a networked environment;

[0008] Figure 3 is an exemplary diagram describing a data integrity system that maintains the integrity of a multi-agent system;

[0009] Figure 4 is an exemplary flow chart illustrating steps taken to evaluate nuances of nodes in a cognitive multi-agent system based on their axioms, compute their reliability, and iteratively update the axioms of the nodes accordingly;

[0010] Figure 5 is an exemplary diagram illustrating the data integrity system 350 determining differences within a node over time and calculating nuances of the node based on the differences;

[0011] Figure 6 is an exemplary diagram describing the relationship between node axioms and nuances; and

[0012] Figure 7 is an exemplary diagram illustrating the relationship between nodes, axioms, nuances, and reliability; and

[0013] Figure 8 are exemplary diagrams depicting two embodiments of iteratively updating node axioms in a cognitive multi-agent system based on nuance and reliability. DETAILED DESCRIPTION

[0014] The terms used herein are for the purpose of describing particular embodiments only and are not intended to limit the present invention. As used herein, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0015] The corresponding structures, materials, actions, and equivalents of all means or step plus function elements in the following claims are intended to include any structure, material, or action for performing a function in combination with other claimed elements as specifically stated. The description of the present disclosure has been presented for the purpose of illustration and description, but it is not intended to be exhaustive or limited to the disclosure in the form disclosed. Many modifications and variations will be clear to those of ordinary skill in the art without departing from the scope and spirit of the present disclosure. The embodiments are selected and described in order to best explain the principles and practical applications of the present disclosure and to enable others of ordinary skill in the art to understand the disclosure of various embodiments with various modifications suitable for the specific purposes envisioned.

[0016] The present invention may be a system, method and / or computer program product at any possible level of technical detail integration. The computer program product may include a computer-readable storage medium (or multiple media) having computer-readable program instructions thereon, the computer-readable program instructions being used to cause a processor to perform various aspects of the present invention.

[0017] Computer readable storage medium can be a tangible device that can retain and store the instructions used by the instruction execution device.Computer readable storage medium can be, for example, but not limited to, electronic storage device, magnetic storage device, optical storage device, electromagnetic storage device, semiconductor storage device or any suitable combination of the foregoing.A non-exhaustive list of more specific examples of computer readable storage medium includes the following: portable computer disk, hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, such as a punch card or a raised structure in a groove with instructions recorded thereon, and any suitable combination of the foregoing.Computer readable storage medium as used herein should not be interpreted as a transient signal itself, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagated by a waveguide or other transmission medium (for example, a light pulse by an optical fiber cable), or an electrical signal transmitted by a wire.

[0018] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to a corresponding computing / processing device, or downloaded to an external computer or external storage device via a network (e.g., the Internet, a local area network, a wide area network, and / or a wireless network). The network can include copper transmission cables, optical transmission fibers, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in a computer-readable storage medium within the corresponding computing / processing device.

[0019] The computer-readable program instructions for performing the operation of the present invention can be assembly instructions, instruction set architecture (ISA) instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, configuration data of integrated circuits, or source code or object code written in any combination of one or more programming languages, the one or more programming languages ​​including object-oriented programming languages ​​(such as Smalltalk, C++, etc.) and procedural programming languages ​​(such as " C " programming language or similar programming languages). The computer-readable program instructions can be performed completely on the user's computer or partly on the user's computer as an independent software package, partly on the user's computer and partly on a remote computer, or completely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any type of network (including local area network (LAN) or wide area network (WAN)), or can be connected to an external computer (such as, using an Internet service provider through the Internet). In certain embodiments, in order to perform various aspects of the present invention, the electronic circuit comprising, for example, a programmable logic circuit, a field programmable gate array (FPGA) or a programmable logic array (PLA) can perform the computer-readable program instructions to personalize the electronic circuit by utilizing the state information of the computer-readable program instructions.

[0020] Aspects of the present invention are described herein with reference to the flowcharts and / or block diagrams of the methods, apparatus (systems) and computer program products according to embodiments of the present invention. It will be understood that each block of the flowcharts and / or block diagrams and the combination of blocks in the flowcharts and / or block diagrams can be implemented by computer-readable program instructions.

[0021] These computer-readable program instructions can be provided to a processor of a computer or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device create a device for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium, which can direct the computer, programmable data processing device and / or other equipment to operate in a specific manner, so that the computer-readable storage medium having the instructions stored therein includes an article of manufacture, which includes instructions for implementing various aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0022] Computer-readable program instructions may 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, so that the instructions executed on the computer, other programmable apparatus, or other device implement the functions / actions specified in one or more boxes of the flowchart and / or block diagram.

