Test and measurement instrument and method for knowledge set identification and selection

Abstract

To operate a AI assistant with limited resources.SOLUTION: The test and measurement instrument 12 includes one or more memories, a generative AI (AI) model 22 having access to the one or more memories, a display, user controls, and one or more processors to access an AI (API) of a AI assistant for the generative AI model 22 and to receive one or more user inputs as prompts via the one or more APIs or a user interface; Access the master vector database 16 to obtain a list of master candidates, compare the prompt to the list of master candidates to select a candidate from the master candidates, transmit the selected candidate to the vector database 18, receive a particular candidate from the vector database 18, transmit the prompt and the particular candidate to the generative AI model 22 to receive a response, and display the response on the display.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to artificial intelligence systems, and more particularly to artificial intelligence (AI) assistants. [Background technology]

[0002] Currently, AI assistants and chatbots generally function using large language models (LLMs) and other forms of generative AI. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Unexamined Japanese Patent Publication No. 2020-118689 [Non-patent literature]

[0004] [Non-Patent Document 1] "Tektronix oscilloscope" introduction site, Tektronix, [online], [searched June 26, 2025], Internet<https: / / www.tek.com / ja / products / oscilloscopes> Summary of the Invention [Problem to be solved by the invention]

[0005] Current LLMs are extremely resource intensive. During training, they require vast amounts of data to train properly. At runtime, they typically require extensive cloud-based processing and memory resources, as well as the power to run these resources. [Means for solving the problem]

[0006] Embodiments of the present application generally use small, large-scale language models (LLMs) in limited computing environments. These embodiments allow these small, offline LLMs to perform as well as or better than large, cloud-based LLMs. This requires careful management of the information provided to the LLMs.

[0007] As used herein, the term "generative model" refers to a generative model, such as an LLM or other type of model, that can respond to prompts and create responses to a user. The term "knowledge set" refers to the building blocks of information that can be loaded into and used by a generative model. The term "prompt manager" refers to the AI ​​assistant interface that receives input from the user, handles messaging between elements of the system, and manages the language of prompts. This interface is accessible through the assistant application programming interface (API).

[0008] Embodiments of the present application define a method for isolating core pieces of information / knowledge and define several possible methods for loading only the information necessary to provide a response to a prompt. These methods, according to embodiments of the disclosed technology, generally involve dividing knowledge into separately loadable pieces and one or more methods for determining when to load which bits of knowledge. Embodiments provide focused information without extraneous information, reducing the likelihood of hallucinations.

[0009] The ability to control the loading of smaller knowledge sets allows for a level of control not currently available. Knowledge sets and their associated information can be loaded as needed, such as in situations where export controls such as the International Traffic in Arms Regulations (ITAR), security (certain parties have access to information, but not others), or other sensitive information needs to be avoided. Embodiments allow for the loading or metering of multiple knowledge chunks based on the user's security level or the user's authorized access to information, such as subscriptions or subscriptions made by a particular user or organization, or the level of administrative privileges on the system. Embodiments allow the system to retain smaller, yet more focused, data, thereby improving performance in smaller computing environments.

[0010] Specific examples of these types of applications may include use in test and measurement environments. For example, controlling multiple knowledge sets allows users to differentiate data sets for specific equipment, models, and license sets, allowing the AI ​​assistant to operate more efficiently and limit irrelevant responses. Managing multiple equipment data sets simultaneously can help troubleshoot complex multi-equipment setups and pinpoint the situation.

[0011] In one embodiment, the system identifies and loads data sets based on user characteristics, which may include the user's security clearance, system access privileges, license status, user role, and user location, such as the user's country. In one embodiment, the system may also identify and load data sets based on external factors, such as the equipment used, applicable regulations, applications used, and licenses associated with the applications.

[0012] For an AI assistant to be useful in many commercial environments, it needs domain knowledge. In the test and measurement equipment domain, this knowledge would cover a wide range of topics. These would include, but are not limited to, programming APIs for specific equipment, general knowledge of signals (amplitude, data rate, rise / fall, etc.), and specific protocol knowledge (for example, how to identify and set up I2C (inter-integrated circuit) signal decoding, or specific problem-solving knowledge such as standard issues with I2C and how to identify them). Other examples include knowledge of specific setups of multiple devices, including Bode plots, compliance verification, and other examples, as well as knowledge of the currently enabled features of the test and measurement equipment (for example, ITAR, licensing, whether the test and measurement equipment has arbitrary function generation capabilities, etc.).

