DESC: FORM 2

THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003

COMPLETE SPECIFICATION
(See section 10 and rule 13)
1. TITLE OF THE INVENTION

SYSTEM AND METHOD FOR INFERENCING ONE OR MORE TRAINED MODELS
2. APPLICANT(S)
NAME NATIONALITY ADDRESS
JIO PLATFORMS LIMITED INDIAN OFFICE-101, SAFFRON, NR. CENTRE POINT, PANCHWATI 5 RASTA, AMBAWADI, AHMEDABAD 380006, GUJARAT, INDIA
3.PREAMBLE TO THE DESCRIPTION

THE FOLLOWING SPECIFICATION PARTICULARLY DESCRIBES THE NATURE OF THIS INVENTION AND THE MANNER IN WHICH IT IS TO BE PERFORMED.

FIELD OF THE INVENTION
[0001] The present invention relates to the field of network data analytics for predictive network management and, more specifically, to a system and a method thereof to store inference data generated by a user while performing inference on selected models which are already trained or are being trained in a user interactive manner.
BACKGROUND OF THE INVENTION
[0002] With the increase in number of users, the network service provisions have been implementing to up-gradations to enhance a service quality so as to keep pace with such high demand. With advancement of technology, there is a demand for a telecommunication service to induce up-to-date features into the scope of provision. To enhance user experience and implement advanced monitoring mechanisms, prediction methodologies are being incorporated in a network management service. An advanced prediction system integrated with an Artificial intelligence (AI)/Machine learning (ML) system excels in executing a wide array of algorithms and predictive tasks. The advanced prediction system is underpinned by the capabilities of Large Language Models (LLMs). Its primary mission centers around the comprehensive analysis of both network data and operational data, capitalizing on the advanced techniques of machine learning (ML) to glean profound insights. The LLM retrains the existing trained model for new data source or existing data source.
[0003] An advanced AI/ML integrated system is capable of data analysis and making predictions. There are trained models developed by the users to perform specific predictions or analyses. Some AI/ML predictive systems implement large language models (LLMs) as a service, enabling the retraining of existing models such as Generative Pre-Trained Transformers-Jumbo (GPT-J), Large Language Model Meta AI2 (LAMA2), Bloom, Generative Pre-Trained Transformers -neo (GPT-neo) and Falcon etc with either existing or new data sources. This approach helps the users save time and optimize system resources. In this regard, storing relevant data like what sources are used for training, what sort of training methodology is used, what output testing and inference records are obtained, are really important for future data analysis. The data analysis can be tedious process and the users are required to know about the data analysis method and machine learning when machine learning models are deployed for this purpose.
[0004] In a system configured to give the user the option of inference after selecting trained models, the users may ask questions to the models, which will generate answers based on the training. However, contemporary practices do not store the generated answers and conversations between different users and models. Only the input and output are retained for future reference. As a result, important inference data is lost. Hence, there is a need for a system and method to store the lost inference data to assist in future analysis, enabling the study of user inference and the analysis of trained models.
SUMMARY OF THE INVENTION
[0005] One or more embodiments of the present disclosure provide a system and a method for inferencing one or more trained models.
[0006] In one aspect of the present invention, the method for inferencing one or more trained models is disclosed. The method includes retrieving, by one or more processors, data from at least one of: an existing data source or a new data source. Further, the method includes categorizing, by the one or more processors, the data as at least one of: historic data and current data. Further, the method includes pre-processing, by the one or more processors, the historic data and the current data. Further, the method includes autotuning, by the one or more processors, one or more hyperparameters for the pre-processed historic data and the current data. Further, the method includes retraining, by the one or more processors, the one or more trained model with the historic data and the current data. The historic data and the current data are pre-processed and autotuned with the one or more hyperparameters. Further, the method includes generating, by the one or more processors, utilizing the one or more re-trained models, one or more inferences based on a user inputting one or more search queries via a user interface. Further, the method includes storing, by the one or more processors, the one or more inferences generated at a storage unit.
[0007] In an embodiment, the method further includes notifying, by the one or more processors, the user, a status of re-training the one or more trained models. The status of notifying the user of training the one or more models, includes at least one of, status of completion of retraining the existing trained one or more models utilizing one or more identifiers including at least one of, training name, model name, version, type and/or name of the data source used and one or more actions including at least one of, retrain or delete the trained model.
[0008] In an embodiment, the one or more inferences are generated in a format including at least one of, a text based format.
[0009] In an embodiment, storing, the one or more inferences generated at the storage unit, further includes the step of generating, by the one or more processors, a searchable inference catalogue of the one or more inferences to enable the one or more users to search for the one or more inferences generated. The searchable inference catalogue enables at least one of, the one or more users or third party users to search the one or more inferences.
[0010] In an embodiment, further the method includes storing, by the one or more processors, one or more conversational chats of the one or more users while re-training the one or more trained models. Further, the method includes generating, by the one or more processors, a searchable conversational catalogue pertaining to the one or more conversational chats of the one or more users while re-training the one or more models. The searchable conversational catalogue enables the one or more users to search the one or more conversational chats irrespective of the one or more conversational chats generated based on the said one or more users providing inputs as one or more prompts via the user interface.
[0011] In an embodiment, the method further includes, the one or more processors enables the one or more users to browse the inference data including conversations of the one or more users during re-training the one or more trained models.
[0012] In one aspect of the present invention, the system for inferencing one or more trained modelsis disclosed. The system includes a retrieving unit, a categorizing unit, a pre-processing unit, a tuning unit, a training unit, a generating unit and a notifying unit. The retrieving unit is configured to retrieve, data from at least one of, an existing data source or a new data source. The categorizing unit is configured to categorize the data as at least one of, historic data and current data. The pre-processing unit is configured to automatically pre-process, the historic data and the current data. The tuning unit is configured to autotune, one or more hyperparameters for the pre-processed historic data and the current data. The training unit is configured to retrain the one or more trained model with the historic data and the current data. The historic data and the current data are pre-processed and autotuned with the one or more hyperparameters. The generating unit is configured to, generate, utilizing the one or more re-trained models, one or more inferences based on a user inputting one or more search queries via a user interface. The generating unit is configured to, store the one or more inferences generated at a storage unit.
[0013] In one aspect of the present invention, a non-transitory computer-readable medium having stored thereon computer-readable instructions is provided. The computer-readable instructions causes the processor to retrieve, data from at least one of: an existing data source or a new data source. Further, the processor categorizes the data as at least one of, historic data and current data. Further, the processor pre-processes the historic data and the current data. Further, the processor autotunes one or more hyperparameters for the pre-processed historic data and the current data. Further, the processor retrains the one or more trained model with the historic data and the current data. The historic data and the current data are pre-processed and autotuned with the one or more hyperparameters. Further, the processor generates, utilizing the one or more re-trained models, one or more inferences based on a user inputting one or more search queries via a user interface. Further, the processor stores the one or more inferences generated at a storage unit.
[0014] Other features and aspects of this invention will be apparent from the following description and the accompanying drawings. The features and advantages described in this summary and in the following detailed description are not all-inclusive, and particularly, many additional features and advantages will be apparent to one of ordinary skill in the relevant art, in view of the drawings, specification, and claims hereof. Moreover, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes and may not have been selected to delineate or circumscribe the inventive subject matter, resort to the claims being necessary to determine such inventive subject matter.

BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The accompanying drawings, which are incorporated herein, and constitute a part of this disclosure, illustrate exemplary embodiments of the disclosed methods and systems in which like reference numerals refer to the same parts throughout the different drawings. Components in the drawings are not necessarily to scale, emphasis instead being placed upon clearly illustrating the principles of the present disclosure. Some drawings may indicate the components using block diagrams and may not represent the internal circuitry of each component. It will be appreciated by those skilled in the art that disclosure of such drawings includes disclosure of electrical components, electronic components or circuitry commonly used to implement such components.
[0016] FIG. 1 is an exemplary block diagram of an environment for inferencing one or more trained models, according to various embodiments of the present disclosure.
[0017] FIG. 2 is a block diagram of a system of FIG. 1, according to various embodiments of the present disclosure.
[0018] FIG. 3 is an example schematic representation of the system of FIG. 1 in which various entities operations are explained, according to various embodiments of the present system.
[0019] FIG. 4 illustrates a system architecture for inferencing the one or more trained models, according to various embodiments of the present system.
[0020] FIG. 5 is an exemplary flow diagram illustrating the method for inferencing the one or more trained models, according to various embodiments of the present disclosure.
[0021] FIG. 6 is an example flow diagram illustrating the method for inferencing the one or more trained models, according to various embodiments of the present disclosure.
[0022] Further, skilled artisans will appreciate that elements in the drawings are illustrated for simplicity and may not have necessarily been drawn to scale. For example, the flow charts illustrate the method in terms of the most prominent steps involved to help to improve understanding of aspects of the present invention. Furthermore, in terms of the construction of the device, one or more components of the device may have been represented in the drawings by conventional symbols, and the drawings may show only those specific details that are pertinent to understanding the embodiments of the present invention so as not to obscure the drawings with details that will be readily apparent to those of ordinary skill in the art having benefit of the description herein.
[0023] The foregoing shall be more apparent from the following detailed description of the invention.