[0023] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functions and operations of possible implementations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram can represent a module, segment or portion of an instruction, which includes one or more executable instructions for implementing a specified logical function. In some alternative embodiments, the functions noted in the block diagram may not occur in the order noted in the figure. For example, the two blocks shown in succession can actually be implemented as a step, simultaneously, substantially simultaneously, in a manner that overlaps part or all of the time, or these blocks can sometimes be performed in reverse order, depending on the functions involved. It will also be noted that each block of the block diagram and / or flowchart illustration and the combination of the blocks in the block diagram and / or flowchart illustration can be implemented by a dedicated hardware-based system that performs a specified function or action or performs a combination of dedicated hardware and computer instructions. The following detailed description will generally follow the summary of the present disclosure as described above, further explaining and expanding the definitions of various aspects and embodiments of the present disclosure when necessary.

[0024] Figure 1An information handling system 100 is shown, which is a simplified example of a computer system capable of performing the computing operations described herein. Information handling system 100 includes one or more processors 110 coupled to a processor interface bus 112. Processor interface bus 112 connects processors 110 to a Northbridge 115, also known as a memory controller hub (MCH). Northbridge 115 connects to system memory 120 and provides a means for processor(s) 110 to access system memory. A graphics controller 125 is also connected to Northbridge 115. In one embodiment, a Peripheral Component Interconnect (PCI) Express bus 118 connects Northbridge 115 to graphics controller 125. Graphics controller 125 is connected to a display device 130, such as a computer monitor.

[0025] Northbridge 115 and Southbridge 135 are connected to each other using bus 119. In some embodiments, the bus is a direct media interface (DMI) bus that transfers data at high speed in each direction between Northbridge 115 and Southbridge 135. In some embodiments, a PCI bus connects the Northbridge and Southbridge. Southbridge 135, also known as an input / output (I / O) controller hub (ICH), is a chip that typically implements the ability to operate at a slower speed than the capabilities provided by the Northbridge. Southbridge 135 typically provides various buses for connecting various components. These buses include, for example, PCI and PCI Express buses, ISA buses, system management buses (SMBus or SMB), and / or low pin count (LPC) buses. The LPC bus typically connects low-bandwidth devices such as boot ROM 196 and "legacy" I / O devices (using a "super I / O" chip). "Legendary" I / O devices (198) can include, for example, serial and parallel ports, keyboards, mice, and / or floppy disk controllers. Other components typically included in Southbridge 135 include a direct memory access (DMA) controller, a programmable interrupt controller (PIC), and a storage device controller that connects Southbridge 135 to non-volatile storage devices 185 (such as a hard drive) using bus 184.

[0026] ExpressCard 155 is a slot for connecting hot-pluggable devices to the information processing system. ExpressCard 155 supports both PCI Express and USB connections when connected to South Bridge 135 using both the Universal Serial Bus (USB) and PCI Express buses. South Bridge 135 includes a USB controller 140, which provides USB connections to devices connected to the USB. These devices include a webcam 150, an infrared (IR) receiver 148, a keyboard and trackpad 144, and a Bluetooth device 146, which provides a wireless personal area network (PAN). USB controller 140 also provides USB connections to various other USB-connected devices 142, such as a mouse, a removable non-volatile storage device 135, a modem, a network card, an Integrated Services Digital Network (ISDN) connector, a fax machine, a printer, a USB hub, and many other types of USB-connected devices. Although the removable non-volatile storage device 145 is shown as a USB connected device, the removable non-volatile storage device 145 may be connected using a different interface, such as a FireWire interface, etc.

[0027] Wireless local area network (LAN) device 175 connects to Southbridge 135 via PCI or PCI Express bus 172. LAN device 175 typically implements one of the Institute of Electrical and Electronics Engineers (IEEE) 802.11 over-the-air technology standards, all of which use the same protocol for wireless communication between information handling system 100 and another computer system or device. Optical storage device 190 connects to Southbridge 135 using Serial Analog Telephone Adapter (ATA) (SATA) bus 188. Serial ATA adapters and devices communicate over a high-speed serial link. The Serial ATA bus also connects Southbridge 135 to other forms of storage devices, such as hard drives. Audio circuitry 160, such as a sound card, connects to Southbridge 135 via bus 158. Audio circuitry 160 also provides functionality associated with audio hardware, such as audio line-in and optical digital audio input port 162, optical digital output and headphone jack 164, internal speakers 166, and internal microphone 168. Ethernet controller 170 connects to Southbridge 135 using a bus, such as a PCI or PCI Express bus. Ethernet controller 170 connects information handling system 100 to a computer network, such as a local area network (LAN), the Internet, and other public and private computer networks.

[0028] Although Figure 1An information handling system is shown, but an information handling system may take many forms. For example, an information handling system may take the form of a desktop computer, server, portable computer, laptop computer, notebook computer, or other form factor computer or data processing system. In addition, an information handling system may take other form factors such as a personal digital assistant (PDA), a gaming device, an automated teller machine (ATM), a portable telephone device, a communication device, or other device that includes a processor and memory.

