WO2023223289A1

WO2023223289A1 - A method for generating ml model and a ml model generating system thereof

Info

Publication number: WO2023223289A1
Application number: PCT/IB2023/055194
Authority: WO
Inventors: Seshadri Manivannan; Satyanarayana THOGALAGANTI; Harit Nagpal; Vishal Arya
Original assignee: Tata Play Limited
Priority date: 2022-05-19
Filing date: 2023-05-19
Publication date: 2023-11-23

Abstract

Embodiments of present disclosure relates to a method and a Machine Learning (ML) model generating system for generating a ML model. The method includes receiving a user request to generate ML model. Thereafter, the method includes providing one or more functionalities on a ML operation platform, associated with the ML model generating system, for user selection after receiving the user request. Further, the method includes initializing one of the one or more functionalities selected by a user. Finally, the method includes generating a ML model based on SDK and launch instance function parameters on the selected functionality. Thus, the present disclosure helps the user to select a best-fit ML model by providing suggestion in terms of functionalities, SDK and function parameters, leading to obtaining accurate results. Also, the present disclosure optimizes infrastructure, security, and a framework-based approach by implementing the ML model generating system on any server including, but not limited to, public cloud server, enterprise datacentre and the like for user access.

Description

A METHOD FOR GENERATING ML MODEL AND A ML MODEL GENERATING SYSTEM THEREOF

TECHNICAL FIELD

The present subject matter is related in general to Machine Learning (ML) techniques, but not exclusively, the present subject matter relates to a method and a system for generating a ML model.

BACKGROUND

Currently Machine Learning (ML) allows organizations, enterprises and data scientists to perform tasks on a scale and scope that was previously impossible to achieve. Because the ML techniques possess a capability to speed up the pace of work, reduce errors and improve accuracy, these techniques assist an employee and a customer alike. Hence, many organizations or enterprises are finding ways to adopt concepts on ML technology or deep learning techniques to achieve benefits.

In the existing ML methods, the organizations, the enterprises and the data scientists perform their task in an ad-hoc manner, and there is no particular platform-based approach for adopting new concepts of ML technologies. The reason being ML techniques are either proprietary products, which are very expensive or there are no certain solutions, which can work on a public cloud.

Hence, there is a need for designing a ML platform for organisations, enterprises and data scientists, where the platform is not tied up with any proprietary technology or cloud solution provider. Also, there is a need of developing an open framework that can work on any data centre/server or any other public cloud.

The information disclosed in this background of the disclosure section is only for enhancement of understanding of the general background of the invention and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art already known to a person skilled in the art.

SUMMARY

In an embodiment, the present disclosure relates to a method for generating a Machine Learning (ML) model. The method comprises receiving a user request to generate an ML model. The method comprises providing one or more functionalities on a ML operation platform for user selection upon receiving the user request. The ML operation platform is associated with the ML model generating system. Further, the method comprises initializing one of the one or more functionalities on the ML operation platform selected by a user. Thereafter, the method comprises generating a ML model based on Software Development Kit (SDK) and launch instance function parameters on the selected functionality.

In an embodiment, the present disclosure relates to a ML model generating system for generating a ML model. The ML model generating system includes a processor and a memory communicatively coupled to the processor. The processor is configured to receive a user request to generate an ML model. Thereafter, the processor is configured to provide one or more functionalities on a ML operation platform for user selection upon receiving the user request. The ML operation platform is associated with the ML model generating system. Subsequently, the processor is configured to initialize one of the one or more functionalities on the ML operation platform selected by a user. Thereafter, the processor is configured to generate a ML model based on Software Development Kit (SDK) and launch instance function parameters on the selected functionality.

The foregoing summary is illustrative only and is not intended to be in any way limiting. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features will become apparent by reference to the drawings and the following detailed description.

BRIEF DESCRIPTION OF DRAWINGS

The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, serve to explain the disclosed principles. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. The same numbers are used throughout the figures to reference like features and components. Some embodiments of device or system and/or methods in accordance with embodiments of the present subject matter are now described, by way of example only, and with reference to the accompanying figures, in which: Figure 1 shows an exemplary environment of a ML model generating system for generating a ML model, in accordance with some embodiments of the present disclosure;

Figure 2 shows a detailed block diagram of a ML model generating system for generating an ML model, in accordance with some embodiments of the present disclosure;

Figure 3a illustrate an exemplary architecture of a ML model generating system for generating an ML model, in accordance with some embodiments of the present disclosure;

Figure 3b shows a flow diagram illustrating a method generating an ML model, in accordance with some embodiments of the present disclosure;

Figure 4 illustrates a flowchart showing exemplary method for generating an ML model, in accordance with some embodiments of present disclosure; and

Figure 5 illustrates a block diagram of an exemplary computer system for implementing embodiments consistent with the present disclosure.

It should be appreciated by those skilled in the art that any block diagrams herein represent conceptual views of illustrative systems embodying the principles of the present subject matter. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and executed by a computer or processor, whether such computer or processor is explicitly shown.

DETAILED DESCRIPTION

In the present document, the word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or implementation of the present subject matter described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.

While the disclosure is susceptible to various modifications and alternative forms, specific embodiment thereof has been shown by way of example in the drawings and will be described in detail below. It should be understood, however that it is not intended to limit the disclosure to the forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternative falling within the spirit and the scope of the disclosure.

