25 March 2020
The 4th annual Artificial Intelligence Dev Conference at DeveloperWeek was a full day of thought leadership talks and technical workshops on the newest technologies and best practices in AI development. Cloud AI services and automated machine learning (AutoML) were the subjects of one of the talks, and in this article we’ll explore why they’re so popular and how they can help your organisation.
Cloud computing has been around for years and most of the key providers now offer cloud-based AI services. These leverage the providers’ extensive expertise and deep pockets, and make it easier for business users (from SMEs to large budget-sensitive corporations) to get started in AI. It makes it less expensive for them in terms of the heavy lifting processor technology that AI demands, and also means they don’t have to buy in costly AI and ML expertise (or not as much anyway).
IBM Cloud, Amazon Web Services, Google Cloud and Microsoft Azure all provide AI as a Service, and are making technological advances all the time. Here are some of the AI shortcuts they can currently give you:
That brings us on to AutoML. Machine learning is at the forefront of today's AI activities but it can require in-house expertise that is not always available. AutoML can help with this. It allows non-experts to make use of ML models without having to learn complex algorithms and advanced data science techniques.
Normally with a ML application, you have a dataset of input data points to train on. Expertise is then needed to perform the data pre-processing, feature engineering, feature extraction and feature selection to make the dataset suitable for machine learning. Once you’ve done that, you then need to perform algorithm selection and hyperparameter optimisation to maximise predictive performance. AutoML dramatically simplifies this.
It is offered to various degrees by the main Cloud AI service providers, as well as other companies. The Azure Machine Learning service, for example, lets you create models with the automated machine learning no-code UI or through a code-first approach. It can then intelligently select from a wide array of algorithms and hyperparameters to help you build highly accurate models. Google Cloud AutoML also enables you to train high quality models, even if you have limited ML expertise.
Cloud AI Services and AutoML can get you up and running fast. They are a growing trend.
JBI Training provides a range of industry-leading AI/ML and Cloud training courses including: