AWS decision-guides documentation change
Summary
Added a comprehensive guide on choosing AWS generative AI services, covering overview, key concepts, use cases, service comparisons, and implementation resources.
Security assessment
The guide includes security considerations in data handling (privacy, security, bias mitigation) and references secure implementation patterns (IAM Identity Center integration). However, no specific vulnerabilities or incidents are addressed.
Diff
diff --git a/decision-guides/latest/generative-ai-on-aws-how-to-choose/guide.md b/decision-guides/latest/generative-ai-on-aws-how-to-choose/guide.md index 8b1378917..a33a477a0 100644 --- a//decision-guides/latest/generative-ai-on-aws-how-to-choose/guide.md +++ b//decision-guides/latest/generative-ai-on-aws-how-to-choose/guide.md @@ -0,0 +1 @@ + @@ -1,0 +3,333 @@ +# Choosing an AWS generative AI service + +Determine which AWS generative AI services are the best fit for your organization. + +## Overview of AWS generative AI services + +Generative AI is a set of artificial intelligence (AI) systems and models designed to generate content such as code, text, images, music, or other forms of data. These systems can produce new content based on patterns and knowledge learned from existing data. + +_This twelve-minute video discusses building generative AI applications on AWS, part one of a four-part series. View[part two ](https://www.youtube.com/watch?v=HUuO9eJbOTk), [part three ](https://www.youtube.com/watch?v=JbizhCFmC20), and [part four ](https://www.youtube.com/watch?v=gNY06JM6m8U). _ + +**Increasingly, organizations and businesses are using generative AI to:** + + * **Automate creative workflows:** Use generative AI services to automate the workflows of time-consuming creative processes such as writing, image or video creation, and graphic design. + * **Customize and personalize content:** Generate targeted content, product recommendations, and customized offerings for an audience-specific context. + * **Augment data:** Synthesize large training datasets for other ML models to unlock scenarios where human-labeled data is scarce. + * **Reduce cost:** Potentially lower costs by using synthesized data, content, and digital assets. + * **Faster experimentation:** Test and iterate on more content variations and creative concepts than would be possible manually. + + + +## Key concepts + +Amazon offers a range of generative AI services, applications, tools, and supporting infrastructure. Which of these you use depends largely on what you’re trying to accomplish, how much choice you need in foundation models, the degree of customization you need in your generative AI applications, and the expertise within your organization. + +**Foundation models and types** + +Foundation models (FMs) are the backbone of generative AI. These pre-trained AI models can be customized for specific tasks, and come in various types: + + * **Text models:** Process and generate natural language + * **Image models:** Work with visual data for tasks like image generation or analysis + * **Multimodal models:** Handle multiple types of data simultaneously + + + +The capabilities of an FM are often related to its size, measured in parameters. Larger models can capture more complex patterns, but require more computational resources. + +After you've decided on a generative AI service, choose the foundation model (FM) that gives you the best results for your use case. Amazon Bedrock has a [model evaluation capability ](/bedrock/latest/userguide/evaluation.html) that can assist in evaluating, comparing, and selecting the best FMs for your use case. + +### Public foundation models + +Amazon Bedrock provides access to a variety of foundation models from different companies. These include [Anthropic Claude ](https://aws.amazon.com/bedrock/claude/), [Cohere Command & Embed ](https://aws.amazon.com/bedrock/cohere-command-embed/), [AI21 Labs Jurassic ](https://aws.amazon.com/bedrock/jurassic/), [Meta Llama ](https://aws.amazon.com/bedrock/llama/), [Mistral AI ](https://aws.amazon.com/bedrock/mistral/), [Stable Diffusion XL ](https://aws.amazon.com/bedrock/stable-diffusion/), [Amazon Nova ](https://aws.amazon.com/ai/generative-ai/nova/), and [Amazon Titan ](https://aws.amazon.com/bedrock/titan/). With this multi-model approach, you can select and compare different models based on your specific use case requirements. Using Amazon Bedrock and Amazon SageMaker AI, you can experiment with a variety of foundation models, and privately customize them with your data. + +[Learn more](https://aws.amazon.com/what-is/foundation-models/) + +### Model