DETAILED DESCRIPTION OF THE INVENTION
[0024] Some embodiments of the present disclosure, illustrating all its features, will now be discussed in detail. It must also be noted that as used herein and in the appended claims, the singular forms "a", "an" and "the" include plural references unless the context clearly dictates otherwise.
[0025] Various modifications to the embodiment will be readily apparent to those skilled in the art and the generic principles herein may be applied to other embodiments. However, one of ordinary skill in the art will readily recognize that the present disclosure including the definitions listed here below are not intended to be limited to the embodiments illustrated but is to be accorded the widest scope consistent with the principles and features described herein.
[0026] A person of ordinary skill in the art will readily ascertain that the illustrated steps detailed in the figures and here below are set out to explain the exemplary embodiments shown, and it should be anticipated that ongoing technological development will change the manner in which particular functions are performed. These examples are presented herein for purposes of illustration, and not limitation. Further, the boundaries of the functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternative boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed. Alternatives (including equivalents, extensions, variations, deviations, etc., of those described herein) will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein. Such alternatives fall within the scope and spirit of the disclosed embodiments.
[0027] Before discussing example, embodiments in more detail, it is to be noted that the drawings are to be regarded as being schematic representations and elements that are not necessarily shown to scale. Rather, the various elements are represented such that their function and general purpose becomes apparent to a person skilled in the art. Any connection or coupling between functional blocks, devices, components, or other physical or functional units shown in the drawings or described herein may also be implemented by an indirect connection or coupling. A coupling between components may also be established over a wireless connection. Functional blocks may be implemented in hardware, firmware, software or a combination thereof.
[0028] Further, the flowcharts provided herein, describe the operations as sequential processes. Many of the operations may be performed in parallel, concurrently or simultaneously. In addition, the order of operations maybe re-arranged. The processes may be terminated when their operations are completed, but may also have additional steps not included in the figured. It should be noted, that in some alternative implementations, the functions/acts/ steps noted may occur out of the order noted in the figured. For example, two figures shown in succession may, in fact, be executed substantially concurrently, or may sometimes be executed in the reverse order, depending upon the functionality/acts involved.
[0029] Further, the terms first, second etc… may be used herein to describe various elements, components, regions, layers and/or sections, it should be understood that these elements, components, regions, layers and/or sections should not be limited by these terms. These terms are used only to distinguish one element, component, region, layer or section from another region, layer, or a section. Thus, a first element, component, region layer, or section discussed below could be termed a second element, component, region, layer, or section without departing form the scope of the example embodiments.
[0030] Spatial and functional relationships between elements (for example, between modules) are described using various terms, including “connected,” “engaged,” “interfaced,” and “coupled.” Unless explicitly described as being “direct,” when a relationship between first and second elements is described in the description below, that relationship encompasses a direct relationship where no other intervening elements are present between the first and second elements, and also an indirect relationship where one or more intervening elements are present (either spatially or functionally) between the first and second elements. In contrast, when an element is referred to as being "directly” connected, engaged, interfaced, or coupled to another element, there are no intervening elements present. Other words used to describe the relationship between elements should be interpreted in a like fashion (e.g., "between," versus "directly between," "adjacent," versus "directly adjacent," etc.).
[0031] The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting. Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which example embodiments belong. It will be further understood that terms, e.g., those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
[0032] 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. As used herein, the terms “and/or” and “at least one of” include any and all combinations of one or more of the associated listed items. It will be further understood that the terms “comprises,” “comprising,” “includes,” and/or “including,” when used herein, 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.
[0033] Unless specifically stated otherwise, or as is apparent from the description, terms such as “processing” or “computing” or “calculating” or “determining” of “displaying” or the like, refer to the action and processes of a computer system, or similar electronic computing device/hardware, that manipulates and transforms data represented as physical, electronic quantities within the computer system's registers and memories into other data similarly represented as physical quantities within the computer system memories or registers or other such information storage, transmission or display devices.
[0034] Various embodiments of the present invention provide a system and a method to store inference data obtained after inferring models selected by users which are already trained or are being trained.
[0035] FIG. 1 illustrates an exemplary block diagram of an environment (100) for inferencing the one or more trained models (e.g., ML model, AI model, supervised learning, deep learning, LLM or the like), according to various embodiments of the present disclosure. The environment (100) comprises a plurality of user equipment’s (UEs) (102-1, 102-2, ……,102-n). The at least one UE (102-n) from the plurality of the UEs (102-1, 102-2, ……102-n) is configured to connect to a system (108) via a communication network (106). Hereafter, label for the plurality of UEs or one or more UEs is 102.
[0036] In accordance with yet another aspect of the exemplary embodiment, the plurality of UEs (102) may be a wireless device or a communication device that may be a part of the system (108). The wireless device or the UE (102) may include, but are not limited to, a handheld wireless communication device (e.g., a mobile phone, a smart phone, a phablet device, and so on), a wearable computer device (e.g., a head-mounted display computer device, a head-mounted camera device, a wristwatch, a computer device, and so on), a laptop computer, a tablet computer, or another type of portable computer, a media playing device, a portable gaming system, and/or any other type of computer device with wireless communication or Voice Over Internet Protocol (VoIP) capabilities. In an embodiment, the UEs (102) may include, but are not limited to, any electrical, electronic, electro-mechanical or an equipment or a combination of one or more of the above devices such as virtual reality (VR) devices, augmented reality (AR) devices, laptop, a general-purpose computer, desktop, personal digital assistant, tablet computer, mainframe computer, or any other computing device, where the computing device may include one or more in-built or externally coupled accessories including, but not limited to, a visual aid device such as camera, audio aid, a microphone, a keyboard, input devices for receiving input from a user such as touch pad, touch enabled screen, electronic pen and the like. It may be appreciated that the UEs (102) may not be restricted to the mentioned devices and various other devices may be used. A person skilled in the art will appreciate that the plurality of UEs (102) may include a fixed landline, and a landline with assigned extension within the communication network (106).
[0037] The communication network (106), may use one or more communication interfaces/protocols such as, for example, Voice Over Internet Protocol (VoIP), 802.11 (Wi-Fi), 802.15 (including Bluetooth™), 802.16 (Wi-Max), 802.22, Cellular standards such as Code Division Multiple Access (CDMA), CDMA2000, Wideband CDMA (WCDMA), Radio Frequency Identification (e.g., RFID), Infrared, laser, Near Field Magnetics, etc.
[0038] The communication network (106) includes, by way of example but not limitation, one or more of a wireless network, a wired network, an internet, an intranet, a public network, a private network, a packet-switched network, a circuit-switched network, an ad hoc network, an infrastructure network, a Public-Switched Telephone Network (PSTN), a cable network, a cellular network, a satellite network, a fiber optic network, or some combination thereof. The communication network (106) may include, but is not limited to, a Third Generation (3G) network, a Fourth Generation (4G) network, a Fifth Generation (5G) network, a Sixth Generation (6G) network, a New Radio (NR) network, a Narrow Band Internet of Things (NB-IoT) network, an Open Radio Access Network (O-RAN), and the like.
[0039] The communication network (106) may also include, by way of example but not limitation, at least a portion of one or more networks having one or more nodes that transmit, receive, forward, generate, buffer, store, route, switch, process, or a combination thereof, etc. one or more messages, packets, signals, waves, voltage or current levels, some combination thereof, or so forth. The communication network (106) may also include, by way of example but not limitation, one or more of a wireless network, a wired network, an internet, an intranet, a public network, a private network, a packet-switched network, a circuit-switched network, an ad hoc network, an infrastructure network, a Public-Switched Telephone Network (PSTN), a cable network, a cellular network, a satellite network, a fiber optic network, a VOIP or some combination thereof.
[0040] One or more network elements can be, for example, but not limited to a base station that is located in the fixed or stationary part of the communication network (106). The base station may correspond to a remote radio head, a transmission point, an access point or access node, a macro cell, a small cell, a micro cell, a femto cell, a metro cell. The base station enables transmission of radio signals to the UE (102) or a mobile transceiver. Such a radio signal may comply with radio signals as, for example, standardized by a 3rd Generation Partnership Project (3GPP) or, generally, in line with one or more of the above listed systems. Thus, a base station may correspond to a NodeB, an eNodeB, a Base Transceiver Station (BTS), an access point, a remote radio head, a transmission point, which may be further divided into a remote unit and a central unit. The 3GPP specifications cover cellular telecommunications technologies, including radio access, core network, and service capabilities, which provide a complete system description for mobile telecommunications.
[0041] The system (108) is communicatively coupled to a server (104) via the communication network (106). The server (104) can be, for example, but not limited to a standalone server, a server blade, a server rack, an application server, a bank of servers, a business telephony application server (BTAS), a server farm, a cloud server, an edge server, home server, a virtualized server, one or more processors executing code to function as a server, or the like. In an implementation, the server (104) may operate at various entities or a single entity (include, but is not limited to, a vendor side, a service provider side, a network operator side, a company side, an organization side, a university side, a lab facility side, a business enterprise side, a defense facility side, or any other facility) that provides service.
[0042] The environment (100) further includes the system (108) communicably coupled to the server (e.g., remote server or the like) (104) and each UE of the plurality of UEs (102) via the communication network (106). The remote server (104) is configured to execute the requests in the communication network (106).
[0043] The system (108) is adapted to be embedded within the remote server (104) or is embedded as an individual entity. The system (108) is designed to provide a centralized and unified view of data and facilitate efficient business operations. The system (108) is authorized to access to update/create/delete one or more parameters of their relationship between the requests for training the model, which gets reflected in real-time independent of the complexity of network.
[0044] In another embodiment, the system (108) may include an enterprise provisioning server (for example), which may connect with the remote server (104). The enterprise provisioning server provides flexibility for enterprises, ecommerce, finance to update/create/delete information related to the requests for the training the model in real time as per their business needs.
[0045] The system (108) may include, by way of example but not limitation, one or more of a standalone server, a server blade, a server rack, a bank of servers, a business telephony application server (BTAS), a server farm, hardware supporting a part of a cloud service or system, a home server, hardware running a virtualized server, one or more processors executing code to function as a server, one or more machines performing server-side functionality as described herein, at least a portion of any of the above, some combination thereof. In an implementation, system (108) may operate at various entities or single entity (for example include, but is not limited to, a vendor side, service provider side, a network operator side, a company side, an organization side, a university side, a lab facility side, a business enterprise side, ecommerce side, finance side, a defense facility side, or any other facility) that provides service.
[0046] However, for the purpose of description, the system (108) is described as an integral part of the remote server (104), without deviating from the scope of the present disclosure. Operational and construction features of the system (108) will be explained in detail with respect to the following figures.
[0047] FIG. 2 illustrates a block diagram of the system (108) provided for inferencing one or more trained models (e.g., AI model, ML model (such as LLMs), or the like), according to one or more embodiments of the present invention. As per the illustrated embodiment, the system (108) includes the one or more processors (202), the memory (204), an input/output interface unit (206), a display (208), an input device (210), and the database (214). Further the system (108) may comprise one or more processors (202). The one or more processors (202), hereinafter referred to as the processor (202) may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, single board computers, and/or any devices that manipulate signals based on operational instructions. As per the illustrated embodiment, the system (108) includes one processor. However, it is to be noted that the system (108) may include multiple processors as per the requirement and without deviating from the scope of the present disclosure.