The terms “comprises”, “comprising”, or any other variations thereof, are intended to cover a nonexclusive inclusion, such that a setup, device, or method that comprises a list of components or steps does not include only those components or steps but may include other components or steps not expressly listed or inherent to such setup or device or method. In other words, one or more elements in a system or apparatus proceeded by “comprises... a” does not, without more constraints, preclude the existence of other elements or additional elements in the system or method.

The terms “includes”, “including”, or any other variations thereof, are intended to cover a nonexclusive inclusion, such that a setup, a device, or a method that includes a list of components or steps does not include only those components or steps but may include other components or steps not expressly listed or inherent to such setup or device or method. In other words, one or more elements in a system or an apparatus proceeded by “includes... a” does not, without more constraints, preclude the existence of other elements or additional elements in the system or the method.

In the following detailed description of the embodiments of the disclosure, reference is made to the accompanying drawings that form a part hereof, and in which are shown by way of illustration specific embodiments in which the disclosure may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the disclosure, and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the present disclosure. The following description is, therefore, not to be taken in a limiting sense.

Present disclosure relates to a method and a ML model generating system for generating a ML model. The method of the present disclosure in order to generate a ML model, initially receives a user request to generate the ML model. Upon receiving the user request, the method of the present disclosure provides one or more functionalities on a ML operation platform for user selection. The ML operation platform is associated with the ML model generating system. Further, the method of the present disclosure initialises one of the one or more functionalities on the ML operation platform that is selected by the user. Upon initializing, the method of the present disclosure generates a ML model based on Software Development Kit (SDK) and launch instance function parameters on the selected functionality. Thus, the present disclosure is able to generate a ML model based on end user request. The method of the present disclosure helps the user to select a best-fit ML model by providing suggestion in terms of functionalities, SDK and launch instance function parameters. As a result, the present disclosure helps in obtaining accurate results. The implementation of the ML model generating system of present disclosure and associated ML operation platform on a cloud server optimizes infrastructure, enhances security, and increases connectivity/accessibility in terms of user outreach (number of users who can access the ML model generating system of the present disclosure). The ML model generating system of the present disclosure is cloud agnostic. As a result, the ML model generating system can be run on any server or any public cloud such as AWS®, Micro Azure®, Google® cloud and the like. The ML model generating system of the present disclosure can be used by users ranging from beginners to proficient data scientists, making the ML model generating system highly scalable and sophisticated system. The ML model generating system of the present disclosure can be implemented on any ML operation platform allowing any users, such as data scientists or statisticians or ML engineers, to write a ML code, run/text the ML code, and deploy the ML code in a secure manner.

Figure 1 shows an exemplary environment 100 for generating a ML model. The environment 100 may include a ML model generating system 101 and a user 102. The user 102 communicates/initiates a user request with the ML model generating system 101 for generating the ML model. The ML model generating system 101 may be implemented on, but is not limited to, a Personal Computer (PC), a laptop computer, a desktop computer, a server, a network server, a cloud-based server, and the like. In one embodiment, the ML model generating system 101 may include a processor 103, an Input/Output (I/O) interface 104, and a memory 105. In some embodiments, the memory 105 may be communicatively coupled to the processor 103. The memory 105 stores instructions, executable by the processor 103, which, on execution, may cause the ML model generating system 101 to generate the ML model. In one embodiment, the user 102 may communicate/initiates a user request via the PC or the laptop computer or the desktop computer with the ML model generating system 101 for generating the ML model when the ML model generating system 101 is implemented on the PC, the laptop computer, the desktop computer, and the like. In another embodiment, the user 102 may communicate/initiates a user request via the PC or the laptop computer or the desktop computer and a communication network (not shown in Figure 1) with the ML model generating system 101 for generating the ML model when the ML model generating system 101 is implemented on the server, the network server, the cloud-based server, and the like. The communication network may include, without limitation, a direct interconnection, Local Area Network (LAN), Wide Area Network (WAN), Controller Area Network (CAN), wireless network (e.g., using Wireless Application Protocol), the Internet, and the like.

The ML model generating system 101 may communicate with a database (not shown in Figure 1). The database stores ML models generated by the ML model generating system 101. The database is communicatively coupled to the ML model generating system 101 using the communication network.

Hereinafter, the operation of the ML model generating system 101 is explained briefly.

Initially, a user 102 may wish to generate a ML model based on his/her requirement. Prior to the generating the ML model, the ML model generating system 101 receives a user request to generate the ML model. The user request may include, but is not limited to, a username and a password, and Active Directory (AD) credentials. In an embodiment, the AD credentials may be stored in an enterprise directory which acts as single repository for all the user authentication and authorization within an enterprise. Upon receiving the user request, the ML model generating system 101 performs the authentication of the user request. Based on authentication i.e., when the authentication is successful, the ML model generating system 101 provides access to an ML operation platform. In case of authentication failure, the ML model generating system 101 denies access to the user 102 to the ML operation platform. Thereafter, the ML model generating system 101 provides one or more functionalities on the ML operation platform for user selection upon receiving the user request and post successful authentication. In an embodiment, the ML operation platform is associated with the ML model generating system 101. The ML operation platform may be, but not limited to, any platform such as AWS® platform, Micro Azure® platform, Google® platform, enterprise datacentres, any public cloud provider and the like on which ML techniques can be implemented or operationalized. The one or more functionalities may include, but is not limited to, GIT repository application, an automatic ML application, and a Jupyter Notebook application. Further, the ML model generating system 101 initializes one of the one or more functionalities on the ML operation platform selected by the user 102. Thereafter, the ML model generating system 101 generates the ML model based on a Software Development Kit (SDK) and launch instance function parameters on the selected functionality. The SDK may be a collection of software development tools, which includes a compiler, debugger and/or a software framework. The launch instance function parameters may include, but are not limited to, at least one of user data, one or more libraries, one or more datasets, a process engine, one or more software applications, one or more configurations, Identity and Access Management (IAM) identity, and one or more tags. In an embodiment, the generated ML model is used to solve a variety of use cases. For instance, when the user 102 starts generating ML model based on the use cases, data is generated, and artefacts is tagged with necessary information for easy retrieval and further processing of the user data. The one or more libraries may be libraries which are available publicly for ML data scientists to develop ML models. Thereafter, in one embodiment, the ML model generating system 101 may store the generated ML model in the database. The process engine may refer to a software framework that enables the execution and maintenance of ML process workflows. The IAM identity refers to a set of processes, and tools for controlling user access. In an embodiment, the ML model generating system 101 determines at least one of run identifier (id), accuracy score and hyper setting parameter for the generated ML model. When the run identifier, the accuracy score and the hyper setting parameter are all above a pre-defined threshold limit, the ML model generating system 101 stores the generated ML model in the database. When at least one of the run identifier, the accuracy score and the hyper setting parameter is below a pre-defined threshold limit, the ML model generating system 101 trains the generated ML model using the launch instance function parameters. After training, when the run identifier, the accuracy score and the hyper setting parameter are all above or equal to the pre-defined threshold limit the ML model generating system 101 stores the generated ML model in the database. In one embodiment, there may be an individual pre-defined threshold limit or value for each of the run identifier, the accuracy score, and the hyper setting parameter. The pre-defined threshold limit may be set by a user or may be as per industry standard. The runtime identifier may be a unique, platform-level identifier for a workflow execution. The accuracy score may be a number that denotes how accurate the prediction of the ML model is. The hyper setting parameter may be a parameter that is set before the learning/training process of the ML model begins. The hyper setting parameter are tunable and can directly affect how well a ML model is trained.

Figure 2 shows a detailed block diagram of a ML model generating system for generating an ML model, in accordance with some embodiments of the present disclosure.

The model generating system 101, in addition to the I/O interface 104 and processor 103 described above, includes data 206 and one or more modules 200, which are described herein in detail. In the embodiment, the data 206 may be stored within the memory 105.

The data 206 in the memory 105 and one or more modules 200 of the model generating system 101 are described herein in detail.

In one embodiment, the data 206 in the memory 105 may include input data 207, functionality data 208, launch instance function parameters 209 and miscellaneous data 210 associated with the ML model generating system 101.

The input data 207 may include a user request comprising information such as username, password, and Active Directory (AD) credentials of the user 102.

The functionality data 208 may include information regarding the one or more functionalities. The one or more functionalities may include a GIT repository application, an automatic ML application, and a Jupyter Notebook application.

The launch instance function parameters 209 may include information regarding user data, one or more libraries, one or more dataset, a process engine, one or more software applications, one or more configurations, IAM identity, and one or more tags.

The miscellaneous data 210 may store data, including temporary data and temporary files, generated by one or more modules 200 for performing the various functions of the ML model generating system 101.

In one implementation, the one or more modules 200 may include, but are not limited to, a receiving module 201, a providing module 202, an initializing module 203, a generating module 204, and miscellaneous modules 205 associated with the ML model generating system 101.

In an embodiment, the data 206 in the memory 105 may be processed by the one or more modules 200 of the ML model generating system 101. The one or more modules 200 may be configured to perform the steps of the present disclosure using the data 206 for generating the ML model. In an embodiment, each of the one or more modules 200 may be a hardware unit which may be outside the memory 105 and coupled with the ML model generating system 101, as shown in Figure 2. In an embodiment, the one or more modules 200 may be implemented as dedicated units. When implemented in such a manner, said modules may be configured with the functionality defined in the present disclosure to result in a novel hardware. As used herein, the term module may refer to an Application Specific Integrated Circuit (ASIC), an electronic circuit, a Field-Programmable Gate Arrays (FPGA), Programmable System-on-Chip (PSoC), a combinational logic circuit, and/or other suitable components that provide the described functionality. In another embodiment, the one or more modules 200 may be inside the memory 105 and may be software units.

During the process of generating a ML model, the receiving module 201 is configured to receive a user request to generate the ML model. In an embodiment, the receiving module 201, upon receiving the user request, is configured to authenticate the user request and provide access to an ML operation platform based on the authentication. Further, the providing module 202 is configured to provide one or more functionalities on the ML operation platform associated with the ML model generating system 101 for user selection. Further, the initializing module 203 is configured to initialize one of the one or more functionalities on the ML operation platform that is selected by user 102. Upon initializing one of the one or more functionalities, the generating module 204 is configured to generate the ML model based on SDK and the launch instance function parameters on the selected functionality. Thereafter, the generating module 204 is configured to store the generated ML model in the database. In an embodiment, the generating module 204 is configured to determine at least one of run identifier, accuracy score and hyper setting parameter for the generated ML model. Further, the generating module 204 is configured to train the generated ML model using the launch instance function parameters when at least one of the run identifier, the accuracy score and the hyper setting parameter is below a pre-defined threshold limit. Thereafter, the generating module 204 is configured to store the generated ML model in a database when the run identifier, the accuracy score and the hyper setting parameter are all above or equal to the pre-defined threshold limit after training.

The one or more modules 200 may also include miscellaneous modules 205 to perform various miscellaneous functionalities of the ML model generating system 101. It will be appreciated that all the modules may be represented/configured as a single module or a combination of different modules.

Figure 3a illustrate an exemplary architecture of a ML model generating system for generating a ML model, in accordance with some embodiments of the present disclosure. In an embodiment, Figure 3a depicts a high-level logical architecture of a Machine Learning Operations (ML Ops) platform (also, referred as the ML operation platform) associated with the ML model generating system 101. According to Figure 3 a, the user 102 create use cases (for example, for chum analytics to evaluate a company’s customer loss rate for reducing it) and corresponding ML models in the ML Ops platform. The user 102 may include, data scientists, ML engineers and the like. The user 102 may have access to all the data points or information stored in a repository/database for creating the ML models. Further, the user 102 may conduct experiments i.e., construct ML models and run experiments i.e., test the constructed ML models in the ML Ops platform to identify accurate ML model. In one embodiment, the user 102 may conduct experiments i.e., construct ML models and run experiments i.e., test the constructed ML models in the ML Ops platform to identify accurate ML model based on the run identifier, the accuracy score, and the hyper setting parameter. The infrastructure or the ML Ops support tools such as Jupyter notebook® is an industrial standard tool for the user 102. The infrastructure or the ML Ops support tools may also be referred as one or more functionalities on the ML Ops platform for user selection. In one embodiment, the one or more functionalities comprise at least one of a GIT repository application, an automatic ML application, and a Jupyter Notebook application. In an embodiment, the ML Ops support libraries including, but not limited to, ‘Python’ libraries and ‘R’ libraries. The libraries may be stored in the repository/database and/or on cloud database (not shown in Figure 3a). In one embodiment, with auto ML libraries, the ML model generating system automatically picks and chooses best ML models based on input data set received from the user 102 and attribute that has to be predicted.

In an embodiment, the ML Ops comprise a code repository which runs on GIT, a distributed version control system, where the user 102 may connect to the GIT and may gain access to the code repository. Further, the ML model is binned/stored in the code repository. Upon binning the ML model, the run time instances are generated in real-time in the repository/database and/or on the cloud database, and subsequently, the ML models are made to run in ML Ops with the data points stored in the repository/database to get an output or result. Further, the outputs or the results are stored in the repository/database. Additionally, for securing access to the ML Ops platform, the user 102 may login to the authentication server of the organization for requesting access to the ML Ops platform.

Figure 3b shows a flow diagram illustrating a method generating an ML model, in accordance with some embodiments of the present disclosure. The following method is carried out by the ML model generating system of the present disclosure.

Initially, at step 1 of Figure 3b, if the user 102 need to access the ML operation platform (also, referred as ML Ops platform) associated with the ML model generating system 101, the user 102 first connects to a secure Virtual Private Network (VPN) using a username and a password. Further, if credentials of the user 102 are valid, then the VPN prompts the user 102 to enter a 6- digit code for authentication of the user 102. Upon successful authentication, a connection is established between the user 102 and the ML Ops platform. In an exemplary embodiment, the user 102 may use domains such as ‘https://mlops.tataplay.com/’ to login to the ML Ops platform.

At step 2, when the user 102 makes a request, the ML model generating system 101 receives the request through an Application Load Balancer (ALB). Further, a Secure Socket Layer (SSL) is attached to the ALB, and subsequently, the ALB is attached to an instance named as ‘ML Ops frontend’. Concurrently, a login page is displayed on the browser of the end users’ device. In an embodiment, the ML Ops platform follows a 3 -tiered architectural model, and deployment is done based on the following features:

• Frontend Technologies: VueJS, Vuetify, NodeJS, etc.

• Backend Technologies: Python 3.7, Django 3.2, DRF, etc.

• Database: MySQL Engine. At step 3, the user 102 may connect to the ML Ops application using their Active Directory (AD) credentials, and subsequently, the ML Ops application may connect to a Lightweight Directory Access Protocol (LDAP) for authentication and authorization. If the credentials are valid, the application may receive the user’s metadata as a response. The metadata may include details such as username, security group, email etc. Further, the user 102 may navigate to the dashboard. To summarize the steps 1, 2 and 3 mentioned above, the ML model generating system 101 receives a user request to generate the ML model. The user request may include, but is not limited to, a username and a password, and Active Directory (AD) credentials. Upon receiving the user request, the ML model generating system 101 performs the authentication of the user request. Based on authentication i.e., when the authentication is successful, the ML model generating system 101 provides access to an ML operation platform. In case of authentication failure, the ML model generating system 101 denies access to the user 102 to the ML operation platform.

At step 4, the ML model generating system 101 may provide one or more functionalities on the ML operation platform for performing the ML operations:

• GIT Repo (also, referred as GIT repository application)

• Auto ML (also, referred as automatic ML application)

• Jupyter Notebook (also, referred as Jupyter Notebook application)

In one embodiment, the ML Ops application in co-ordination with the ML model generating system 101 may provide the above-mentioned one or more functionalities on the ML operation platform. To summarize the step 4 mentioned above, the ML model generating system 101 provides one or more functionalities on the ML operation platform for user selection upon receiving the user request. The ML operation platform is associated with the ML model generating system. The one or more functionalities comprise one of a GIT repository application, an automatic ML application, and a Jupyter Notebook application.

