Press "Enter" to skip to content

Category: Machine Learning

AutoML with pycaret

Published 2022-03-01 by Kevin Feasel

Brendan Tierney looks at the pycaret library:

In this post we will have a look at using the AutoML feature in the Pycaret Python library. AutoML is a popular topic and allows Data Scientists and Machine Learning people to develop potentially optimized models based on their data. All requiring the minimum of input from the Data Scientist. As with all AutoML solutions, care is needed on the eventual use of these models. With various ML and AI Legal requirements around the World, it might not be possible to use the output from AutoML in production. But instead, gives the Data Scientists guidance on creating an optimized model, which can then be deployed in production. This facilitates requirements around model explainability, transparency, human oversight, fairness, risk mitigation and human in the loop.

Read on for a tutorial as well as additional resources.

Comments closed

Form Recognizer Updates

Published 2022-02-25 by Kevin Feasel

Vinod Kurpad shares some news:

Form Recognizer continues to improve product capabilities with improved models, support for additional document types and containerized solutions that run in the cloud or on premises either connected or fully disconnected for scenarios where containers need to run in an isolated environment. Recent updates to pricing include commitment tiers for customers who have a predictable volume of documents. Starting February 15th, the pricing for Invoices and General Document API will drop to $10 per 1000 pages, an 80% reduction, making it possible for customers to use invoices and the general document APIs for high volume scenarios to significantly lower cost while providing additional value.

That’s a pretty big improvement.

Comments closed

Multivariate Anomaly Detection in SynapseML

Published 2022-02-17 by Kevin Feasel

Louise Han has an announcement:

Today, we are excited to announce a wonderful collaborated feature between Multivariate Anomaly Detector and  SynapseML , which joined together to provide a solution for developers and customers to do multivariate anomaly detection in Synapse. This new capability allows you to detect anomalies quickly and easily in very large datasets and databases, perfectly lighting up scenarios like equipment predictive maintenance. For those who is not familiar with predictive maintenance, it is a technique that uses data analysis tools and techniques to detect anomalies in the operation and possible defects in equipment and processes so customers can fix them before they result in failure. Therefore, this new capability will benefit customers who have a huge number of sensor data within hundreds of pieces of equipment, to do equipment monitor, anomaly detection, and even root cause analysis.

Click through for more details and a demonstration on how to use it.

Comments closed

The Architecture of Project Bansai

Published 2022-02-17 by Kevin Feasel

Tsuyoshi Matsuzaki takes us through the architecture for Project Bansai:

Project Bonsai is a reinforcement learning framework for machine teaching in Microsoft Azure.

In generic reinforcement learning (RL), data scientists will combine tools and utilities (such like, Gym, RLlib, Ray, etc) which can be easily customized with familiar Python code and ML/AI frameworks, such as, TensorFlow or PyTorch.
But, in engineering tasks with machine teaching for autonomous systems or intelligent controls, data scientists will not always explore and tune attributes for AI. In successful practices, the professionals for operations or engineering (non-AI specialists) will tune attributes for some specific control systems (simulations) to train in machine teaching, and data scientists will assist in cases where the problem requires advanced solutions.

Read on to see how it works.

Comments closed

Azure ML and MLOps

Published 2022-02-15 by Kevin Feasel

I continue a series on Azure ML:

We ended the prior series with model deployment via the Azure ML Studio UI. This is entirely manual and UI-driven. Then, we looked at model deployment via manually-run notebooks. This is still manual but at least offers the possibility of automation as we control the code to run.

From there, we moved to model deployment via the Azure CLI and Python SDK. Now we have the capability to run, train, register, and deploy models via scripts. This leads to the next phase in the process, in which we can perform continuous integration and continuous deployment of models using a tool like Azure DevOps or GitHub Actions. This is where MLOps starts to shine.

Read on for a few thoughts about MLOps and software maturity.

Comments closed

Azure ML and the Python SDK in VS Code

Published 2022-02-08 by Kevin Feasel

I continue a series on getting beyond the basics with Azure ML. First up, we get up close and personal in development:

Notebooks are great for ad hoc work or simple data analysis but we will want more robust tools if we wish to perform proper code development, testing, and deployment. This is where Visual Studio Code comes into play, particularly the Azure Machine Learning extension.

Then, I get into the Python SDK:

Over the past two posts, we have started using the Azure Machine Learning SDK for Python but I’ve only touched on the topic. In this post, we are going to dive into the topic.

Read on for more info on each.

Comments closed

Sentiment Analysis with Python

Published 2022-02-04 by Kevin Feasel

Sanil Mhatre performs a bit of sentiment analysis:

Previous articles in this series have focused on platforms like Azure Cognitive Services and Oracle Text features to perform the core tasks of Natural Language Processing (NLP) and Sentiment Analysis. These easy-to-use platforms allow users to quickly analyze their text data with easy-to-use pre-built models. Potential drawbacks of this approach include lack of flexibility to customize models, data locality & security concerns, subscription fees, and service availability of the cloud platform. Programming languages like Python and R are convenient when you need to build your models for Natural Language Processing and keep your code as well as data contained within your data centers. This article explains how to do sentiment analysis using Python.

Python is a versatile and modern general-purpose programming language that is powerful, fast, and easy to learn. Python runs on interpreters, making it compatible with multiple platforms, and is widely used in applications for web platforms, graphical interfaces, data science, and machine learning. Python is increasingly gaining popularity in data analysis and is one of the most widely used languages for data science. You can learn more about Python from the official Python Software Foundation website.

Click through to see what’s available in the NLP world for Python. The short version is “a lot.”

Comments closed

Run Spark within Azure ML Compute

Published 2022-02-04 by Kevin Feasel

James Nguyen makes an announcement:

Following the blog post on Turning AML compute into Ray and Dask , we’ve added a new exciting capability to run Spark within AML compute where Spark shares the same context with your ML code. The Spark version is 3.2.1 with support for Delta Lake and Synapse SQL read/write. This enables users of AML to perform powerful data transformation and even Spark ML within AML interactive notebook or in a job run. 

Traditionally, Azure ML integrates with Spark Synapse or external compute services via a pipeline step or better via magic command like %synapse, but the computing context is separate from your AML logic so you still need to run Spark in a separate step and persist the output to some storage and load it in your AML script.

With this approach, Spark is available right within your AML code whether it’s AML notebook, python script or pipeline step. It shares the common computing context and most of the cases you can just directly convert the Spark Dataframe to Pandas and Dask Dataframe without persisting first to an intermediary storage.

I’ll have to try this out to see if it makes up for their getting rid of the Spark-based curated environments last year.

Comments closed