[0048] An information related to the trained model may be provided or stored in the memory (204) of the system (108). Among other capabilities, the processor (202) is configured to fetch and execute computer-readable instructions stored in the memory (204). The memory (204) may be configured to store one or more computer-readable instructions or routines in a non-transitory computer-readable storage medium, which may be fetched and executed to create or share data packets over a network service. The memory (204) may include any non-transitory storage device including, for example, volatile memory such as RAM, or non-volatile memory such as disk memory, EPROMs, FLASH memory, unalterable memory, and the like.
[0049] The memory (204) may comprise any non-transitory storage device including, for example, volatile memory such as Random-Access Memory (RAM), or non-volatile memory such as Electrically Erasable Programmable Read-only Memory (EPROM), flash memory, and the like. In an embodiment, the system (108) may include an interface(s). The interface(s) may comprise a variety of interfaces, for example, interfaces for data input and output devices, referred to as input/output (I/O) devices, storage devices, and the like. The interface(s) may facilitate communication for the system. The interface(s) may also provide a communication pathway for one or more components of the system. Examples of such components include, but are not limited to, processing unit/engine(s) and the database (214). The processing unit/engine(s) may be implemented as a combination of hardware and programming (for example, programmable instructions) to implement one or more functionalities of the processing engine(s).
[0050] The information related to the trained model may further be configured to render on the user interface (206). The user interface (206) may include functionality similar to at least a portion of functionality implemented by one or more computer system interfaces such as those described herein and/or generally known to one having ordinary skill in the art. The user interface (206) may be rendered on the display (208), implemented using Liquid Crystal Display (LCD) display technology, Organic Light-Emitting Diode (OLED) display technology, and/or other types of conventional display technology. The display (208) may be integrated within the system (108) or connected externally. Further the input device(s) (210) may include, but not limited to, keyboard, buttons, scroll wheels, cursors, touchscreen sensors, audio command interfaces, magnetic strip reader, optical scanner, etc.
[0051] The database (214) may be communicably connected to the processor (202) and the memory (204). The database (214) may be configured to store and retrieve the request pertaining to features, or services or workflow of the system (108), access rights, attributes, approved list, and authentication data provided by an administrator. In another embodiment, the database (214) may be outside the system (108) and communicated through a wired medium and a wireless medium.
[0052] Further, the processor (202), in an embodiment, may be implemented as a combination of hardware and programming (for example, programmable instructions) to implement one or more functionalities of the processor (202). In the examples described herein, such combinations of hardware and programming may be implemented in several different ways. For example, the programming for the processor (202) may be processor-executable instructions stored on a non-transitory machine-readable storage medium and the hardware for the processor (202) may comprise a processing resource (for example, one or more processors), to execute such instructions. In the present examples, the memory (204) may store instructions that, when executed by the processing resource, implement the processor (202). In such examples, the system (108) may comprise the memory (204) storing the instructions and the processing resource to execute the instructions, or the memory (204) may be separate but accessible to the system (108) and the processing resource. In other examples, the processor (202) may be implemented by an electronic circuitry.
[0053] In order for the system (108) to inference the one or more trained models, the processor (202) includes a retrieving unit (216), a categorizing unit (218), a pre-processing unit (220), a tuning unit (222), a training unit (224), a notifying unit (226), a generating unit (228) and a storage unit (230). The retrieving unit (216), the categorizing unit (218), the pre-processing unit (220), the tuning unit (222), the training unit (224), the notifying unit (226), the generating unit (228) and the storage unit (230) may be implemented as a combination of hardware and programming (for example, programmable instructions) to implement one or more functionalities of the processor (202). In the examples described herein, such combinations of hardware and programming may be implemented in several different ways. For example, the programming for the processor (202) may be processor-executable instructions stored on a non-transitory machine-readable storage medium and the hardware for the processor (202) may comprise a processing resource (for example, one or more processors), to execute such instructions. In the present examples, the memory (204) may store instructions that, when executed by the processing resource, implement the processor. In such examples, the system (108) may comprise the memory (204) storing the instructions and the processing resource to execute the instructions, or the memory (204) may be separate but accessible to the system (108) and the processing resource. In other examples, the processor (202) may be implemented by the electronic circuitry.
[0054] In order for the system (108) to inference the one or more trained models, the retrieving unit (216), the categorizing unit (218), the pre-processing unit (220), the tuning unit (222), the training unit (224), the notifying unit (226), the generating unit (228) and the storage unit (230) are communicably coupled to each other.
[0055] The retrieving unit (216) retrieves data from an existing data source or a new data source. The existing data source can be, for example, but not limited to Generative Pre-Trained Transformers-Jumbo (GPT-J), Large Language Model Meta AI2 (LAMA2), Bloom, Generative Pre-Trained Transformers -neo (GPT-neo) and Falcon etc. In an embodiment, the retrieving unit (216) retrieves the data from the existing data source by selecting the existing data source via the user interface (206) by the user. In an example, if the user selects the existing data source then, the user of the system (108) will select the data source from the list of existing data sources.
[0056] In another embodiment, the retrieving unit (216) creates the new data source when the user selects one or more data sources from a list of data sources via the user interface (206). Further, the retrieving unit (216) pulls the data from one or more data sources and stores the data at the new data source. Further, the retrieving unit (216) retrieves the data from the new data source. In simple terms, if the user of the system (108) selects the new data source then, the user of the system (108) creates the new data source by retrieving the data from data as file input or the data from the source path, an input stream, a Hypertext Transfer Protocol (HTTP), Hadoop Distributed File Systems (HDFS) and a network attached storage (NAS).
[0057] Further, the categorizing unit (218) categorizes the data as at least one of historic data and current data. In an embodiment, the data is categorized as at least one of: the historic data and the current data based on a time of generation of the data.
[0058] In an example, consider a retail shop company that uses the machine learning model to predict sales trends based on the historical data and the current data. The historic data means sales data from the last five years (e.g., daily sales figures, customer details or the like). This data is categorized based on the time it was generated, so any data prior to the last six months is labeled as historic. The current data means the sales data from the last six months, which reflects recent trends and changes in consumer behavior. The current data includes daily sales figures, promotional events, seasonal influences or the like. When the new sales data is generated, the categorizing unit (218) analyzes the timestamp of this data. If the data is from today or the past six months, the categorizing unit (218) categorizes the data as current data. If the data is older than six months, the categorizing unit (218) categorizes the data as the historic data.
[0059] The pre-processing unit (220) automatically pre-processes the historic data and the current data. In an embodiment, the pre-processing unit (220) cleans the historic data and the current data by removing at least one of, duplicates and correcting data types. Further, the pre-processing unit (220) transforms the cleaned historic and the current data by at least one of converting categorical variables into numerical format and encoding the data. Further, the pre-processing unit (220) removes irrelevant or redundant features from the historic and the current data. Further, the pre-processing unit (220) normalizes the historic data and the current data by converting one or more tokens to a base format. The one or more tokens are present in at least one of the existing data source or the new data source.
[0060] In an example, consider, a company wants to improve its sentiment analysis model based on the customer reviews. The dataset includes both historic data (from previous years) and current data (recent submissions). The reviews from the previous years may contain inconsistent formatting (e.g., "Great product!" vs. "great Product"). The pre-processing step would normalize these to a consistent format, such as converting all text to lowercase. Also, the recent reviews may include slang or abbreviations (e.g., "LOL" or "BFF"). The normalizing could involve replacing these with their full meanings or using a predefined dictionary to standardize expressions.
[0061] Further, older reviews might contain HTML or XML tags (e.g., Great product!). The pre-processing would involve stripping these tags to retain only the text content. Similar formatting issues can occur with recent reviews, especially if the users copy and paste from different platforms. These tags would also be removed to standardize the dataset.
[0062] Further, the reviews might include irrelevant information such as product codes or author details that do not contribute to sentiment analysis. These would be removed to streamline the data. Also, the recent reviews may include promotional links or unrelated comments. The pre-processing would involve identifying and removing these elements (e.g., promotional links or unrelated comments).
[0063] In an example, the reviews might contain repeated phrases or emojis that add noise. The pre-processing step would involve removing the repeated phrases or emojis to ensure the model focuses on the sentiment conveyed.
[0064] In an example, the pre-processing unit (220) has normalized both the historic and current data by converting the purchase amounts to a base format (for example, USD) and standardizing product categories. This ensures consistency and allows the machine learning model to effectively learn from the combined dataset, improving its accuracy in predicting customer behavior.
[0065] The tuning unit (222) autotunes one or more hyperparameters for the pre-processed historic data and the current data based. In an example, the tuning unit (222) effectively automates a hyperparameter tuning process, so as to improve the ML model accuracy by utilizing both historical and current datasets. This approach helps ensure that the ML model is robust and well-suited to predict customer churn accurately.
[0066] For example, during the hyperparameter tuning, the tuning unit (222) uses a GridSearchCV to autotune the hyperparameters. In an example,
from sklearn.model_selection import GridSearchCV
param_grid = {
'n_estimators': [100, 200],
'max_depth': [None, 10, 20],
'min_samples_split': [2, 5, 10]
}
grid_search = GridSearchCV(model, param_grid, cv=5)
grid_search.fit(historical_data.drop('target', axis=1), historical_data['target'])
[0067] Further, the training unit (224) retrains the one or more trained models with the historic data and the current data. The historic data and the current data are pre-processed and autotuned with the one or more hyperparameters. After the pre-processing and autotuning, the cleaned dataset (comprising both historic and current data) can be fed into the machine learning model. This improves the model's accuracy and effectiveness in understanding sentiment by reducing variability and noise in the data.
[0068] The generating unit (228) generates, utilizing the one or more re-trained models, one or more inferences based on the user inputting one or more search queries via the user interface (206). Further, the generating unit (228) stores the one or more inferences generated at the storage unit (230). The one or more inferences stored in the storage unit (230) includes inference data generated by the one or more users for the one or more re-trained models. In an embodiment, the generating unit (228) enables the one or more users to browse the inference data including conversations of the one or more users during re-training the one or more trained models. In an embodiment, the generating unit (228) generates a searchable inference catalogue of the one or more inferences to enable the one or more users or a third party user to search for the one or more inferences generated. The searchable inference catalogue enables at least one of, the one or more users or third party users to search the one or more inferences. The one or more inferences are generated in a format including at least one text based format.
[0069] In an example, consider, a medical diagnosis system powered by the ML model that analyzes patient symptoms to provide potential diagnoses. The generating unit (228) is responsible for creating the searchable inference catalogue based on the inferences (e.g., diagnoses) generated by the ML model. The inference catalogue contains various diagnoses (e.g., inferences) generated from the ML model. Each entry includes details like diagnosis name (e.g., "Flu", "Diabetes"), Symptoms associated with the diagnosis, and recommended tests or treatments. The searchable inference catalogue is designed to allow the users (e.g., doctors, healthcare professionals or the like) to search for inferences based on user requirement (e.g. treatments, test or the like).
[0070] Based on the proposed method, regardless of whether the inferences were generated based on a specific patient’s symptoms or a general dataset, the searchable inference catalogue allows the doctor to retrieve relevant diagnoses. The system (108) quickly pulls up the list of inferences (e.g., "Flu", "Mononucleosis", "Chronic Fatigue Syndrome") that match the search terms. Further, the results show multiple potential diagnoses with details like symptoms. The doctor can click on each diagnosis to view more information, such as treatment options or additional tests that might be necessary.
[0071] Further, the generating unit (228) stores one or more conversational chats of the one or more users while re-training the one or more trained models. Further, the generating unit (228) generates a searchable conversational catalogue pertaining to the one or more conversational chats of the one or more users while re-training the one or more models. The searchable conversational catalogue enables the one or more users to search the one or more conversational chats irrespective of the one or more conversational chats generated based on the said one or more users providing inputs as one or more prompts via the user interface (206).
[0072] In an example, consider, the customer support chatbot uses the ML model to handle the user inquiries. While retraining the ML model, the generating unit (228) stores multiple conversational chats between the users and the chatbot. The conversational catalogue includes user queries (e.g., "How do I reset my password?"), chatbot responses, timestamps, user satisfaction ratings and user interaction via the user interface (206). Further, if the user types "password reset" in the search bar then, the system (108) retrieves relevant past chats that include discussions about password resets, regardless of whether they originated from that user or others. Further, the user sees a list of previous conversations that include "How do I reset my password?" "Can you help me with my account access?" based on the proposed method, the users can quickly find helpful information from past interactions without generating new conversations, improving the efficiency of support queries.
[0073] The notifying unit (226) notifies a status of retraining the one or more trained model to the user. In an embodiment, the notifying unit (226) notifies status of completion of retraining the one or more trained models utilizing one or more identifiers. The one or more identifier can be, for example, but not limited to training name, model name, version, type and/or name of the data source used. The notifying unit (226) performs the one or more actions. The one or more action can be for example but not limited to retrain the trained model or delete the trained model.
[0074] In an example, a company regularly updates its machine learning model for predicting customer behaviour based on the new data or the existing data. The training process is automated, and the team needs to be notified once the training is completed. The notifying unit (226) is responsible for sending the notifications to the users about the status of the model training. The notifying unit (226) is used to specify which model has been trained and relevant details (e.g., Model Name: CustomerChurnPredictor, Version: 2.1, Training Name: Analysis_2024, Data Source: Customer_Dataset_2024, Status: Completed Successfully). The example for training the model is explained in FIG. 4 to FIG. 6.
[0075] FIG. 3 is an example schematic representation of the system (300) of FIG. 1 in which various entities operations are explained, according to various embodiments of the present system. It is to be noted that the embodiment with respect to FIG. 3 will be explained with respect to the first UE (102-1) and the system (108) for the purpose of description and illustration and should nowhere be construed as limited to the scope of the present disclosure.
[0076] As mentioned earlier, the first UE (102-1) includes one or more primary processors (305) communicably coupled to the one or more processors (202) of the system (108). The one or more primary processors (305) are coupled with a memory (310) storing instructions which are executed by the one or more primary processors (305). Execution of the stored instructions by the one or more primary processors (305) enables the UE (102-1). The execution of the stored instructions by the one or more primary processors (305) further enables the UE (102-1) to execute the requests in the communication network (106).
[0077] As mentioned earlier, the one or more processors (202) is configured to transmit a response content related to the trained model to the UE (102-1). More specifically, the one or more processors (202) of the system (108) is configured to transmit the response content to at least one of the UE (102-1). A kernel (315) is a core component serving as the primary interface between hardware components of the UE (102-1) and the system (108). The kernel (315) is configured to provide the plurality of response contents hosted on the system (108) to access resources available in the communication network (106). The resources include one of a Central Processing Unit (CPU), memory components such as Random Access Memory (RAM) and Read Only Memory (ROM).
[0078] As per the illustrated embodiment, the system (108) includes the one or more processors (202), the memory (204), the input/output interface unit (206), the display (208), and the input device (210). The operations and functions of the one or more processors (202), the memory (204), the input/output interface unit (206), the display (208), and the input device (210) are already explained in FIG. 2. For the sake of brevity, we are not explaining the same operations (or repeated information) in the patent disclosure. Further, the processor (202) includes retrieving unit (216), the categorizing unit (218), the pre-processing unit (220), the tuning unit (222), the training unit (224), the notifying unit (226), the generating unit (228) and the storage unit (230). The operations and functions of the retrieving unit (216), the categorizing unit (218), the pre-processing unit (220), the tuning unit (222), the training unit (224), the notifying unit (226), the generating unit (228) and the storage unit (230) are already explained in FIG. 2. For the sake of brevity, we are not explaining the same operations (or repeated information) in the patent disclosure.
[0079] FIG. 4 illustrates a system architecture (400) for inferencing one or more trained models, according to various embodiments of the present system. The system architecture (400) includes the one or more processors (202), the memory (204), the input/output interface unit (206), the display (208), and the input device (210). The operations and functions of the one or more processors (202), the memory (204), the input/output interface unit (206), the display (208), and the input device (210) are already explained in FIG. 2. For the sake of brevity, we are not explaining the same operations (or repeated information) in the patent disclosure. Further, the processor (202) includes retrieving unit (216), the categorizing unit (218), the pre-processing unit (220), the tuning unit (222), the training unit (224), the notifying unit (226), the generating unit (228) and the storage unit (230). The operations and functions of the retrieving unit (216), the categorizing unit (218), the pre-processing unit (220), the tuning unit (222), the training unit (224), the notifying unit (226), the generating unit (228) and the storage unit (230) are already explained in FIG. 2. For the sake of brevity, we are not explaining the same operations (or repeated information) in the patent disclosure.
[0080] Further, the system architecture (400) includes the system (108) configured to interact with an integrated system (402) via a load balancer (404) and a data-lake (406). The system (108) is integrated with a Large Language Model (LLM) to provide provisions for optimal retaining.
[0081] The integrated system (402) collects raw data from different data sources. The load balancer (404) distributes the data source request traffic across the system architecture (400). The input device (210) is for taking the inputs from the user. By means of the input device (210), the user will give the training name for the model name. The model name is selected by the user from the list of models like GPT-J, LAMA2, Bloom, GPT-neo and Falcon etc. The LLM as a service sets the training version by default for a specified training name (for example). The user will select the execution group where the training is going to execute from the list provided by the LLM as a service. Further, the system (108) allows an entry to a new data source. In simple terms, if the user selects the new data source then, the system (108) creates a new data source by retrieving the data from data as file input, data from the source path, data as the input stream, data from the HTTP2, data from the HDFS (Hadoop Distributed File Systems) and data from NAS (network attached storage).
[0082] If the user selects existing data source then, the user will select the data source from the list of existing data sources. Further, the system (108) pre-processes the data received to normalize and clean it. On the basis of the historic data and the current data, the LLM as a service does the preprocessing such as cleaning and normalizing the text content while removing formatting tags and irrelevant elements from the data source which contain various formatting elements such as headings, tables, footnotes, page numbers, other structural components and images. The cleaning and normalizing the data is performed by removing the noise from extra rows which contain invalid column values, such as NaN, None, 0, null, or empty strings. By using the tuning unit (222), the system (108) with LLM as a service automatically evaluates the user's data, identifies the most suitable hyper-parameters to the ML algorithm that is most likely to yield the best results. The LLM as a service auto tune the hyper-parameters based on data size, model weight size and suitable for provided input. The training unit (224) executes a training schedule for the ML models. The ML training is performed on the input data. The display (208) displays a training status list which contains tabular view of the training name, the model name, the version, the data source type, the status and action like retrain and delete.
[0083] The generating unit (228) selects the training name and version from the created models list on which inference is to be performed. The generating unit (228) are configured to enable the user to ask questions to the trained model and LLM as a service gives the answer to that questions. Also, LLM as a service can generate the text to the context provided by the user. Also, the generating unit (228) inferences the ML model selected by the user. The generating unit (228) selects the training name and version from the created models list on which inference is to be performed. Further, the generating unit (228) is configured to enable the user to ask questions to the trained model and LLM as a service gives the answer to that question. Also, the LLM as a service can generate the text to the context provided by user.
[0084] Further, the system (108) is connected to the data-lake (406) which is a distributed database used to store the processed data and algorithm outputs. The LLM as a service stores the trained model by performing retrain of existing model on new data source or existing data source which can be used by other users also for further retraining and inference. The data-lake (406) also stores all conversations between different users and different models. This helps the user to study the different types of data sources and their trends and patterns after inferencing the model. The data security is also maintained for the stored data which gives user the assurance that data will be maintained properly.
[0085] The system (108) introduces LLM as a service to create a new data source and train the data source on the provided input and to enable a user to retrain the model with the existing data source or the new data source on different existing trained model. This helps the user to get better results for provided data source with the new data source or the existing data source. The data analysis can be a tedious process and users require knowledge about this. Further, the system (108) having LLM as a service, gives the option of inference to the user after selecting created trained models. The user asks questions to the trained models and it will generate the answer on the basis of trained models. Also, the text generation is possible from all available trained models in the network (106). The LLM also configured to store acquired inference data incorporating all the conversation in between the user and the inference unit for future analysis and this data is accessible by every authenticated users using the network.
[0086] The system (108) is configured to interact with an external and internal data sources (not shown). The system (108) may further include one or more database and is capable of interacting with one or more application server in the network (106).
[0087] Further, the system (108) may be configured to interact with various component of the network and external network by means of various Application Programming Interface (API), databases and servers or any other compatible element. The databases (214)/data lakes (406) are configured to store past data, dynamic data, and trained models for future necessity. The databases (214)/data lakes (406) are configured to store past data, dynamic data, and trained models for future necessity. The present system may further be configured to incorporate even more data into pre-processing steps if required to refine the data analysis. The pre-processing step involves extracting and normalizing the data by applying suitable operation filter, normalization, cleaning and standardization of data.
[0088] The system (108) may further be configured to incorporate even more data into pre-processing steps if required to refine the data analysis. The pre-processing step involves extracting and normalizing the data by applying suitable operation filter, normalization, cleaning and standardization of data.
[0089] In an example, the user, by means of the present system (108), first creates a training name and selects the model name. The user will give the training name. The training name (or model name) is selected by the user from the list of models like GPT-J, LAMA2, Bloom, GPT-neo and Falcon. The system (108) sets the version by the default. The LLM as a service sets the training version by default for the specified training name. For example, if the user given training name as trainingname1 then, a first time version will be set to version1.1 by default. When a second time training name is given as TrainingName1 then, the version will be set as version1.2 by default.
[0090] After creating the training name and setting of the version, the user will select the execution group where the training is going to execute from the list provided by the LLM as a service. The execution group is the group of different servers. Whatever execution groups are accessible to that particular user is displayed in the list and user has to select the execution from the list. The ML training will be performed on the selected execution group. After that, the user has to select the data source. If the user selects a new data source then, the system (108) creates a new data source by retrieving the data from the data as file input, the data from source path, the data as input stream, the data from HTTP2, the data from HDFS and the data from NAS.
[0091] If the user selects existing data source then, the user will select the data source from the list of existing data sources. The user has to select the source path. The data path is determined by providing valid credentials for the new data source and for the existing data source. Further, the user has to select the data source from drop down list of the LLM. On the basis of historic and current data, the LLM as a service does the preprocessing such as cleaning and normalizing the text content while removing formatting tags and irrelevant elements from the data source which contain various formatting elements such as headings, tables, footnotes, page numbers, other structural components and images.
