AWS wellarchitected documentation change
Summary
Updated publication date, added distinction from Generative AI Lens, expanded ML use cases, and changed lens access method from AWS catalog to GitHub repository
Security assessment
The changes are editorial updates and content expansions without security-specific content. The mention of responsible AI implementation is a rewording of existing guidance, not new security documentation. No vulnerabilities, security patches, or security incidents are addressed.
Diff
diff --git a/wellarchitected/latest/machine-learning-lens/machine-learning-lens.md b/wellarchitected/latest/machine-learning-lens/machine-learning-lens.md index aeb761ae6..cbc360432 100644 --- a//wellarchitected/latest/machine-learning-lens/machine-learning-lens.md +++ b//wellarchitected/latest/machine-learning-lens/machine-learning-lens.md @@ -5 +5 @@ -IntroductionLens availability +IntroductionDistinction from the Generative AI LensLens availability @@ -7 +7 @@ IntroductionLens availability -# Machine Learning Lens +# Machine Learning Lens - AWS Well-Architected Framework @@ -9 +9 @@ IntroductionLens availability -Publication date: **July 5th, 2023** ([Document history](./document-history.html)) +Publication date: **November 19, 2025** ([Document revisions](./document-revisions.html)) @@ -11 +11 @@ Publication date: **July 5th, 2023** ([Document history](./document-history.html -In recent years, machine learning (ML) has moved from research and development to the mainstream, driven by the increasing number of data sources and scalable cloud-based compute resources. AWS’ customers currently use AI/ML for a wide variety of applications such as call center operations, personalized recommendations, identifying fraudulent activities, social media content moderation, audio and video content analysis, product design services, and identity verification. Industries using AI/ML include healthcare and life sciences, industrial and manufacturing, financial services, media and entertainment, and telecom. +Machine learning (ML) has evolved from research and development to the mainstream, driven by the exponential growth of data sources, generative AI and scalable cloud-based compute resources. AWS customers use AI/ML for a wide variety of applications, ranging from foundation model development and fine-tuning to sophisticated computer vision implementations. Common use cases include call center operations, personalized recommendations, fraud detection, social media content moderation, audio and video content analysis, product design services, and identity verification. These applications use both custom-built models and pre-trained solutions to address specific business needs. AI/ML adoption has become common across nearly every industry, including healthcare and life sciences, automotive, industrial and manufacturing, financial services, media and entertainment, and telecommunications. @@ -13 +13 @@ In recent years, machine learning (ML) has moved from research and development t -Machine learning, through its use of algorithms to find patterns in data, can bring considerable power to its customers and thus recommends responsibility in its use. AWS is committed to developing fair and accurate AI and ML services and providing you with the tools and guidance needed to build AI and ML applications responsibly. For more information on this important topic, refer to [AWS' Responsible AI](https://aws.amazon.com/machine-learning/responsible-machine-learning/). +Machine learning harnesses algorithms to discover patterns in data, delivering considerable value to customers while requiring responsible implementation. AWS is committed to developing fair and accurate AI and ML services and providing you with the tools and guidance needed to [build AI and ML applications responsibly](https://aws.amazon.com/machine-learning/responsible-machine-learning/). @@ -15 +15 @@ Machine learning, through its use of algorithms to find patterns in data, can br -This whitepaper provides you with a set of proven best practices. You can apply this guidance and architectural principles when designing your ML workloads, and after your workloads have entered production as part of continuous improvement. Although the guidance is cloud- and technology-agnostic, the paper also includes guidance and resources to help you implement these best practices on AWS. +This whitepaper provides you with a set of best practices. You can apply this guidance and architectural principles when designing your ML workloads and after your workloads have entered production as part of continuous improvement. Although the guidance is cloud- and technology-agnostic, the paper also includes guidance and resources to assist you to implement these best practices on AWS. @@ -19 +19 @@ This whitepaper provides you with a set of proven best practices. You can apply -The [AWS Well-Architected Framework](https://aws.amazon.com/architecture/well-architected/) helps