[0013] The knowledge needed in this domain can be thought of as a relatively small set of knowledge. Being able to identify these small sets, as well as store them as small, identifiable pieces of information, allows the AI ​​assistant to focus only on the knowledge sets necessary for the current request. This limits the opportunities for hallucination. As used herein, "hallucination" refers to instances in which a generative AI model generates information that is erroneous, inaccurate, or completely fabricated. Typically, this situation results from the generative AI model misinterpreting the input. One way that hallucination can occur is when there is too much information for the AI ​​model to properly absorb. Limiting the amount of information and focusing the scope of information can mitigate this problem. Smaller knowledge sets speed up the validation of candidate responses, further eliminating hallucination.

[0014] In addition to the above benefits, breaking down knowledge into smaller, more manageable pieces allows these knowledge sets to be dynamically loaded as needed. This reduces bandwidth issues within the system, making the model more agile and responsive. The system can also track the use of knowledge sets used in telemetry or measurement. Knowledge set measurement can generate usage fees for billing. Furthermore, these smaller knowledge sets, and the systems that employ them, may be packaged and sold as options, including ITAR compliance, or bundled with other options for equipment. These smaller knowledge sets may also be updated with new data and individually redistributable. Note that while the knowledge sets are smaller and more manageable, access to the knowledge base itself as a whole remains unchanged. All information remains available, but it is accessed in a more manageable and faster manner. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 shows an embodiment of a test and measurement device in an artificial intelligence system. [Figure 2] Figure 2 shows a message flow diagram for an embodiment of the artificial intelligence system. [Figure 3] FIG. 3 shows a flowchart of an embodiment of a method for identifying and selecting a data set in an artificial intelligence system. DETAILED DESCRIPTION OF THE INVENTION

[0016] Embodiments of the disclosed technology may be implemented in a variety of ways. Below are some example embodiments according to various embodiments, but embodiments of the disclosed technology are not necessarily limited to this list of possible embodiments. In a first embodiment, a master knowledge dictionary maps requests to knowledge sets. This master database does not contain response information, but only references to the knowledge sets needed to respond to an answer. In this implementation, the mapping allows the system to load the referenced knowledge sets for telemetry to understand the most used knowledge sets, measure access to the referenced knowledge sets, and look up responses within the specified knowledge sets.

[0017] Another embodiment may include knowledge data sets combined with knowledge for identifying source information by implementing a unified master knowledge base that returns both responses and knowledge categories, allowing for telemetry as described above, measuring access to referenced knowledge sets, examining responses within specified knowledge sets, verifying that responses to requests are from referenced knowledge sets, and ensuring that responses are not hallucinations.

[0018] Yet another embodiment may include a hybrid of the two, which may monitor the accuracy of the response. Having both a combined knowledge base and smaller, more focused knowledge allows for validation of the user's potential responses, thus ensuring that the answer contains the knowledge that would be expected to be included in the response.

[0019] One specific use case for retrieval-based dynamic data set identification is to combine it with another search to optimize generation. Retrieval Augmented Generation (RAG) uses a vector database to accept a prompt from the user and respond with candidates containing information about the prompt's topic. As described below with respect to Figure 2, these candidates are fed back to the model along with the initial prompt to improve the generative AI. Importantly, this approach can be used to inexpensively provide data to the model without retraining. A dynamic data set identification system allows the use of many partitioned vector databases within a single RAG system. Furthermore, the prompt and the identified candidates, as well as the prompt and knowledge set used in response to the prompt, may be stored as vectors in the vector database for further reference on future prompts that may be similar.

[0020] FIG. 1 illustrates an embodiment of an artificial intelligence system. While many of the examples described herein involve a user in a test and measurement environment, a user 10 may operate in any environment that has a computing device with a generative AI component and some interface to that component. Furthermore, the test and measurement system may consist of what many users would consider a test and measurement device, such as an oscilloscope or multimeter. This test and measurement device 12 may include any of these types of devices. Test and measurement devices typically include one or more processors, memory, a user interface for user input and operation, and a display.