[0092] The LLM also performs cleaning and normalizing on the data by removing the noise from extra rows which contain invalid column values, such as NaN, None, 0, null, or empty strings. The system (108) then executes ML training. Further, the system (108) then performs the auto-tuning of the hyper-parameters based on the data size, the model weight size and suitable for provided input. Then, the required ML training is performed on the given data by the system (108). The LLM as a service stores the trained model by performing retrain of existing model on new data source or existing data source which can be used by other users also for further retraining and inference. The LLM as a service displays the training status list which contains tabular view of the training name, the model name, the version, the data source type, the status and the action like retrain and delete. For inferring, the user has to select the training name and version from the created models list on which inference is to be performed then LLM as service inferences the selected created models. The user may ask questions to the trained model and LLM as a service gives answer to those questions. Also, LLM as a service can generate the text to the context provided by user. The system (108) having the LLM as service stores all conversations between different users and different models. This helps the user to study the different types of data sources and their trends and patterns after inferencing the model. The data security is also maintained for the stored data which gives user the assurance that data will be maintained properly in the data-lake (406) associated with the system (108).
[0093] The most unique aspect of this invention is the capability to use data from different sources with different format and type. The various data including data as file input, the data from the source path, the data as input stream, the data from the HTTP2, the data from the HDFS and the data from the NAS is received by the present system (108) with LLM as a service to retrain the existing model with the new data source or the existing data source. Another inventive feature is the ability to retrain an already trained model with existing data or new data as per requirement. The system (108) with LLM as a service provides the option for the user to retrain the existing model GPT-J, LAMA2, Bloom, GPT-neo and Falcon on new data source as well as existing data source. The system (108) gives the trained model after retraining the existing model on the data as file input, the data from source path, the data as input stream, the data from HTTP2, the data from HDFS and the data from NAS. Yet another inventive feature lies in the ability of the present system with LLM as a service is that the system (108) is configured to provide the option to the user to inference the trained model. The questions can be asked by the user of the trained model in the network (106) by the present system with LLM as a service. The system (108) is configured to auto-suggest some questions which can be asked on the trained data which helps the user to infer the trained model easily. The text generation is also possible by the present system (108) when required for further analysis. The system (108) integrated with LLM is configured to store all conversations between different users and different models. The key advantage of this is that all the data is securely stored and can be accessed by the users when required and therefore gives the better approach for inferencing the trained model and get better results for the different trained models.
[0094] The system (108) is configured to receive various data including the data as file input, the data from source path, the data as input stream, the data from HTTP2, the data from HDFS and the data from NAS to retrain the existing model with the new data source or the existing data source. The integration with LLM as a service, provides the option for the user to retrain the existing model like GPT-J, LAMA2, Bloom, GPT-neo and Falcon etc. on the new data source as well as the existing data source. The system (108) having LLM as a service gives the trained model after retraining the existing model on the data as file input, the data from source path, the data as input stream, the data from HTTP2, the data from HDFS and the data from NAS. As the system (108) is configured to automatically preprocess and tune data, system (108) becomes easier for the user to train/retrain the ML model. The system (108) having LLM as a service, also provides a training status list which contains tabular view of the training name, the model name, the version, the data source type, the status and the action. In the action column, the user can delete or retrain the trained model by seeing the training name, the model name and the version. This helps the user to retrain the model as many times on same model with new data source or existing data source with different versions.
[0095] The system (108) may further be configured to interact with the application servers, an IPM (integrated performance management) system, an FMS (Fulfillment Management System), an NMS (network management system) modules in the network (106) via an application programming interface (API) as medium of communication and may perform the process by means of various formats like JSON(JavaScript Object Notation), Python or any other compatible formats.
[0096] FIG. 5 is an exemplary flow diagram (500) illustrating the method for inferencing one or more trained models, according to various embodiments of the present disclosure.
[0097] At 502, the method includes retrieving the data from at least one of: the existing data source or the new data source. In an embodiment, the method allows the retrieving unit (216) to retrieve data from at least one of: the existing data source or the new data source.
[0098] At 504, the method includes categorizing the data as at least one of the historic data and the current data. In an embodiment, the method allows the categorizing unit (218) to categorize, the data as at least one of: the historic data and the current data.
[0099] At 506, the method includes pre-processing the historic data and the current data. In an embodiment, the method allows the pre-processing unit (220) to pre-process the historic data and the current data.
[00100] At 508, the method includes autotuning the one or more hyperparameters for the pre-processed historic data and the current data. In an embodiment, the method allows the tuning unit (222) to autotune the one or more hyperparameters for the pre-processed historic data and the current data.
[00101] At 510, the method includes retraining the one or more trained model with the historic data and the current data. The historic data and the current data are pre-processed and autotuned with the one or more hyperparameters. In an embodiment, the method allows the training unit (224) to retrain the one or more trained model with the historic data and the current data.
[00102] At 512, the method includes generating, utilizing the one or more re-trained models, one or more inferences based on the one or more users inputting the one or more search queries via the user interface (206). In an embodiment, the method allows the generating unit (228) to, generate, utilizing the one or more re-trained models, the one or more inferences based on the one or more users inputting the one or more search queries via the user interface (206).
[00103] At 514, the method includes storing the one or more inferences generated at the storage unit (230). In an embodiment, the method allows the generating unit (228) to store the one or more inferences generated at the storage unit (230).
[00104] FIG. 6 is an example flow diagram (600) illustrating the method for inferencing one or more trained models, according to various embodiments of the present disclosure. The method is executed by the present system (108) is integrated with LLM as a service to infer selected models which are already trained or are being trained, by any authenticated user as well as to retrain the existing trained model with efficient pre-processing of data obtained from existing or new data source.
[00105] At 602, the user of the system (108) creates the training name and selects the model name from the list of models like GPT-J, LAMA2, Bloom, GPT-neo and Falcon, by using the user interface (206). At 604, the system (108) automatically sets the version by default for the particular training name. For example, if the user is given the training name as TrainingName1 then, the first time version will be set to version1.1 by default. When the second time training name is given as TrainingName1 then, the version will be set as version1.2 by default.
[00106] At 606, the user of the system (108) selects the execution group. The execution group is the group of different servers. Whatever the execution groups are accessible to that particular user is displayed in the list and the user has to select the execution from the list. The ML training will be performed on selected execution group.
[00107] At 608, the user of the system (108) selects the data source as existing or new. If the user selects the new data source then, the system (108) creates the new data source by retrieving the data from data as file input, the data from source path, the data as input stream, the data from HTTP2, the data from HDFS and the data from NAS. If the user selects the existing data source then user will select the data source from the list of existing data sources.
[00108] At 610, the system (108) with LLM as service preprocesses the data source. On the basis of historic and current data does the preprocessing. The cleaning and normalizing the text content is performed while removing formatting tags and irrelevant elements from the data source which contain various formatting elements such as headings, tables, footnotes, page numbers, other structural components and images. The cleaning and normalizing the data is performed by removing the noise from extra rows which contain invalid column values, such as NaN, None, 0, null, or empty strings.
[00109] At 612, the system (108) automatically tunes the hyper-parameters based on the data size, the model weight size and suitable for provided input. At 614, the system (108) executes the ML training. At 616, the system (108) with LLM as a service stores the trained model by performing retrain of existing model on the new data source or the existing data source. At 618, the system (108) displays the training status list which contains tabular view of training name, model name, version, data source type, status and action like retrain and delete. At 620, the user selects the training name and version from the created models list on which inference is to be performed. Further, the system (108) with LLM as service inferences selected created models. The user can ask questions and LLM as a service will provide answer on the basis of selected trained model and generates text generation if prompted by the user.
[00110] At 622, the system (108) stores the one or more inferences of the trained model (e.g., first inference of trained model, second inference of trained model, nth inference of trained model or the like) at the storage unit (230). At 624, the system (108) having LLM as service stores all conversations between different users and different models. This helps the user to study the different types of data sources and their trends and patterns after inferencing the model. The data security is also maintained for the stored data which gives user the assurance that data will be maintained properly in data-lake.
[00111] In preferred embodiments, the method may also include various steps to collect information from network elements like servers and other network functions, trigger consecutive operational procedures etc., improve learning methodology for retraining the machine learning models and may not be considered strictly limited to the above method steps.
[00112] Below is the technical advancement of the present invention:
[00113] The proposed system (108) can be used to store the inference data securely. The proposed system (108) enables the authenticated user to perform inference on the selected models and to browse stored inference data including conversation of the user while training the ML models. The proposed system (108) provides an interactive inference option and stores the changes as well as user conversations for future analysis. The method executed by the proposed system (108) involving steps to store data related to model inferences existing for future requirement, thus optimizing data analysis.
[00114] The proposed system (108) has a capability to train the ML model using various data sources. The proposed system (108) with the LLM as a service provides retraining of the existing trained model on the new data source or the existing data source. After training, the system (108) is configured to provide the option to save the trained model and can be used by other users also for further inference.
[00115] The proposed system (108) with the LLM as service gives user the option of inference of the trained model. The user can ask questions to the selected trained model and can get results for the same. The system (108) is also configured for generating the text for given prompt/context. This helps user to analyze the efficiency of the data based on the results. The inference data is stored by the present system with LLM as a service. The system (108) stores all conversations between different users and different models. This helps user to understand different types of data source by different data source and retraining of the data source on existing trained model by different users. Also, inference on different models can also be observed by the other users and can analyze the trained models.
[00116] The proposed system (108) offers improved efficiency and faster operation. This efficiency improvement not only saves time but also enhances the overall performance of the algorithms. By auto-tuning the hyper-parameters, the system having LLM as a service achieve better accuracy, robustness, and adaptability to different data sources.
[00117] A person of ordinary skill in the art will readily ascertain that the illustrated embodiments and steps in description and drawings (FIGS. 1-6) are set out to explain the exemplary embodiments shown, and it should be anticipated that ongoing technological development will change the manner in which particular functions are performed. These examples are presented herein for purposes of illustration, and not limitation. Further, the boundaries of the functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternative boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed. Alternatives (including equivalents, extensions, variations, deviations, etc., of those described herein) will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein. Such alternatives fall within the scope and spirit of the disclosed embodiments.
[00118] Method steps: A person of ordinary skill in the art will readily ascertain that the illustrated steps are set out to explain the exemplary embodiments shown, and it should be anticipated that ongoing technological development will change the manner in which particular functions are performed. These examples are presented herein for purposes of illustration, and not limitation. Further, the boundaries of the functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternative boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed. Alternatives (including equivalents, extensions, variations, deviations, etc., of those described herein) will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein. Such alternatives fall within the scope and spirit of the disclosed embodiments.
[00119] The present invention offers multiple advantages over the prior art and the above listed are a few examples to emphasize on some of the advantageous features. The listed advantages are to be read in a non-limiting manner.