At step 5, the ML Ops application may create a wrapper file with user data, and the created user data is passed as a payload to the Application Programming Interface (API) gateway trigger point to invoke a launch of Lambda function. Further, the configuration of the API gateway is carried out on the back-end server using temporary credentials such as Security Token Service (STS) assume role policy. To summarize the step 4 mentioned above, the ML model generating system 101 initializes one of the one or more functionalities on the ML operation platform selected by the user.

At step 6, the Lambda function may receive the user data as a parameter for the instance launch function and it may launch an on-demand instance, for example, with predefined Amazon® Machine Image (AMI), using Boto3 (AWS SDK) script. The instance details may include details like Internet Protocol (IP) address and instance ID etc., and they are recorded in the database. The following parameters are the arguments for the boto3 script.

• User data

• AMI - contains all required libraries, software and configurations (Refer step 23)

• Identity and Access Management (IAM) role

• Tags

At step 7, the ML Ops application may offer a version control system GitLab to the data scientists and/or ML engineers to maintain their code repositories. The on-demand instance may download the code from the GitLab to train the model.

At step 8, the dataset may be downloaded from the Bigdata production S3 bucket to the on-demand instance using the IAM role. The above said dataset may also be utilized to train the model.

At step 9, a process engine executes Python or R scripts and subsequently performs the ML operations like training, model fitting, calculating the accuracy score and making predictions on a test data etc.

At step 10, the dataset, from S3 bucket, which is a type of parquet and is used to train the model. For instance, the parquet is an efficient columnar data storage format that supports complex nested data structures in a flat columnar format.

At step 11, the result of the Python or R script, which is a pkl object file, is stored in a file with ‘.pkl’ or ‘.dump’ file extension. Further, during the execution of the script, logs are captured in a Tog. txt’ file.

At step 12, the pkl object file may be uploaded back to the ML Ops production S3 bucket for later use. Specifically, actual predictions are made using the pkl object file.

At step 13, the log.txt file, which may contain errors or exception information, if any, are sent to the back-end server using the Client Uniform Resource Locator (CURL) command.

At step 14, if the training model completes its training successfully, the script may send response back to the backend server using the CURL command. Upon receiving the response, the details are saved into a database.

At step 15, after the completion of ML model training, the backend server may invoke a ‘terminate’ Lambda function using the API gateway. As a result, the on-demand instance may also terminate automatically. To summarize the steps 6 to 15 mentioned above, the ML model generating system 101 generates a ML model based on Software Development Kit (SDK) and launch instance function parameters on the selected functionality.

At step 16, the results of the ML model training are collected. In an embodiment, the results of the model training are shown in the table with the necessary information along with a deployment button. The necessary information may include, without limiting to, a run identifier (ID), an accuracy score and hyper settings parameter, etc. Further, once a best-fit model is determined by accuracy score, the run ID, and hyper settings parameter, the user 102 may wish to have this ML model deployed on the server to perform batch predictions. For instance, if the user 102 clicks on the deployment button, the backend server automatically takes the user’s data depending on the runtime ID. Subsequently, the Lambda function is invoked using API gateway with the user data as a Java Script Object Notations (JOSN) pay load. Subsequently, the instance may be launched using a predefined AMI (similar to step 6). To summarize the step 16 mentioned above, the ML model generating system 101 determines at least one of run identifier, accuracy score and hyper setting parameter for the generated ML model. When at least one of the run identifier, the accuracy score and the hyper setting parameter is below a pre-defined threshold limit, the ML model generating system 101 trains the generated ML model using launch instance function parameters.

At step 17, both the pkl object file and the prediction data set file may be downloaded from the ML Ops production S3 bucket into the on-demand instance. At step 18, the user data may include scripts to make batch predictions, and the process engine may initiate the execution of the script by applying the parameters of the pkl object file and dataset. In an embodiment, the final outcome of the script may be the predictions.

At step 19, the prediction results are appended to prediction dataset by adding one extra column named as ‘predictions’ and this file is uploaded back to the ML Ops production S3 bucket.

At step 20, the prediction results are sent back to the back-end server, and subsequently, stored in the database and shown in the frontend. The application may also have an option to download the prediction results to the local machine.

At step 21, the terminate Lambda is invoked to terminate the server once the results are saved to database. If any exception occurs, the application displays the exception details to the user 102.

At step 22, the application may also provide a functionality to monitor the cost of each individual user and group. As an example, the organisation may share the cost explorer file on 6th of every month to the ML Ops production S3 bucket at the specified location using the bucket replication policy. Further, the application may execute cron jobs to read the cost explorer file and save the cost details in the database.

At step 23, the predefined AMI mentioned in step 6 may contain the required libraries and software. These libraries are installed by taking a temporary internet access to the AMI server by allowing requested URLs only. This is done by application team in the organization by raising a request with AWS team and network team.

Figure 4 illustrates a flowchart showing exemplary method for generating a ML model, in accordance with some embodiments of present disclosure.

As illustrated in Figure 4, the method 400 may include one or more blocks for executing processes in the ML model generating system 101. The method 400 may be described in the general context of computer executable instructions. Generally, computer executable instructions can include routines, programs, objects, components, data structures, procedures, modules, and functions, which perform functions or implement abstract data types. The order in which the method 400 are described may not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method. Additionally, individual blocks may be deleted from the methods without departing from the scope of the subject matter described herein. Furthermore, the method can be implemented in any suitable hardware, software, firmware, or combination thereof.