[0021] As noted above, deployments of this system may include small-scale environments in which all elements of the system reside on the test and measurement instrument 12. Alternatively, only the interface may reside on the test and measurement instrument 12, with the remaining elements of the system residing on one or more separate computing devices, each with one or more processors. The above examples represent two ends of the spectrum; the system may reside anywhere in between. For example, in one embodiment, the AI ​​assistant / prompt manager 14 resides on the test and measurement instrument 12, along with the generative model 22, master vector database 16, and vector database 18. The knowledge set 20 may reside on the test and measurement instrument 12 and in a distributed manner elsewhere. This example merely highlights how dividing the knowledge set into smaller pieces can alleviate bandwidth and speed issues when the entire knowledge base, consisting of all the knowledge sets, is loaded onto the test and measurement instrument 12 for each user query.

[0022] FIG. 2 provides a message flow diagram of one embodiment of how the prompt manager (which may be part of or may constitute the entire AI assistant 14) uses an initial user prompt to obtain a "master candidate," which is itself a purpose-built vector database. The master vector database 16 consists of one database of databases or vectors of vectors in the AI ​​space. The master vector database loads only the obtained master candidates, ensuring that the model receives only the most relevant information in the second phase of the search. The master candidate may be a single vector or may consist of multiple vectors of information, or other structures, where each vector in the master vector contains anywhere from one to many knowledge sets.

[0023] The master vector database can select master candidates in a variety of ways. When a system implements "explicit configuration," it may query configured devices to determine details such as model, license, enabled features, and location (as required by ITAR). "Implicit configuration" takes a broader approach, less specific than the explicit version. This may include loading master candidates depending on the test measurement device connected to the current test measurement device, rather than directly querying the connected test measurement device. This configuration may also consider input / output usage as tracked by telemetry, or it may load the "standard" workspace for the test measurement device.

[0024] Similarly, a selection may be made from an explicit user context, such as a prompt asking, "What do I do with my Tektronix MSO6 series?", which loads a specific master candidate. A more implicit context may load a roughly correct master candidate from a suggestion involving a function or action, such as loading an oscilloscope with arbitrary waveform generator (AWG) capability when the user asks, "How do I configure my AWG to generate a jittered clock to test the robustness of my digital input signal?" In one embodiment, it may be useful to create master candidates for generic devices, such as oscilloscope candidates, AWG candidates, etc.

[0025] FIG. 2 illustrates an embodiment of a "message" flow, where the term "message" is used for requests and responses between the various components of FIG. 1. The process begins with a user 10 entering an initial prompt, such as the example prompt described above, on a test and measurement instrument or other computing device 12. A prompt manager 14 then indexes the prompt, identifies keywords within the prompt, and sends the prompt to a master vector database 16. The master vector database 16 then identifies master candidates in some manner, as described above, and retrieves a set of master candidates, including candidates composed of other candidates. The prompt manager 14 compares the retrieved master candidates with the prompt and sends those associated with the selected prompt to a vector database 18. The vectors in the vector database 18 contain a list of knowledge sets 20. The selected candidates from the vector database 18, referred to herein as "specific candidates," are then sent back to the prompt manager 14. The prompt manager 14 then sends the indexed prompt and the retrieved specific candidate to the generative model 22. The generative model 22 then loads the information from the specific candidate and uses that information in combination with other information the model already has to generate a response to the user 10. This response is returned through the prompt manager, which displays the response on the display of the test and measurement instrument 12.

[0026] For ease of explanation and understanding, the following example provides a more concrete example, in which the test and measurement equipment 12 comprises an oscilloscope and the AI ​​assistant 14 comprises an oscilloscope AI assistant. The prompt manager 14 collects user prompts, such as, "How do I configure my oscilloscope to decode I2C serial communication on channel number 1?" The prompt manager 14 then indexes the prompts and compares them to the master vector database 16. The master vector database compares the indexed prompts with the indexes of master candidates and returns the best master vector candidate. In this example, the specific Tektronix product used by the user should be the most likely candidate. Depending on the use case and how the master vector database makes its selection, the master vector database may return another Tektronix oscilloscope, or even an oscilloscope from another company that is most similar to the Tektronix oscilloscope used.

[0027] The Prompt Manager 14 receives the indexed initial prompt and compares it to specific candidates within each of the retrieved master candidates. For example, specific candidates could be SCPI (Standard Commands for Programmable Instruments) commands, user manual documentation, or even general waveform and electronics information. In this case, the most similar result might be the SCPI command "AFG:OUTPut:STATE" or the reference to the AFG page in the MSO64 user manual, which would be located in the MSO6 Series master candidate vector database 16. At this point, the initial prompt and the retrieved information are passed to the Model 22. With knowledge of the AFG command set and the MSO64 user manual, the generative AI model should respond more accurately to the initial prompt.