REFERENCE NUMERALS
[00120] Environment - 100
[00121] UEs– 102, 102-1-102-n
[00122] Server - 104
[00123] Communication network – 106
[00124] System – 108
[00125] Processor – 202
[00126] Memory – 204
[00127] User Interface – 206
[00128] Display – 208
[00129] Input device – 210
[00130] Database – 214
[00131] Retrieving unit– 216
[00132] Categorizing unit – 218
[00133] Pre-processing unit – 220
[00134] Tuning unit - 222
[00135] Training unit – 224
[00136] Notifying unit – 226
[00137] Generating unit – 228
[00138] Storage unit - 230
[00139] System - 300
[00140] Primary processors -305
[00141] Memory– 310
[00142] Kernel– 315
[00143] System architecture – 400
[00144] Integrated system – 402
[00145] Load balancer – 404
[00146] Data-lake - 406

,CLAIMS:CLAIMS:
We Claim:
1. A method for inferencing one or more trained models, the method comprising the steps of:
retrieving, by one or more processors, data from at least one of, an existing data source or a new data source;
categorizing, by the one or more processors, the data as at least one of, historic data and current data;
pre-processing, by the one or more processors, the historic data and the current data;
autotuning, by the one or more processors, one or more hyperparameters for the pre-processed historic data and the current data;
re-training, by the one or more processors, the one or more trained models with the historic data and the current data, wherein the historic data and the current data are pre-processed and autotuned with the one or more hyperparameters;
generating, by the one or more processors, utilizing the one or more re-trained models, one or more inferences based on one or more users inputting one or more search queries via a user interface; and
storing, by the one or more processors, the one or more inferences generated at a storage unit.