[0029] Figure 2 Provided Figure 1 2. The information processing system environment shown is an expansion of the information processing system environment to illustrate that the methods described herein can be performed on various information processing systems operating in a networked environment. The types of information processing systems range from small handheld devices such as handheld computers / mobile phones 210 to large systems such as mainframe computers 270. Examples of handheld computers 210 include personal digital assistants (PDAs), personal entertainment devices such as Moving Picture Experts Group Layer 3 audio (MP3) players, portable televisions, and compact disc players. Other examples of information processing systems include pen or tablet computers 220, laptop or notebook computers 230, workstations 240, personal computer systems 250, and servers 260. Figure 2 Other types of information handling systems not shown separately are represented by information handling system 280. As shown, various information handling systems can be networked together using computer network 200. Types of computer networks that can be used to interconnect various information handling systems include local area networks (LANs), wireless local area networks (WLANs), the Internet, public switched telephone networks (PSTNs), other wireless networks, and any other network topology that can be used to interconnect information handling systems. Many information handling systems include non-volatile data storage devices, such as hard disk drives and / or non-volatile memory. Figure 2 The embodiment of the information handling system shown in FIG includes separate non-volatile data storage devices (more specifically, server 260 utilizes non-volatile data storage device 265, mainframe computer 270 utilizes non-volatile data storage device 275, and information handling system 280 utilizes non-volatile data storage device 285). The non-volatile data storage device can be a component external to the various information handling systems, or it can be a component internal to one of the information handling systems. In addition, removable non-volatile storage device 145 can be shared between two or more information handling systems using various techniques, such as connecting the removable non-volatile storage device 145 to a USB port or other connector of the information handling system.

[0030] As described above, a fundamental aspect of nuanced nodes / sensors in cognitive multi-agent networks is their axioms, and nodes in cognitive multi-agent networks interact with each other by assigning weights to their neighbors based on the data they receive from them. One challenge with existing systems is that, while existing systems have mechanisms for ranking nodes, they do not consider the axioms of the nodes on the data points. Additionally, while existing systems can evaluate the reputation of nodes, they do not consider the nuances of each node based on the independent cognitive capabilities of each node to maintain axioms. In other words, existing systems do not consider how a node's cognitive process differs from other nodes with respect to the different data points it receives. In order to detect faulty sensors / devices, it is necessary to analyze their behavior when detecting the desired phenomenon. Typically, the digital readings of a faulty sensor will differ from the majority of sensors in the subnetwork or neighborhood. This persistent difference in reading values ​​makes the nuances of the faulty node less apparent.

[0031] A fundamental aspect of a node / sensor in a cognitive multi-agent network, with subtle differences (small distinctions), is its axioms. A node's axioms are inherent properties of the node that represent the node's local view of the world for any data point it receives. Such axioms are analogous to generally accepted facts about the node and are considered fundamental to the node's decision-making process. For example, for the weather processing node in the scenario above, the node's axiom might be "the temperature of the boiler cannot fall below 0 degrees Celsius." The node takes this belief or axiom at face value because it has been programmed to do so or because of its experience.

[0032] Nodes in cognitive multi-agent networks interact with each other by assigning weights to their neighbors based on the data they receive from them. These weights are a measure of the importance a node assigns to the data it receives from its neighbors. The weights they assign to another node depend on: (a) their own axioms, and (b) the data they generate. If the weight given to a particular node by its neighbors is high, then that particular node is less likely to be a faulty node because the data points sent by that node are consistent with (a) the neighboring node's axioms and (b) the data points generated by the neighboring node.

[0033] Figures 3 to 8 A method, executable on an information processing system, is described for maintaining data integrity in a large autonomous cognitive multi-agent system. The method evaluates nuances of nodes in the cognitive multi-agent system based on axioms. The method then computes the reliability of nodes in the cognitive multi-agent system based on the axioms and nuances. Furthermore, the method iteratively updates the axioms of nodes in the cognitive multi-agent system based on the nuances.

[0034] Figure 3FIG3 is an exemplary diagram illustrating a data integrity system for maintaining the integrity of a multi-agent system. Integrity system 350 provides a mechanism for computing nuances of nodes 310 in cognitive multi-agent system 300 using axioms on different input data points. In one embodiment, data integrity system 350 assigns nuances based on axioms on input data points in applications such as: reliable weather monitoring and forecasting over large areas; reliable maintenance and monitoring of temperature in heat-sensitive equipment; reliable navigation and global positioning using satellite data; and large-scale reliable deployment of edge devices.

[0035] In one embodiment, the network of the cognitive multi-agent system includes two different types of embedded structures, such as an infrastructure network and an interaction network. In this embodiment, the infrastructure network includes the actual connections between nodes, which essentially configures the infrastructure for further data exchange (e.g., a satellite constellation for a global positioning system). The interaction network is an overlay network that uses the interactions between nodes of the underlying infrastructure network (e.g., the weights assigned by the nodes to each other as a result of data exchange).

[0036] Data integrity system 350 employs algorithms that define metrics such as axioms, variances, and nuances across the network of interactions underlying cognitive multi-agent system 300. Once data integrity system 350 evaluates the axioms of node 310 in cognitive multi-agent system 300 based on its axioms, it provides a mechanism for calculating the reliability of node 310 in maintaining data integrity based on the nuances. A node's reliability is an indicator of the independent nature of its cognitive processes, based on the axioms that the node holds across various data points. Furthermore, reliability is a measure of data integrity and is defined iteratively based on the reliability and nuances of the node's neighbors and its own nuances.