At step 401, the receiving module 201 of the ML model generating system 101 receivesa user request to generate an ML model. The user request comprises at least one of username and password, and Active Directory (AD) credentials.

At step 402, the providing module 202 of the ML model generating system 101 provides one or more functionalities on a ML operation platform for user selection upon receiving the user request. The ML operation platform is associated with the ML model generating system 101. The one or more functionalities comprise one of a GIT repository application, an automatic ML application, and a Jupyter Notebook application.

At step 403, the initialization module 203 of the ML model generating system 101 initializes one of the one or more functionalities on the ML operation platform selected by a user 102.

At step 404, the generating module 204 of the ML model generating system 101 generates a machine learning model based on Software Development Kit (SDK) and launch instance function parameters on the selected functionality. The launch instance function parameters comprise at least one of user data, one or more libraries, one or more dataset, a process engine, one or more software applications, one or more configurations, IAM identity, and one or more tags.

Computing System

Figure 5 illustrates a block diagram of an exemplary computer system 500 for implementing embodiments consistent with the present disclosure. In an embodiment, the computer system 500 is used to implement the ML model generating system 101. The ML operation platform associated with the ML model generating system may be implemented in the computer system 500 or on a cloud server (not shown in Figure 5). The computer system 500 may include a central processing unit (“CPU” or “processor”) 502. The processor 502 may include at least one data processor for processing a file for generating a ML model. The processor 502 may include specialized processing units such as, integrated system (bus) controllers, memory management control units, floating point units, graphics processing units, digital signal processing units, etc.

The processor 502 may be disposed in communication with one or more input/output (I/O) devices 509 and 510 via I/O interface 501. The I/O interface 501 may employ communication protocols/methods such as, without limitation, audio, analog, digital, monaural, Radio Corporation of America (RCA) connector, stereo, IEEE®-1394, serial bus, universal serial bus (USB), infrared, PS/2, BNC, coaxial, digital visual interface (DVI), high-definition multimedia interface (HDMI), Radio Frequency (RF) antennas, S-Video, video graphics array (VGA), IEEE® 8O2.n /b/g/n/x, Bluetooth, cellular (e.g., code-division multiple access (CDMA), high-speed packet access (HSPA+), global system for mobile communications (GSM), long-term evolution (LTE®), Worldwide interoperability for Microwave access (WiMax®), or the like), etc.

Using the I/O interface 501, the computer system 500 may communicate with one or more I/O devices 509 and 510. For example, the input devices 509 may be an antenna, keyboard, mouse, joystick, (infrared) remote control, camera, card reader, fax machine, dongle, biometric reader, microphone, touch screen, touchpad, trackball, stylus, scanner, storage device, transceiver, video device/source, etc. The output devices 510 may be a printer, fax machine, video display (e.g., cathode ray tube (CRT), liquid crystal display (LCD), light-emitting diode (LED), plasma, Plasma display panel (PDP), Organic light-emitting diode display (OLED) or the like), audio speaker, etc.

In some embodiments, the computer system 500 may consist of the ML model generating system 101. The processor 502 may be disposed in communication with the communication network 511 via a network interface 503. The network interface 503 may communicate with the communication network 511. The network interface 503 may employ connection protocols including, without limitation, direct connect, Ethernet (e.g., twisted pair 10/100/1000 Base T), transmission control protocol/internet protocol (TCP/IP), token ring, IEEE® 802.1 la/b/g/n/x, etc. The communication network 511 may include, without limitation, a direct interconnection, local area network (LAN), wide area network (WAN), wireless network (e.g., using Wireless Application Protocol), the Internet, etc. Using the network interface 503 and the communication network 511, the computer system 500 may communicate with user device/user 102 for generating a ML model and a database (not shown in Figure 5). The network interface 503 may employ connection protocols include, but not limited to, direct connect, Ethernet (e.g., twisted pair 10/100/1000 Base T), transmission control protocol/internet protocol (TCP/IP), token ring, IEEE® 802.1 l a/b/g/n/x, etc.

The communication network 511 includes, but is not limited to, a direct interconnection, an e- commerce network, a peer to peer (P2P) network, local area network (LAN), wide area network (WAN), wireless network (e.g., using Wireless Application Protocol), the Internet, Wi-Fi, and such. The communication network 511 may either be a dedicated network or a shared network, which represents an association of the different types of networks that use a variety of protocols, for example, Hypertext Transfer Protocol (HTTP), Transmission Control Protocol/internet Protocol (TCP/IP), Wireless Application Protocol (WAP), etc., to communicate with each other. Further, the communication network 511 may include a variety of network devices, including routers, bridges, servers, computing devices, storage devices, etc.

In some embodiments, the processor 502 may be disposed in communication with a memory 505 (e.g., RAM, ROM, etc. not shown in Figure 5) via a storage interface 504. The storage interface 504 may connect to memory 505 including, without limitation, memory drives, removable disc drives, etc., employing connection protocols such as, serial advanced technology attachment (SATA), Integrated Drive Electronics (IDE), IEEE- 1394, Universal Serial Bus (USB), fibre channel, Small Computer Systems Interface (SCSI), etc. The memory drives may further include a drum, magnetic disc drive, magneto-optical drive, optical drive, Redundant Array of Independent Discs (RAID), solid-state memory devices, solid-state drives, etc.