2. The method as claimed in claim 1, wherein the method further comprises the step of:
notifying, by the one or more processors, the user, a status of re-training the one or more models, wherein the status of notifying the user of training the one or more models, includes at least one of, status of completion of retraining the existing trained one or more models utilizing one or more identifiers including at least one of, training name, model name, version, type and/or name of the data source used and one or more actions including at least one of, retrain or delete the trained model.

3. The method as claimed in claim 1, wherein the step of, storing, the one or more inferences generated at a storage unit, further includes the step of:
generating, by the one or more processors, a searchable inference catalogue of the one or more inferences to enable the one or more users to search for the one or more inferences generated, wherein the searchable inference catalogue enables at least one of, the one or more users or third party users to search the one or more inferences.

4. The method as claimed in claim 1, wherein the method further comprises the steps of:
storing, by the one or more processors, one or more conversational chats of the one or more users while re-training the one or more trained models; and
generating, by the one or more processors, a searchable conversational catalogue pertaining to the one or more conversational chats of the one or more users while re-training the one or more models, wherein the searchable conversational catalogue enables the one or more users to search the one or more conversational chats irrespective of the one or more conversational chats generated based on the said one or more users providing inputs as one or more prompts via the user interface.

5. The method as claimed in claim 1, wherein the one or more inferences stored in the storage unit includes inference data generated by the one or more users for the one or more re-trained models.

6. The method as claimed in claim 5, wherein the one or more processors enables the one or more users to browse the inference data including conversations of the one or more users during re-training the one or more trained models.

7. A system (108) for inferencing one or more trained models, the system (108) comprising:
a retrieving unit (216) configured to retrieve, data from at least one of, an existing data source or a new data source;
a categorizing unit (218), configured to categorize, the data as at least one of, historic data and current data;
a pre-processing unit (220), configured to pre-process, the historic data and the current data;
a tuning unit (222), configured to, autotune, one or more hyperparameters for the pre-processed historic data and the current data;
a training unit (224), configured to, re-train, the one or more trained models with the historic data and the current data, wherein the historic data and the current data are pre-processed and autotuned with the one or more hyperparameters;
a generating unit (228), configured to, generate, utilizing the one or more re-trained models, one or more inferences based on one or more users inputting one or more search queries via a user interface; and
the generating unit (228), configured to, store, the one or more inferences generated at a storage unit.

8. The system (108) as claimed in claim 7, wherein the system (108) further comprising:
a notifying unit (226), configured to, notify the user, a status of re-training the one or more models, wherein the status of notifying the user of training the one or more models, includes at least one of, status of completion of retraining the existing trained one or more models utilizing one or more identifiers including at least one of, training name, model name, version, type and/or name of the data source used and one or more actions including at least one of, retrain or delete the trained model.

9. The system (108) as claimed in claim 7, wherein the generating unit (228), generates a searchable inference catalogue of the one or more inferences to enable the one or more users to search for the one or more inferences generated, wherein the searchable inference catalogue enables at least one of, the one or more users or third party users to search the one or more inferences.

10. The system (108) as claimed in claim 7, wherein the generating unit (228) is further configured to:
store, one or more conversational chats of the one or more users while re-training the one or more trained models; and
generate, a searchable conversational catalogue pertaining to the one or more conversational chats of the one or more users while re-training the one or more models, wherein the searchable conversational catalogue enables the one or more users to search the one or more conversational chats irrespective of the one or more conversational chats generated based on the said one or more users providing inputs as one or more prompts via the user interface.

11. The system (108) as claimed in claim 7, wherein the one or more inferences stored in the storage unit includes inference data generated by the one or more users for the one or more re-trained models.

12. The system (108) as claimed in claim 11, wherein the generating unit (228) enables the one or more users to browse the inference data including conversations of the one or more users during re-training the one or more trained models.