[0037] In one embodiment, data integrity system 350 scores nodes 310 based on the convergence of reliability. The reliability score determines which nodes are most reliable in maintaining the network's data integrity. The higher the reliability of a node, the less likely it is to corrupt data and the greater the likelihood of maintaining data integrity. In turn, integrity system 350 iteratively updates the axioms of nodes 310 based on the nuances and reliability of nodes 310 in cognitive multi-agent system 300, enabling the evolution of nodes 310 in multi-agent system 300 over time.

[0038] Figure 4 is an exemplary flow chart illustrating steps taken to evaluate nuances of nodes, compute their reliability, and iteratively update the axioms of nodes accordingly based on the axioms of the nodes in a cognitive multi-agent system.

[0039] Figure 4 Processing begins at 400, whereupon, at step 420, the process evaluates the nuances of the nodes 310 in the cognitive multi-agent system 300 based on their axioms. In one embodiment, the data integrity system 350 solves the problem of computing the nuances of the nodes in the cognitive multi-agent system 300 based on their axioms on the various input data points they receive. Specifically, the data integrity system 350 focuses on the tendency of the nodes to correctly analyze the data points they receive and include the data points in their decision making process based on the axioms they hold. This ability to analyze data points is referred to herein as the nuance of the node. As defined herein, nuance is the variance that a node exhibits in its ability to analyze a data point. In one embodiment, the variance is in the range [0, 1] (see further details in [0, 1] for more information). Figure 6 and corresponding text).

[0040] In one embodiment, to assess nuance and help characterize the analytical capabilities of nodes based on axioms, data integrity system 350 determines whether a node exhibits tendencies such as: i) whether the node holds the same axioms regardless of the data points it receives; ii) whether the node tends to be compulsively different from other nodes regardless of the data points it receives; or iii) whether the node tends to be the same or different from other nodes without following a specific pattern in the data point set. In another embodiment, integrity system 350 defines variance as the difference between each node's axioms and those of other nodes. Variance is modeled in two ways. The first type of variance is variance from the aggregated consensus. In the first type of variance, for a given data point, data integrity system 350 calculates the aggregated consensus of the population and then calculates the difference between each node's axioms and the consensus. The second type of variance is variance from individual nodes. In the second type of variance, for a collection of data points, data integrity system 350 calculates the total divergence of the node's axioms from those of other nodes. For either type of variance, data integrity system 350 defines the node's nuance as the aggregate (statistical moment) of the variance distribution.

[0041] At step 440, the process computes the reliability of the node based on the axioms and corresponding nuances of the node in the cognitive multi-agent system 300. The data integrity system 350 defines node reliability based on the axioms held by a particular node and its corresponding nuances on the data points. (For further details, see Figure 7 and corresponding text).

[0042] At step 460, the process iteratively updates the axioms of the nodes in the cognitive multi-agent system based on their corresponding nuances and reliability. In one embodiment, the integrity system 350 uses one of the three models described herein to perform simultaneous computation of nuances and reliability and iteratively update the axioms of the nodes (see further details in Figure 7 、 Figure 8 and corresponding text). The data integrity system 350 continues to score the nodes and determines which nodes are most reliable for preserving data integrity in the cognitive multi-agent system 300. Figure 4 The processing then ends at 495.

[0043] Figure 5 is an exemplary diagram illustrating the data security system 350 determining differences within a node over time and calculating nuances of the node based on the differences.

[0044] The data points 500 of the neighboring nodes are fed into both node A 510 and node B 540. Axiom A 515 of node A 510 applies interpretation to the input and outputs a value "A." Similarly, axiom B 545 of node B 540 applies interpretation to the input and outputs a value "B."

[0045] The difference analyzer 520 of node A 510 evaluates the difference between its value A and value B, as well as the difference between its value A and the values ​​C, D, and E from neighboring nodes C, D, and E. The difference analyzer 520 stores these differences (e.g., A=4, B=5, difference=2) in a data storage 525. Over time, the difference analyzer 520 stores multiple differences in the data storage 525. The nuance calculation module 530 analyzes the differences in the data storage 525 and calculates nuance A 535, which, in one embodiment, ranges from 0 to 1, where nuances closer to 0 are low nuances and nuances closer to 1 are high nuances.

[0046] Similarly, a variance analyzer 550 of node B 540 evaluates the variance between its value B and value A, as well as the variance between its value B and the values ​​F, G, and H from neighboring nodes F, G, and H. Variance analyzer 550 stores these variances in a data store 555, and over time, variance analyzer 550 stores multiple variances in data store 555. A nuance calculation module 560 analyzes the variances in data store 555 and calculates nuance B 565, which, in one embodiment, ranges from 0 to 1, where a nuance closer to 0 is a low nuance and a nuance closer to 1 is a high nuance.

[0047] As discussed herein, the data integrity system 350 uses the calculated nuances to evaluate nodes and adjust their axioms as needed to improve the overall reliability of the cognitive multi-agent system 300 .