The memory 505 may store a collection of program or database components, including, without limitation, user interface 506, an operating system 507, web browser 508 etc. In some embodiments, computer system 500 may store user/application data, such as, the data, variables, records, etc., as described in this disclosure. Such databases may be implemented as fault-tolerant, relational, scalable, secure databases such as Oracle® or Sybase®.

The operating system 507 may facilitate resource management and operation of the computer system 500. Examples of operating systems include, without limitation, APPLE® MACINTOSH® OS X, UNIX®, UNIX-like system distributions (E.G., BERKELEY SOFTWARE DISTRIBUTION™ (BSD), FREEBSD™, NETBSD™, OPENBSD™, etc.), LINUX DISTRIBUTIONS™ (E.G., RED HAT™, UBUNTU™, KUBUNTU™, etc.), IBM™ OS/2, MICROSOFT™ WINDOWS™ (XP™, VISTA™/7/8, 10 etc.), APPLE® IOS™, GOOGLE® ANDROID™, BLACKBERRY® OS, or the like.

In some embodiments, the computer system 500 may implement a web browser 508 stored program component. The web browser 508 may be a hypertext viewing application, such as Microsoft® Internet Explorer®, Google™ Chrome™, Mozilla® Firefox®, Apple® Safari®, etc. Secure web browsing may be provided using Hypertext Transport Protocol Secure (HTTPS), Secure Sockets Layer (SSL), Transport Layer Security (TLS), etc. Web browsers 508 may utilize facilities such as AJAX, DHTML, Adobe® Flash®, JavaScript®, Java®, Application Programming Interfaces (APIs), etc. In some embodiments, the computer system 500 may implement a mail server stored program component. The mail server may be an Internet mail server such as Microsoft Exchange, or the like. The mail server may utilize facilities such as ASP, ActiveX®, ANSI® C++/C#, Microsoft® .NET, Common Gateway Interface (CGI) scripts, Java®, JavaScript®, PERL®, PHP, Python®, WebObjects®, etc. The mail server may utilize communication protocols such as Internet Message Access Protocol (IMAP), Messaging Application Programming Interface (MAPI), Microsoft Exchange, Post Office Protocol (POP), Simple Mail Transfer Protocol (SMTP), or the like. In some embodiments, the computer system 500 may implement a mail client stored program component. The mail client may be a mail viewing application, such as Apple® Mail, Microsoft® Entourage®, Microsoft® Outlook®, Mozilla® Thunderbird®, etc.

Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present disclosure. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term “computer-readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., be non-transitory. Examples include Random Access Memory (RAM), Read-Only Memory (ROM), volatile memory, non-volatile memory, hard drives, CD ROMs, DVDs, flash drives, disks, and any other known physical storage media.

Some of the advantages of the present disclosure are listed below:

The ML model generating system of the present disclosure helps the user to select a best-fit ML model by providing suggestion in terms of functionalities, Software Development Kit (SDK) and launch instance function parameters. As a result, the present disclosure helps in obtaining accurate results.

The implementation of the ML model generating system of present disclosure and associated ML operation platform on a cloud server optimizes infrastructure, enhances security, and increases connectivity/accessibility in terms of user outreach (number of user who can access the ML model generating system of the present disclosure).

The ML model generating system of the present disclosure is cloud agnostic. As a result, the ML model generating system can be run on any server or any public cloud such as AWS®, Micro Azure®, Google® cloud and the like.

The ML model generating system of the present disclosure can be used by users ranging from beginners to proficient data scientists, making the ML model generating system highly scalable and sophisticated system.

The ML model generating system of the present disclosure can be implemented on any ML operation platform allowing any users, such as data scientists or statisticians or ML engineers, to write a ML code, run/text the ML code, and deploy the ML code in a secure manner.

The described operations may be implemented as a method, system or article of manufacture using standard programming and/or engineering techniques to produce software, firmware, hardware, or any combination thereof. The described operations may be implemented as code maintained in a “non-transitory computer readable medium”, where a processor may read and execute the code from the computer readable medium. The processor is at least one of a microprocessor and a processor capable of processing and executing the queries. A non-transitory computer readable medium may include media such as magnetic storage medium (e.g., hard disk drives, floppy disks, tape, etc.), optical storage (CD-ROMs, DVDs, optical disks, etc.), volatile and non-volatile memory devices (e.g., EEPROMs, ROMs, PROMs, RAMs, DRAMs, SRAMs, Flash Memory, firmware, programmable logic, etc.), etc. Further, non-transitory computer-readable media may include all computer-readable media except for a transitory. The code implementing the described operations may further be implemented in hardware logic (e.g., an integrated circuit chip, Programmable Gate Array (PGA), Application Specific Integrated Circuit (ASIC), etc.).

An “article of manufacture” includes non-transitory computer readable medium, and /or hardware logic, in which code may be implemented. A device in which the code implementing the described embodiments of operations is encoded may include a computer readable medium or hardware logic. Of course, those skilled in the art will recognize that many modifications may be made to this configuration without departing from the scope of the invention, and that the article of manufacture may include suitable information bearing medium known in the art.

The terms “an embodiment”, “embodiment”, “embodiments”, “the embodiment”, “the embodiments”, “one or more embodiments”, “some embodiments”, and “one embodiment” mean “one or more (but not all) embodiments of the invention(s)” unless expressly specified otherwise.