you understand the benefits and risks of decisions you make while building workloads on AWS. By using the Framework, you learn operational and architectural best practices for designing and operating workloads in the cloud. It provides a way to consistently measure your operations and architectures against best practices and identify areas for improvement. +The AWS Well-Architected Framework assists you to understand the benefits and risks of decisions made while building workloads on AWS. Through the Framework, you learn operational and architectural best practices for designing and operating cloud workloads. It provides a consistent method to measure your operations and architectures against best practices and identify improvement opportunities. @@ -21 +21 @@ The [AWS Well-Architected Framework](https://aws.amazon.com/architecture/well-ar -Your ML models depend on the quality of input data to generate accurate results. As data changes with time, monitoring is required to continually detect, correct, and mitigate issues with accuracy and performance. This monitoring step might require you to retrain your model over time using the latest refined data. +Your ML models depend on high-quality input data to generate accurate results. As data evolves over time, continuous monitoring is essential to detect, correct, and mitigate accuracy and performance issues, often requiring model retraining with refined datasets. @@ -23 +23 @@ Your ML models depend on the quality of input data to generate accurate results. -While application workloads rely on step-by-step instructions to solve a problem, ML workloads enable algorithms to learn from data through an iterative and continuous cycle. The ML Lens complements and builds upon the Well-Architected Framework to address this difference between these two types of workloads. +While application workloads rely on deterministic, step-by-step instructions to solve problems, ML workloads enable algorithms to learn from data through iterative and continuous cycles. The ML Lens complements and builds upon the Well-Architected Framework to address the fundamental differences between traditional application workloads and machine learning workloads. @@ -26,0 +27,6 @@ This paper is intended for those in a technology role, such as chief technology +## Distinction from the Generative AI Lens + +The Machine Learning Lens addresses the broad spectrum of ML workloads, including traditional supervised and unsupervised learning, predictive analytics, classification, regression, and clustering tasks. Common ML use cases covered by this lens include computer vision for object detection and image classification, fraud detection and risk scoring, recommendation engines, predictive maintenance, demand forecasting, customer churn prediction, anomaly detection, and medical diagnosis systems. In contrast, the Generative AI Lens focuses on foundation models and generative AI applications that create content, such as text generation, image synthesis, and conversational AI systems. + +While both lenses share common ML principles, the Generative AI Lens emphasizes prompt engineering, foundation model selection, retrieval-augmented generation (RAG) architectures, and the specific governance challenges of generative AI systems. The Machine Learning Lens provides comprehensive guidance for the full ML lifecycle across ML paradigms, making it the foundational lens for ML workloads. + @@ -29 +35 @@ This paper is intended for those in a technology role, such as chief technology -The Machine Learning Lens is available as an AWS-official lens in the [Lens Catalog](https://docs.aws.amazon.com/wellarchitected/latest/userguide/lens-catalog.html) of the [AWS Well-Architected Tool](https://docs.aws.amazon.com/wellarchitected/latest/userguide/intro.html). +Custom lenses extend the best practice guidance provided by AWS Well-Architected Tool. AWS WA Tool allows you to create your own [custom lenses](https://docs.aws.amazon.com/wellarchitected/latest/userguide/lenses-custom.html), or to use lenses created by others that have been shared with you. @@ -31 +37 @@ The Machine Learning Lens is available as an AWS-official lens in the [Lens Cata -To get started, follow the steps in [Adding a lens to a workload](https://docs.aws.amazon.com/wellarchitected/latest/userguide/lenses-add.html) and select the **Machine Learning Lens**. +To begin reviewing your machine learning workload, download and import the [Machine Learning Lens](https://github.com/aws-samples/sample-well-architected-custom-lens/blob/main/machine-learning-lens/machine-learning-lens.json) into AWS Well-Architected Tool from the public [AWS Well-Architected custom lens GitHub repository](https://github.com/aws-samples/sample-well-architected-custom-lens). @@ -39 +45 @@ To use the Amazon Web Services Documentation, Javascript must be enabled. Please -Well-Architected Framework pillars +Design principles