[0048] Figure 6 is an exemplary diagram illustrating the relationship between axioms and nuances of a node. Data integrity system 350 employs a novel approach to scoring nodes 310 in cognitive multi-agent system 300 by considering the axioms held by nodes 310 on different data points and the nuances exercised by nodes 310. As described below, the framework of integrity system 350 includes a signed network that captures the approval and disapproval relationships between nodes as factors in calculating their corresponding nuances.

[0049] Figure 6 Node 610, axioms 620, and nuances 630 are shown. As previously discussed, axioms 620 are intrinsic properties of node 610 that represent the node's 610 local view of the world for any data point it receives. Axioms 620 are generally accepted facts about node 610 and are considered to be the basis for the node's 610 decision-making process. Each node 310 has an autonomous cognitive process based on their corresponding axioms on different data points. When node 610 receives a data point, node 610 applies axioms 620 and generates a perception for that data point to include in its decision-making process. In this way, for each data point, node 610 generates a new perception using axioms 620. As discussed herein, axioms 620 can be changed by the influence of its neighbors (see further details in the ). Figure 8 and corresponding text).

[0050] Nuance 630 is the ability of node 630 to analyze different data points received by node 610 based on axioms 620. The output of node 610's cognitive process changes based on the data points it receives. In this way, integrity system 350 describes a cognitive system in which independent nodes 310 in multi-agent system 310 interact with each other and exchange data points, and use their nuances as a scoring mechanism for nodes 310.

[0051] The data integrity system 350 presents a model for scoring the nuances of nodes based on the axioms held by nodes 310 in the cognitive multi-agent system 300. The data integrity system 350 temporarily takes the axioms of the nodes on the data points as input; presents the basic mathematical model for computing the nuances; and then describes the computation of the nuances using the nuances. In one embodiment, the data integrity system 350 takes a directed, unsigned graph G U Defined as G u =(V,E,w u ),in:

[0052] V is the set of nodes in the graph;

[0053] is a set of ordered pairs of nodes representing edges in the graph; and

[0054] w u :E→[0,1] indicates the weight of the edge.

[0055] The data integrity system 350 then defines G = (V, E, w) as a directed, signed graph where the weights on the edges are given by w: E → [-1, 1]. The function w u The range of is a subset of the range of the function w, so signed networks (graphs) are generalizations of unsigned networks.

[0056] In one embodiment, integrity system 350 uses the following as a basis for determining nuance 620:

[0057] i) An interaction network modeled as a directed graph represented by an adjacency matrix. Edges can be signed - for example, the edge weight represents the vote of approval (+1) or opposition (-1) placed by the source node of the edge in the destination node.

[0058] ii) A collection of data points exchanged over the network. The data integrity system 350 assumes that these data points are injected into the network in batches.

[0059] iii) The axioms that each node holds for each data point exchanged through the network. The axioms are represented as a matrix where each column represents the axiom vector of the node on all data points. b (i) is set as the axiom for node i on data point n, and the data integrity system 350 assumes X n (i)∈[0,1], where the value quantifies the axiom that the node holds on the data point n.

[0060] For each data point n, the data integrity system 350 calculates the difference between the node and the other nodes. First, the data integrity system 350 describes the difference between the calculated node and the aggregated general opinion. For each data point, the data integrity system 350 calculates the difference between the general opinion O and the aggregated general opinion O. n It is defined as the average of the axioms of all nodes that receive data point n:

[0061]

[0062] where |S n | is the total number of nodes that received the data point. Next, the data integrity system 350 calculates the difference between the node's axiom and the general opinion for each data point and calls it the difference d n (i)

[0063] dn (i) = X n (i)-O n (2)

[0064] In one embodiment, the data integrity system 350 calculates the difference between a node and individual nodes rather than general consensus. For each node, the data integrity system 350 defines the difference as the pairwise cosine dissimilarity or Euclidean distance between its own axiom vector and the axiom vectors of other nodes. Any other distance measure (normalized to the interval [0, 1]) can also be used instead of cosine or Euclidean distance:

[0065] d(i,j)=1-cos(X(i),X(j))(3)

[0066] where X(i) is the axiom vector of node i for a batch of data points, and:

[0067]

[0068] The summation is performed on the batch of data points.

[0069] Next, the data integrity system 350 calculates a difference vector. In one embodiment, the data integrity system 350 determines the difference using equation (2) above, where the difference vector of a node is defined as the set of differences for that node for all data points received by that node:

[0070] H(i)={d n (i)} n (5)

[0071] In another embodiment, data integrity system 350 determines the differences using Equations 3 and 4 above, where a node's difference vector is defined as the set of differences between the node and other nodes in the network:

[0072] H(i)={d(i,j)} j (6)

[0073] For each node (node ​​610), data integrity system 350 calculates its corresponding nuance, where nuance 630 is defined as the variance of its difference vector:

[0074] η(i)=σ 2 (H(i))(7)

[0075] Nuance 630 represents the ability of node 610 to analyze different data points received by node 610 based on axioms 620 .