[0028] FIG. 3 shows a flowchart of the process from the perspective of the prompt manager / AI assistant 14. At 30, the prompt manager receives a prompt from a user. The prompt manager then sends the prompt to a master vector database at 32 and receives a set of master candidates as a response at 34. The prompt manager then compares the master candidates to the indexed prompt and selects the best-matching master candidate to send to the vector database at 36. The prompt manager then receives the specific candidate from the vector database at 38. The prompt manager then sends the initial prompt and the specific candidate to a generative model, which then uses the uploaded knowledge set from the specific candidate to generate a response to the user and send it to the user at 42.

[0029] In this way, the system employs smaller generative AI instances and gives users a way to control which knowledge sets are loaded as part of the overall knowledge base, allowing for better control over who has access to what data, allowing the system to run faster without loading the entire knowledge base, and preventing halcation by staying focused on relevant data.

[0030] Aspects of the disclosed technology may operate on specially created hardware, firmware, digital signal processors, or specially programmed general-purpose computers, including processors that operate according to programmed instructions. The terms "controller" or "processor" herein contemplate microprocessors, microcomputers, ASICs, and dedicated hardware controllers, among others. Aspects of the disclosed technology may be implemented with computer-usable data and computer-executable instructions, such as one or more program modules, executed by one or more computers (including a monitoring module) or other devices. Generally, program modules include routines, programs, objects, components, data structures, etc., which, when executed by a processor in a computer or other device, perform particular tasks or implement particular abstract data types. Computer-executable instructions may be stored in computer-readable storage media, such as hard disks, optical disks, removable storage media, solid-state memory, RAM, etc. Those skilled in the art will appreciate that the functionality of the program modules may be combined or distributed as desired in various embodiments. Furthermore, such functionality may be embodied in whole or in part in firmware or hardware equivalents, such as integrated circuits, field programmable gate arrays (FPGAs), etc. Certain data structures may be used to more effectively implement one or more aspects of the disclosed technology, and such data structures are considered within the scope of the computer-executable instructions and computer-usable data described herein.

[0031] The disclosed aspects may, in some cases, be implemented in hardware, firmware, software, or any combination thereof. The disclosed aspects may also be implemented as instructions carried by or stored on one or more computer-readable media, which may be read and executed by one or more processors. Such instructions may be referred to as a computer program product. As used herein, computer-readable media refers to any medium that can be accessed by a computing device. By way of example, and not limitation, computer-readable media may include computer storage media and communication media.

[0032] "Computer storage media" means any medium that can be used to store computer-readable information. By way of example and not limitation, computer storage media may include random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory and other memory technologies, compact disc read-only memory (CD-ROM), digital video disc (DVD) and other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage and other magnetic storage devices, and any other volatile or nonvolatile, removable or non-removable medium implemented in any technology. "Computer storage media" excludes signals themselves and transitory forms of signal transmission.

[0033] A communication medium means any medium usable for communicating computer-readable information. By way of example, and not limitation, communication media may include coaxial cable, fiber optic cable, air, or any other medium suitable for communicating electrical, optical, radio frequency (RF), infrared, acoustic, or other types of signals. Example

[0034] The following examples are provided to aid in understanding the technology disclosed in this application. Embodiments of the technology may include one or more of the examples described below, and any combination thereof.

[0035] Example 1 is a test and measurement device comprising: one or more memories; a generative artificial intelligence (AI) model that can access the one or more memories; a display; a user interface that allows a user to provide input; and one or more processors, the one or more processors being configured to execute a program that causes the one or more processors to perform the following operations: access an application programming interface (API) of an AI assistant for the generative AI model to allow a user to interact with the AI ​​assistant; receive one or more user inputs, including a prompt, via one or more of the API or the user interface; access a master vector database using the prompt to obtain a list of master candidates; compare the prompt with the list of master candidates to identify a selected candidate from the master candidates; send the selected candidate from the master candidates to a vector database; receive a specified candidate from the vector database; send the prompt and the specified candidate to the generative AI model; receive a response from the generative AI model; and display the response on the display.

[0036] Example 2 is a test and measurement instrument of Example 1, wherein the program that causes the one or more processors to perform a process of accessing the master vector database includes a program that causes the one or more processors to perform a process of accessing the master vector database based on explicit information.

