Automated Machine Learning: Between Chat-GPT and die-hard coding

With the recent hype of Chat-GPT, Artificial Intelligence (AI) and Machine Learning (ML) is getting a lot of media attention. Many people see how useful a generic, pre-trained AI model can be: Although Chat-GPT is not trained specifically on your corporate data, it can provide a bunch of useful services to you and your employees.

But of course, Chat-GPT will not be able to predict how many items of product X your company will sell next week. Many down-to-earth decisions in companies can be optimized by machine learning models. However, companies are afraid that a long running project with expensive ML experts is needed to build such models, which stops them from taking the ML path. Although, in reality, many of the mainstream machine learning models can be created without the need to embark on an expensive and long running project.

In Microsoft Azure, there are two code-free paths to building customized ML models: Cognitive Services ( https://azure.microsoft.com/en-us/products/cognitive-services/, a collections of services, some of which use transfer learning to take a centrally trained model and provide easy, graphical tools to customize the model for your data) and Automated ML.

In Automated ML (https://learn.microsoft.com/en-us/azure/machine-learning/concept-automated-ml) Microsoft automates the process of building a machine learning model. Using cloud compute resources, the service tries out a wide variety of data preprocessing and machine learning techniques. After a few minutes to a few hours (depending on your patience and budget) it returns a report of all its efforts. With a click on a button you can turn the best-performing model in a webservice, such that it can be consumed in many other applications.

Interested to learn the basics about Automated Machine Learning? This weekend the Global AI bootcamp is held. You attend a free one-hour session on Automated ML by Nico Jacobs, one of our U2U trainers. Register at https://www.meetup.com/global-ai-channel/events/290917551/

If you want to have more in-depth knowledge and hands-on experience, you can participate in our two day Citizen Data Science class: https://www.u2u.be/training/citizen-data-science

Or for a full list of our AI oriented training offerings, check https://www.u2u.be/courses/ai

The future is AI, make sure you're part of it!