Nov 12, 2020

Building Machine Learning Pipelines without Coding

In the era of Industry 4.0 where data and predictive analytics players the major role, developing machine learning pipelines have become an essential in intelligent application development.

Azure Machine Learning Designer start page

In the era of Industry 4.0 where data and predictive analytics players the major role, developing machine learning pipelines have become an essential in intelligent application development.

The need of having specialized talent with extensive knowledge in programming is one of the major block factors for organizations in the process of building and testing such machine learning models.

Azure Machine Learning Studio(preview) offers a zero-code solution for developing machine learning pipelines. This is the evolution of legacy Azure ML Studio (Old drag and drop interface).

The prebuilt modules can be used for building the ML pipeline

The idea behind Azure ML Designer is quite straight forward. It allows the users to create a complete machine learning pipeline by interconnecting pre-built modules which has been designed according to the most used theories and methodologies of machine learning.

Aligning to the Cross-Industry Process for Data Mining (CRISP-DM), Designer allows to work on datasets from initial data wrangling steps of the process. You can use a dataset which is already there on Azure or you can import your own dataset using an Azure ML datastore for the pipeline.

Once the data has been imported, the data cleansing and transformation steps can be performed onto the data just by connecting the appropriate modules selected from the pre-built modules.

Right now, Azure ML Designer offers a range of machine learning algorithms that can be used to perform Regression, Clustering and Classification tasks. Model training and evaluation can be done with hyperparamter tuning with your desired method of parameter tuning as well as with scoring visualizations.

User can configure the compute resource to be used for running the experiments. It is advisable to use a high-end resource with larger computation power for processing complex machine learning scenarios.

One of the major advantages of Azure ML designer is the compatibility with R and Python scripts. If the user wants to add any custom script into the pipeline, it can be easily performed through the Python and R language modules.

A sample pipeline created using Azure ML Designer

When the pipeline is ready for the deployment, Designer allows you to deploy is as a web service endpoint which is running on an Azure Kubernetes Service (AKS). The configurations of the deployment target can be managed through Azure portal.

Azure ML Designer is available on the Enterprise version of Azure Machine Learning Studio (Preview). Since no-code development paradigms are getting popular, Designer would be a better solution for prototype development as well as for easy adaptation of intelligent application components for your system. Will discuss on developing a simple machine learning pipieline using Azure ML Designer in the next blog post.  

Interested in hearing more?

Lets connect.

Thoughts, stories and ideas.

June 11, 2021
Using Bicep and Azure DevOps multi stage pipeline to deploy a Function App

Treating your infrastructure as code is becoming more and more necessary these days. Writing these instructions becoming challenging too. In Azure we use ARM templates to define the resources and associate them with a deployment pipeline. But ARM templates are quite complicated and they are not everybody’s cup of tea.

learn more
March 8, 2021
Guide to choosing an Compute option in Azure Machine Learning service

Azure MachineLearning Service provides four main compute options each with a specific purpose attached to it. In this post we will go through each of those and see where we can occupy them in your ML experiments.

learn more
July 6, 2020
Building Machine Learning Pipelines without Coding

In the era of Industry 4.0 where data and predictive analytics players the major role, developing machine learning pipelines have become an essential in intelligent application development.

learn more