[0076] The output of a node's cognitive process varies depending on the data points it receives. In one embodiment, nuance is bounded between [0, 1], where values ​​closer to zero indicate a node with less nuance, while values ​​closer to 1 mean that the node is very nuanced. If node 610 consistently agrees with the general population on every data point, then node 610 exhibits consistent behavior. If node 610 consistently differs from the general population on every data point, then node 610 exhibits prohibited behavior. Neither of these behaviors is classified as nuanced because in both cases, node 610's difference vector has a small variance.

[0077] On the other hand, if node 610 does not exhibit any particular pattern in terms of agreement or differences with other nodes, then node 610 is considered to be nuanced (nuance 630), and therefore the variance of the difference vector of node 610 is high. Nodes in the network are scored based on their corresponding nuances 630. In one embodiment, to calculate the generalization of the differences of the signed network, the integrity system 350 defines w′ as the edge weights scaled to the range [0, 1]:

[0078]

[0079] where t is the edge weight in the graph G. The data integrity system 350 defines the difference as:

[0080] d′(i,j)=w′(i,j)*d(i,j)+(1-w′(i,j))*(1-d(i,j)) (10)

[0081] in

[0082] d'(i, j): weighted difference;

[0083] w'(i, j): the linearly transformed voting value t belongs to [-1, 1]->t belongs to [0, 1]; and

[0084] d(i,j): the difference value as defined in (ii).

[0085] Figure 7 is an exemplary diagram illustrating the relationship between nodes, axioms, nuances, and reliability. Figure 7 Nodes 610, axioms 620, nuances 630, and reliability 710 are shown. For a data point n, the data integrity system 350 calculates nuances 630 as described above. Using nuances 630 and edge weights, the data integrity system 350 calculates the reliability 710 of each node, where, in one embodiment, the reliability of a node is defined as the eigenvector centrality of the (unsigned) graph's adjacency matrix modified with the nuance value:

[0086] R=A T R(1)

[0087] where A is the edge weight w ik Defined and determined by the nuance η of the destination node j j The modified adjacency matrix of the scaled graph can be defined by any of the following equations:

[0088] 1.A[i,j]=w ij *η j

[0089] 2.A[i,j]=w ij 1 / ψ *η j 1 / Φ ,in

[0090] 3.A[i,j]=ψ*w ij +φ*η j ,in

[0091] 4.A[i,j]=ψ*w ij +Φ*η j θ ,in and

[0092] The data integrity system 350 iteratively calculates the feature vector, wherein after each iteration (720), the data integrity system 350 scores the node based on reliability. When the Spearman correlation score of the reliability value of the node exceeds 1-ε for a number of iterations for an arbitrarily small ε, the data integrity system 350 reaches a convergence condition. The data integrity system 350 scores the node based on the reliability after the iteration converges.

[0093] As described above, the convergence criteria is defined on the Spearman correlation scores of the reliability values. When all iterative calculations converge, the data integrity system 350 scores the nodes of the network based on reliability to determine which nodes are most effective in preserving data integrity.

[0094] Figure 8is an exemplary diagram illustrating two embodiments for iteratively updating node axioms in a cognitive multi-agent system based on nuance and reliability. As described above, for a data point n, data integrity system 350 uses nuance and edge weights to calculate the nuance and reliability of each node. After each iteration, data integrity system 350 scores the node based on reliability, and convergence is achieved when the Spearman correlation score of the node's reliability value exceeds 1-ε for a number of iterations for an arbitrarily small ε.

[0095] When all iterative calculations converge, the data integrity system 350 scores the nodes 310 based on their respective reliability and nuance. The purpose of the scoring is twofold, i) to determine the convergence of the iterative calculations of reliability and nuance; and ii) to determine which nodes are most reliable for maintaining data integrity.

[0096] The model 800 updates the axioms 620 using the following formula, which uses a weighted sum of the axiom value in the previous iteration and the influence of its neighbors:

[0097]

[0098] X k : Axioms in the kth iteration

[0099] i: the node where the axiom is being updated

[0100] N in : The inner neighbors of the node

[0101] δ: The weight given to the influence of inner neighbors in the kth iteration

[0102] η: Nuances of Nodes

[0103] The degree of influence depends on the axioms and nuances of its neighbors and its own nuances. Model 800 has two iterative processes. The first iterative process leads to convergence of nuances and axioms (810). Once the axioms converge, the second iterative process calculates reliability (720). Convergence of nuances always precedes convergence of reliability. Data integrity system 350 defines this behavior as binary convergence of axioms and nuances, followed by unary convergence of reliability.

[0104] Model 850 is similar to model 800, except that the influence term now depends on the reliability of its neighbors instead of the nuance 630:

[0105]

[0106] X k : axioms in the kth iteration;

[0107] i: the node where the axiom is being updated;

[0108] N in : inner neighbors of the node;

[0109] δ: the weight given to the influence of inner neighbors in the kth iteration; and

[0110] R: Node reliability.

