AWS expands on SageMaker capabilities with end-to-end features for machine learning

Almost three years after it was first released in 2020, the SageMaker platform of Amazon Web Services has received a major update in the form of new functionality, making it easier for developers to automate and scale each step of the process to create new capabilities for automation and machine learning, the company said.
Business units around industries can find applications for automation as machine learning progresses into the mainstream, and AWS is trying to make it easier for its customers to create such personalised applications.
AWS vice president of machine learning, Swami Sivasubramanian, said that one of the best aspects of providing such a widely used service as SageMaker is that  get tonnes of customer feedback that fuel their next collection of deliverables. On 09 Dec 2020, the company announced a collection of Amazon SageMaker tools that make it much easier for developers to create end-to-end machine learning pipelines to plan, build, train, illustrate, inspect, track, debug and run custom machine learning models with greater visibility, clarity and scale automation.
According to AWS, companies such as 3M, ADP, AstraZeneca, Avis, Bayer, Capital One, Cerner, Domino's Pizza, Fidelity Investments, Lenovo, Lyft, T-Mobile and Thomson Reuters are now using SageMaker instruments in their own operations.
Amazon SageMaker Data Wrangler includes the company's latest products, which the company said offered a way to normalise data from various sources so that the data is consistently easy to use. To highlight those types of data, Data Wrangler can also ease the process of grouping disparate data sources into features. There are more than 300 built-in data transformers in the Data Wrangler tool that can help clients normalise, transform and combine features without having to write any code.
The Feature Store was also introduced by Amazon, which enables consumers to build repositories that make it easier to store, update, retrieve and share training and inference features of machine learning.
Pipelines, the process management and automation toolkit, was another new tool that Amazon Web Services touted. The Pipelines technology is designed to include features of orchestration and automation that are not unlike conventional programming. Developers can describe each stage of an end-to-end machine learning workflow using pipelines, the firm said in a statement. Developers can use SageMaker Studio's software to re-run an end-to-end workflow using the same settings to get the same model every time, or they can re-run the workflow to upgrade their models with new data.
Amazon introduced SageMaker Explain to resolve the longstanding problems with data bias in artificial intelligence and machine learning models. This method, first announced today, allegedly provides bias detection through the workflow of machine learning, so developers can draw on how models were set up with an eye towards better transparency. According to Amazon, there are open-source tools that can do these tests, but the instruments are manual and according to the company, require a lot of lifting from developers.
Another product aimed at simplifying the development process for machine learning applications is the SageMaker Debugger. This allows developers to train their models more quickly by monitoring system resource usage and alerting developers about potential bottlenecks. Distributed training allows you to train large, complex deep learning models faster than current methods by automatically splitting your data across multiple GPUs to reduce training time. SageMakerEdgeManager is a type of machine learning for edge devices. Model management tool. Developers can optimize, protect, monitor, and manage models deployed across edge device clusters.
Last but not least, Amazon has announced the launch of SageMaker Jump Start. It provides developers with a searchable interface to find algorithms and example notebooks so they can start their machine learning journey. The company said it will provide developers unfamiliar with machine learning with the ability to select multiple off-the-shelf machine learning solutions and deploy them to the SageMaker environment.