The terms “including”, “comprising”, “having” and variations thereof mean “including but not limited to”, unless expressly specified otherwise.

The enumerated listing of items does not imply that any or all of the items are mutually exclusive, unless expressly specified otherwise.

The terms “a”, “an” and “the” mean “one or more”, unless expressly specified otherwise.

A description of an embodiment with several components in communication with each other does not imply that all such components are required. On the contrary a variety of optional components are described to illustrate the wide variety of possible embodiments of the invention.

When a single device or article is described herein, it will be readily apparent that more than one device/article (whether or not they cooperate) may be used in place of a single device/article. Similarly, where more than one device or article is described herein (whether or not they cooperate), it will be readily apparent that a single device/article may be used in place of the more than one device or article, or a different number of devices/articles may be used instead of the shown number of devices or programs. The functionality and/or the features of a device may be alternatively embodied by one or more other devices which are not explicitly described as having such functionality /features. Thus, other embodiments of the invention need not include the device itself.

The illustrated operations of Figure 4 shows certain events occurring in a certain order. In alternative embodiments, certain operations may be performed in a different order, modified, or removed. Moreover, steps may be added to the above-described logic and still conform to the described embodiments. Further, operations described herein may occur sequentially or certain operations may be processed in parallel. Yet further, operations may be performed by a single processing unit or by distributed processing units.

Finally, the language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter. It is therefore intended that the scope of the invention be limited not by this detailed description, but rather by any claims that issue on an application based here on. Accordingly, the disclosure of the embodiments of the invention is intended to be illustrative, but not limiting, of the scope of the invention, which is set forth in the following claims.

While various aspects and embodiments have been disclosed herein, other aspects and embodiments will be apparent to those skilled in the art. The various aspects and embodiments disclosed herein are for purposes of illustration and are not intended to be limiting, with the true scope being indicated by the following claims.

Referral numerals:

Claims

We claim:

1. A method for generating a Machine Learning (ML) model (101), the method comprising: receiving (401), by a ML model generating system (101), a user request to generate an ML model; providing (402), by the ML model generating system (101), one or more functionalities on a ML operation platform for user selection upon receiving the user request, wherein the ML operation platform is associated with the ML model generating system (101); initializing (403), by the ML model generating system (101), one of the one or more functionalities selected by a user; and generating (404), by the ML model generating system (101), a ML model based on Software Development Kit (SDK) and launch instance function parameters on the selected functionality.

2. The method as claimed in claim 1, wherein the user request comprises at least one of username and password, and Active Directory (AD) credentials.

3. The method as claimed in claim 1, wherein providing the ML operation platform upon receiving the user request comprises: authenticating, by the ML model generating system (101), the user request; and providing, by the ML model generating system (101), access to the ML operation platform based on the authentication.

4. The method as claimed in claim 1, wherein the one or more functionalities comprise at least one of a GIT repository application, an automatic ML application, and a Jupyter Notebook application.

5. The method as claimed in claim 1, wherein the launch instance function parameters comprise at least one of user data, one or more libraries, one or more dataset, a process engine, one or more software applications, one or more configurations, IAM identity, and one or more tags.

6. The method as claimed in claim 1, further comprising: determining, by the ML model generating system (101), at least one of run identifier, accuracy score and hyper setting parameter for the generated ML model; and training, by the ML model generating system (101), the generated ML model using launch instance function parameters when at least one of the run identifier, the accuracy score and the hyper setting parameter is below a pre-defined threshold limit.

7. The method as claimed in claim 6, further comprising: storing, by the ML model generating system (101), the generated ML model in a database when the run identifier, the accuracy score and the hyper setting parameter are above or equal to the pre-defined threshold limit.

8. A Machine Learning (ML) model generating system (101) for generating a ML model, comprising: a processor (103); and a memory (105) communicatively coupled to the processor (103), wherein the processor (103) is configured to: receive a user request to generate an ML model; provide one or more functionalities on a ML operation platform for user selection upon receiving the user request, wherein the ML operation platform is associated with the ML model generating system; initialize one of the one or more functionalities selected by the user; and generate a ML model based on Software Development Kit (SDK) and launch instance function parameters on the selected functionality.

9. The ML model generating system (101) as claimed in claim 8, wherein the user request comprises at least one of username and password, and Active Directory (AD) credentials.

10. The ML model generating system (101) as claimed in claim 8, wherein the processor (103) provides the ML operation platform upon receiving the user request by: authenticating the user request; and providing access to the ML operation platform based on the authentication.

11. The ML model generating system (101) as claimed in claim 8, wherein the one or more functionalities comprise at least one of a GIT repository application, an automatic ML application, and a Jupyter Notebook application.

12. The ML model generating system ( 101) as claimed in claim 8, wherein the launch instance function parameters comprise at least one of user data, one or more libraries, one or more dataset, a process engine, one or more software applications, one or more configurations, IAM identity, and one or more tags.

13. The ML model generating system (101) as claimed in claim 8, wherein the processor (103) is configured to: determine at least one of run identifier, accuracy score and hyper setting parameter for the generated ML model; and train the generated ML model using launch instance function parameters when at least one of the run identifier, the accuracy score and the hyper setting parameter is below a pre-defined threshold limit.

14. The ML model generating system (101) as claimed in claim 13, wherein the processor (103) is configured to: store the generated ML model in a database when the run identifier, the accuracy score and the hyper setting parameter are above or equal to the pre-defined threshold limit.