[0111] Model 850 has two nested iterative processes. The inner iterative process computes reliability based on nuances (720). When the reliability converges, the axioms are updated (860), which is the outer iterative process. In turn, the newly computed nuances 630 lead back to the iterative process of computing reliability. Data integrity system 350 refers to this behavior as the ternary convergence of axioms, nuances, and reliability.

[0112] In one embodiment, the data integrity system 350 uses a third model, which is a time-based model, to iteratively update the axioms of the node 310. The third model tracks how the axioms change over time:

[0113] X'=dX / dt

[0114] Where X' is the first derivative of the node's axiom vector with respect to time. To discover the consistency of a node, the data integrity system 350 tracks how often the node's axioms change. The data integrity system 350 finds the second derivative of the axioms with time and calculates the root mean square error (RMSE) with the zero vector (which represents the origin), which is given by:

[0115]

[0116] In this embodiment, the data integrity system 350 defines the consistency C of a node as:

[0117]

[0118] The lower the RMSE, the higher the node consistency. The data integrity system 350 uses the node consistency as an additional parameter in the axiom update process of the data integrity system 350, using the following equation:

[0119]

[0120] X t : axiom at time t;

[0121] k: number of iterations;

[0122] i: the node where the axiom is being updated;

[0123] N in : inner neighbors of the node;

[0124] η j : subtle differences in node j;

[0125] P t : the reliability of node j; and

[0126] δ: The weight given to the influence of inner neighbors at the kth iteration.

[0127] Data integrity system 350 may use model 800 , model 850 , a time-based model, and / or other models to maintain data integrity for nodes 310 in cognitive multi-agent system 300 .

[0128] While specific embodiments of the present disclosure have been shown and described, it will be apparent to those skilled in the art, based on the teachings herein, that changes and modifications may be made without departing from the present disclosure and its broader aspects. The appended claims are therefore intended to cover within their scope all such changes and modifications that come within the true spirit and scope of the present disclosure. Furthermore, it should be understood that the present disclosure is limited solely by the appended claims. Those skilled in the art will understand that if a specific number of an introduced claim element is intended, such intent will be expressly recited in the claim, and in the absence of such a recitation, no such limitation exists. By way of non-limiting example, to aid understanding, the appended claims contain the use of the introductory phrases "at least one" and "one or more" to introduce claim elements. However, the use of such phrases should not be construed to imply that the introduction of a claim element by the indefinite article "a" or "an" limits any particular claim containing such introduced claim element to a disclosure containing only one such element, even when the same claim includes the introductory phrases "one or more" or "at least one" and an indefinite article such as "a" or "an"; the same applies to the use of definite articles in a claim.

Claims

1. A computer-implemented method for maintaining data integrity, comprising: Applying a first axiom by a first node to a set of data points to generate a first set of outputs, wherein the first axiom represents a local world view of the first node for the data points received by the first node; applying, by a second node, a second axiom to the set of data points to generate a second set of outputs, wherein the first node and the second node are part of a computer network comprising a plurality of nodes, wherein the second axiom represents a local world view of the second node for the data points received by the second node; calculating a first nuance of the first node based on a set of differences between the first output set and the second output set; calculating the reliability of the first node in the computer network based on the first nuance; as well as The first axiom is iteratively updated based on the first nuance and the reliability of the first node.

2. The computer-implemented method of claim 1 , further comprising: capturing the first set of outputs and the second set of outputs within a time period; as well as The set of differences between the first set of outputs and the second set of outputs over the time period is calculated, wherein the first nuance is a variance of the set of differences between the first set of outputs and the second set of outputs over the time period.

3. The computer-implemented method of claim 1 , wherein calculating the reliability of the first node in the computer network based on the first nuance comprises: The reliability of the first node is iteratively calculated based on a set of edge weights corresponding to a set of neighboring nodes and the first nuance, until the reliability meets a convergence condition based on the set of edge weights.

4. The computer-implemented method of claim 3 , wherein iteratively updating the first axiom based on the first nuance and the reliability of the first node comprises: Before iteratively calculating the reliability of the first node, the first axiom is iteratively adjusted based on a weighted sum of a set of iterative nuance calculations to converge the first axiom and the first nuance.

5. The computer-implemented method of claim 3 , wherein iteratively updating the first axiom based on the first nuance and the reliability of the first node comprises: In response to iteratively calculating the reliability of the first node, the first axiom is iteratively adjusted to converge the first axiom, the first nuance, and the reliability.

6. The computer-implemented method of claim 5 , further comprising: tracking a set of adjustments to the first axiom over a period of time; determining consistency of the first node based on the set of adjustments; as well as The consistency is included as a factor during iterative adjustment of the first axiom.

7. The computer-implemented method of claim 1 , wherein the computer network is a cognitive multi-agent system comprising a set of independent nodes, and wherein the method further comprises: Based on an axiom set including the first axiom and the second axiom, enabling the group of independent nodes including the first node and the second node to maintain an autonomous cognitive process; enabling the set of independent nodes to interact with each other and exchange the set of data points; as well as A set of networks including the set of independent nodes in the cognitive multi-agent system is enabled to capture a set of voting approval relationships and a set of voting disapproval relationships between the set of independent nodes.

