Amazon Web Services Inc. expanded its artificial intelligence portfolio at AWS re:Invent today by launching SageMaker Canvas, a tool that enables business users to create machine learning models without writing any code.

Usually, building a machine learning model requires not only coding know-how but also familiarity with AI-specific development tools such as TensorFlow. The need for specialized skills can make enterprise AI projects challenging in several ways. 

If a company doesn’t have any in-house AI expertise, it may have to hire specialists to support its machine learning projects. Meanwhile, firms that already have the necessary technical know-how can encounter challenges as well. Business users often require assistance from developers to launch machine learning projects, which can limit the pace at which companies can roll out AI in their operations. 

“As a business user or data analyst, you’d like to build and use prediction systems based on the data that you analyze and process every day, without having to learn about hundreds of algorithms, training parameters, evaluation metrics, and deployment best practices,” AWS developer advocate Alex Casalboni wrote in a blog post today. 

With SageMaker Canvas, AWS promises to ease AI development by removing the need for users to write any code. The cloud giant says that the tool doesn’t require extensive knowledge of machine learning technologies either.

To build an AI model, a SageMaker Canvas user must first provide a training dataset. Workers can upload the training dataset as a spreadsheet or import information from their company’s internal systems. SageMaker Canvas can draw on records stored in Amazon S3, other cloud sources such as the Amazon Redshift data warehouse or on-premises systems.

Users that import multiple training datasets can optionally integrate them into a single file for their AI projects. SageMaker Canvas automates key data preparation tasks. The tool helps identify issues such as missing spreadsheet fields and streamlines the manual work involved in combining information from different files. 

Once the training dataset is ready, workers can start building their AI model. Before spinning up a neural network, SageMaker Canvas provides an estimate of how accurately the neural network will produce results. Users can review the estimate and, if there’s room to improve accuracy, they may adjust their datasets as needed. 

SageMaker Canvas evaluates hundreds of AI models and picks the one that would prove most effective at the processing task the user is looking to automate. Workers can then have the neural network trained on their datasets with a few clicks. 

“SageMaker Canvas leverages the same technology as Amazon SageMaker to automatically clean and combine your data, create hundreds of models under the hood, select the best performing one, and generate new individual or batch predictions,” Casalboni wrote.

The SageMaker machine learning platform of which SageMaker Canvas is part also includes a variety of other tools. Those tools provide features for tasks ranging from creating AI training datasets to deploying neural networks in production. 

Another element of SageMaker’s value proposition is that it can reduce hardware costs and related maintenance overhead. SageMaker automatically scales the infrastructure customers use for AI projects. It also comes with built-in cybersecurity and compliance features that enterprises would otherwise have to implement from scratch. According to an analysis that AWS published last year, the platform has a lower total cost of ownership than manually setting up and maintaining an AI development environment. 

The newly announced SageMaker Canvas tool is available in general availability today. 

Image: AWS