[0037] Example 3 is the test and measurement device of Example 2, wherein the explicit information includes at least one of information from a query to the test and measurement device, information from the prompt, a license associated with the test and measurement device, and user context information.

[0038] Example 4 is a test and measurement instrument of any of Examples 1 to 1, wherein the program that causes the one or more processors to perform a process of accessing the master vector database includes a program that causes the one or more processors to perform a process of accessing the master vector database based on implicit information.

[0039] Example 5 is the test and measurement instrument of example 4, wherein the implicit information includes at least one of information regarding devices connected to the test and measurement instrument, input and output usage, and functionality of the test and measurement instrument.

[0040] Example 6 is the test and measurement device of any of Examples 1 to 5, wherein the program that causes the one or more processors to identify the specific candidates includes a program that causes the one or more processors to identify the specific candidates based on user characteristics including one or more of user security clearance, user system access privileges, user license status, user role, and user location.

[0041] Example 7 is the test and measurement device of any of Examples 1 to 6, wherein the program that causes the one or more processors to identify specific candidates includes a program that causes the one or more processors to identify specific candidates based on external factors including one or more of the device used, applicable regulations, software applications used, and licenses associated with the software applications.

[0042] Example 8 is the test and measurement instrument of any of Examples 1 to 7, wherein at least one of the master vector database and the generative AI model resides locally on the test and measurement instrument.

[0043] Example 9 is the test and measurement instrument of any of Examples 1-8, wherein the one or more processors are further configured to identify and package knowledge sets for sale.

[0044] Example 10 is the test and measurement device of any of Examples 1 to 9, wherein the one or more processors are further configured to execute a program that causes the one or more processors to perform a process of measuring knowledge set usage for billing purposes.

[0045] Example 11 is the test and measurement device of any of Examples 1 to 10, wherein the one or more processors are further configured to execute a program that causes the one or more processors to perform a process of updating the model using the prompt and the particular candidate in retrieval augmented generation.

[0046] Example 12 is a method for identifying and selecting a knowledge set for use with a generative AI model, the method comprising: accessing an AI assistant application programming interface (API) for the generative AI model to enable a user to interact with the AI ​​assistant; receiving one or more user inputs, including a prompt, via one or more of the API or the user interface; accessing a master vector database using the prompt to obtain a list of master candidates; comparing the prompt with the list of master candidates to identify a selected candidate from among the master candidates; sending the selected candidate from among the master candidates and the prompt to a vector database; receiving the identified candidate from the vector database; sending the prompt and the identified candidate to the generative AI model; receiving a response from the generative AI model; and displaying the response to the user on the display.

[0047] Example 13 is the method of example 12, wherein the step of accessing the master vector database includes a step of accessing the master vector database based on explicit information.

[0048] Example 14 is the method of Example 13, wherein the explicit information includes at least one of information from a query to the test and measurement device, information from the prompt, a license associated with the test and measurement device, and user context information.

[0049] Example 15 is a method of any one of Examples 12 to 14, wherein the process of accessing the master vector database includes a process of accessing the master vector database based on implicit information.

[0050] Example 16 is the method of example 15, wherein the implicit information includes at least one of information regarding devices connected to the test and measurement device, usage of the inputs and outputs, and functionality of the test and measurement device.

[0051] Example 17 is the method of any of Examples 12 to 16, wherein identifying the specific candidate includes identifying the specific candidate based on user characteristics including one or more of user security clearance, user system access privileges, user license status, user role, and user location.

[0052] Example 18 is any of the methods of Examples 12 to 17, wherein identifying the specific candidate includes identifying the specific candidate based on external factors including one or more of the device used, applicable regulations, software applications used, and licenses associated with the software applications.

[0053] Example 19 is the method of any of Examples 12-18, further comprising identifying and packaging knowledge sets for sale.

[0054] Example 20 is the method of any of examples 12 to 19, further comprising metering usage of the knowledge set for billing purposes.

[0055] Example 21 is the method of any of Examples 12 to 20, further comprising, in search expansion generation, updating the generative AI model using the prompt and the particular candidate.

[0056] All features disclosed in the specification, claims, abstract and drawings, and all steps in any disclosed method or process, may be combined in any combination, except where at least some of such features or steps are mutually exclusive combinations. Each feature disclosed in the specification, abstract, claims and drawings may be replaced by an alternative feature serving the same, equivalent or similar purpose, unless expressly stated otherwise.

[0057] Additionally, the description of this application refers to specific features. It should be understood that the disclosure herein includes all possible combinations of these specific features. When a specific feature is disclosed in connection with a particular aspect or example, that feature can also be used in connection with other aspects and examples, to the extent possible.

[0058] Furthermore, when this application refers to a method having two or more defined steps or processes, these defined steps or processes may be performed in any order or simultaneously, unless the circumstances preclude this possibility.

[0059] Although specific embodiments of the invention have been illustrated and described for purposes of illustration, it will be appreciated that various modifications can be made therein without departing from the spirit and scope of the invention. Accordingly, the invention should not be limited except as by the appended claims. [Explanation of symbols]

[0060] 10 users 12 Test and measurement equipment 14 AI Assistant / Prompt Manager 16 Master Vector Database 18 Vector Database 20 Knowledge Sets 22 Generative Model

Claims

1. 1. A test and measurement device comprising: one or more memories; a generative artificial intelligence (AI) model having access to the one or more memories; The display and a user control that allows a user to provide input; one or more processors Equipped with the one or more processors accessing an AI assistant application programming interface (API) for the generative AI model to enable the user to interact with the AI ​​assistant; receiving one or more user inputs, including prompts, via one or more of the API or the user interface; Accessing the master vector database using the above prompt to obtain a list of master candidates; comparing the prompt with the list of master candidates to identify a selected candidate from among the master candidates; sending a selected candidate from the master candidates to a vector database; receiving a particular candidate from the vector database; sending the prompt and the specific candidate to the generative AI model; receiving a response from the generative AI model; a process of displaying the response on the display; a test and measurement instrument configured to execute a program that causes the one or more processors to:

2. 2. The test and measurement instrument of claim 1, wherein the program that causes the one or more processors to perform processing to access the master vector database includes a program that causes the one or more processors to perform processing to access the master vector database based on explicit information including at least one of information from an inquiry to the test and measurement instrument, information from the prompt, a license associated with the test and measurement instrument, and user contextual information, or based on implicit information including at least one of information regarding devices connected to the test and measurement instrument, input / output usage, and test and measurement instrument functionality.

3. 2. The test and measurement instrument of claim 1, wherein the program that causes the one or more processors to identify specific candidates includes a program that causes the one or more processors to identify specific candidates based on user characteristics including one or more of user security clearance, user system access privileges, user license status, user role, and user location, or based on external factors including one or more of the device used, applicable regulations, software applications used, and licenses associated with the software applications.

4. The test and measurement instrument of claim 1 , wherein the one or more processors are further configured to identify and package knowledge sets for sale.

5. 10. The test and measurement instrument of claim 1, wherein the one or more processors are further configured to execute a program that causes the one or more processors to measure knowledge set usage for billing purposes.

6. 10. The test and measurement instrument of claim 1, wherein the one or more processors are further configured to execute a program that causes the one or more processors to update the model using the prompt and the particular candidate in search expansion generation.

7. Accessing an AI assistant's application programming interface (API) for the generative AI model to allow a user to interact with the AI ​​assistant; receiving one or more user inputs, including prompts, via one or more of the API or the user interface; Accessing the master vector database using the above prompt to obtain a list of master candidates; comparing the prompt with the list of master candidates to identify a selected candidate from among the master candidates; sending the selected candidate from the master candidates and the prompt to a vector database; receiving a particular candidate from the vector database; sending the prompt and the specific candidate to the generative AI model; receiving a response from the generative AI model; A process of displaying the response to the user on a display. A method for identifying and selecting knowledge sets for use in a generative AI model, comprising:

8. 8. The method of claim 7, wherein the process of accessing the master vector database includes a process of accessing the master vector database based on explicit information including at least one of information from a query to the test and measurement device, information from the prompt, a license associated with the test and measurement device, and user contextual information, or based on implicit information including at least one of information about devices connected to the test and measurement device, input and output usage, and capabilities of the test and measurement device.

9. 8. The method of claim 7, wherein identifying specific candidates includes identifying specific candidates based on user characteristics including one or more of user security clearance, user system access privileges, user license status, user role, and user location, or based on external factors including one or more of equipment used, applicable regulations, software applications used, and licenses associated with the software applications.

10. 8. The method of claim 7, further comprising updating the generative AI model using the prompt and the identified candidate in search expansion generation.

Images