8. An information processing system for maintaining data integrity, comprising: one or more processors; a memory coupled to at least one of the processors; A set of computer program instructions stored in the memory and executed by at least one of the processors to perform the following actions: Applying a first axiom by a first node to a set of data points to generate a first set of outputs, wherein the first axiom represents a local world view of the first node for the data points received by the first node; applying, by a second node, a second axiom to the set of data points to generate a second set of outputs, wherein the first node and the second node are part of a computer network comprising a plurality of nodes, wherein the second axiom represents a local world view of the second node for the data points received by the second node; calculating a first nuance of the first node based on a set of differences between the first output set and the second output set; calculating the reliability of the first node in the computer network based on the first nuance; as well as The first axiom is iteratively updated based on the first nuance and the reliability of the first node.

9. The information processing system of claim 8, wherein the processor performs further actions comprising: capturing the first set of outputs and the second set of outputs within a time period; and calculating the set of differences between the first set of outputs and the second set of outputs over the time period, wherein: The first nuance is a variance of the set of differences between the first set of outputs and the second set of outputs over the time period.

10. The information processing system of claim 8, wherein calculating the reliability of the first node in the computer network based on the first subtle difference comprises: The reliability of the first node is iteratively calculated based on a set of edge weights corresponding to a set of neighboring nodes and the first nuance, until the reliability meets a convergence condition based on the set of edge weights.

11. The information processing system of claim 10 , wherein iteratively updating the first axiom based on the first nuance and the reliability of the first node comprises: Before iteratively calculating the reliability of the first node, the first axiom is iteratively adjusted based on a weighted sum of a set of iterative nuance calculations to converge the first axiom and the first nuance.

12. The information processing system of claim 10 , wherein iteratively updating the first axiom based on the first nuance and the reliability of the first node comprises: In response to iteratively calculating the reliability of the first node, the first axiom is iteratively adjusted to converge the first axiom, the first nuance, and the reliability.

13. The information processing system of claim 12, wherein the processor performs further actions comprising: tracking a set of adjustments to the first axiom over a period of time; determining consistency of the first node based on the set of adjustments; and The consistency is included as a factor during iterative adjustment of the first axiom.

14. The information processing system of claim 8, wherein the computer network is a cognitive multi-agent system comprising a set of independent nodes, and wherein the processor performs further actions comprising: Based on an axiom set including the first axiom and the second axiom, enabling the group of independent nodes including the first node and the second node to maintain an autonomous cognitive process; enabling the set of independent nodes to interact with each other and exchange the set of data points; and A set of networks including the set of independent nodes in the cognitive multi-agent system is enabled to capture a set of voting approval relationships and a set of voting disapproval relationships between the set of independent nodes.

15. A computer program product comprising computer program code, which, when executed by an information processing system, causes the information processing system to perform actions comprising: Applying a first axiom by a first node to a set of data points to generate a first set of outputs, wherein the first axiom represents a local world view of the first node for the data points received by the first node; applying, by a second node, a second axiom to the set of data points to generate a second set of outputs, wherein the first node and the second node are part of a computer network comprising a plurality of nodes, wherein the second axiom represents a local world view of the second node for the data points received by the second node; calculating a first nuance of the first node based on a set of differences between the first output set and the second output set; calculating the reliability of the first node in the computer network based on the first nuance; as well as The first axiom is iteratively updated based on the first nuance and the reliability of the first node.

16. The computer program product of claim 15, wherein the information processing system performs further actions comprising: capturing the first set of outputs and the second set of outputs within a time period; and calculating the set of differences between the first set of outputs and the second set of outputs over the time period, wherein: The first nuance is a variance of the set of differences between the first set of outputs and the second set of outputs over the time period.

17. The computer program product of claim 15, wherein calculating the reliability of the first node in the computer network based on the first nuance comprises: The reliability of the first node is iteratively calculated based on a set of edge weights corresponding to a set of neighboring nodes and the first nuance, until the reliability meets a convergence condition based on the set of edge weights.

18. The computer program product of claim 17, wherein iteratively updating the first axiom based on the first nuance and the reliability of the first node comprises: Before iteratively calculating the reliability of the first node, the first axiom is iteratively adjusted based on a weighted sum of a set of iterative nuance calculations to converge the first axiom and the first nuance.

19. The computer program product of claim 17, wherein iteratively updating the first axiom based on the first nuance and the reliability of the first node comprises: In response to iteratively calculating the reliability of the first node, the first axiom is iteratively adjusted to converge the first axiom, the first nuance, and the reliability.

20. The computer program product of claim 19, wherein the information processing system performs further actions comprising: tracking a set of adjustments to the first axiom over a period of time; determining consistency of the first node based on the set of adjustments; and The consistency is included as a factor during iterative adjustment of the first axiom.

21. A system for maintaining data integrity, comprising modules respectively configured to perform the steps of the method according to any one of claims 1 to 7.

Images