customization and enhancement + +Adapting FMs to specific use cases is crucial for optimal performance. Key customization techniques include: + + * **Fine-tuning** : Further training the model on domain-specific data + * **Continuous pre-training** : Extending the model's knowledge with new data over time + * **RAG ([ retrieval augmented generation](/sagemaker/latest/dg/jumpstart-foundation-models-customize-rag.html)) ** : Enhancing responses by retrieving external information + + + +Prompt engineering is another vital skill, allowing users to craft inputs that guide the model to produce desired outputs effectively. + +[Get started](/bedrock/latest/userguide/custom-models.html) + +### Performance considerations + +When you evaluate generative AI model performance, consider two factors: + + * **Inference latency** : The speed at which a model generates responses + * **Context window** : Amount of text a model can consider at once + + + +[Get started](/prescriptive-guidance/latest/gen-ai-workload-assessment/introduction.html) + +### Data considerations + +The quality and quantity of data significantly impact model performance. Key factors include: + + * **Proper data handling** : Ensuring data security and privacy, effective annotation and labeling processes, and the right balance between data quantity with quality + * **Data relevance** : How well the data matches the intended use case + * **Data accuracy and consistency** + * **Bias and fairness** : Ensuring training data doesn't perpetuate harmful biases + + + +[Learn more]( https://aws.amazon.com/blogs/big-data/data-governance-in-the-age-of-generative-ai/) + +## Use cases + +Transform business processes and content creation with AI-powered automation and insights + +Enterprise knowledge management and support + + * Querying and analyzing company data across multiple enterprise systems + * Generating summaries and insights from internal documents + * Creating automated responses for customer service + * Accessing and synthesizing information from various business data sources like Salesforce, SharePoint, and Confluence + + + +**Learn more:** [ New Amazon Bedrock capabilities enhance data processing and retrieval ](https://aws.amazon.com/blogs/aws/new-amazon-bedrock-capabilities-enhance-data-processing-and-retrieval/) + +[  Amazon Q Business ](/amazonq) [  Amazon Bedrock ](/bedrock) [  Amazon Sagemaker AI ](/sagemaker) + +Software development and DevOps + + * Code generation and optimization + * Application testing and debugging + * Infrastructure management and AWS resource optimization + * Automated code upgrades and security fixes + * Technical documentation creation and maintenance + + + +**Get started:** [ How generative AI is transforming developer workflows at Amazon ](https://aws.amazon.com/blogs/devops/how-generative-ai-is-transforming-developer-workflows-at-amazon/) + +[  Amazon Q Developer ](/amazonq) + +Content creation & customizations + + * Automated writing and content generation + * Image and multimedia content creation + * Personalized content and recommendations for specific audiences + * Document classification and analysis + * Converting data into visualizations and reports + + + +**Get started:** [ Experiment with Amazon Nova foundation models ](https://nova.amazon.com/chat) + +[  Amazon Bedrock ](/bedrock) [  Amazon Nova ](/nova) + +Business process automation + + * Automating creative workflows and time-consuming manual tasks + * Building conversational AI applications and chatbots + * Data analysis and pattern recognition + * Creating synthetic training data for ML models + * Process automation through multi-step reasoning and planning + + + +**Get started:** [ Amazon Q Business is adding new workflow automation capability and 50+ action integrations ](https://aws.amazon.com/blogs/aws/amazon-q-business-is-adding-new-workflow-automation-capability-and-50-action-integrations/) + +[  Amazon Bedrock ](/bedrock) [  Amazon SageMaker AI ](/sagemaker) [  Amazon Q ](/amazonq) + +## Compare services + +The AWS generative AI stack provides options at every layer of implementation, from ready-to-use applications to foundational infrastructure. This diagram shows how these services relate to each other. + + + +As you choose services, consider your team’s technical expertise, development timeline, and specific use case requirements. The following table helps you match your business needs to the appropriate AWS service. + +If you need to... | To get started +---|--- +Generate code and get answers to business questions across your enterprise data | [ Amazon Q Business ](/amazonq/latest/qbusiness-ug/setting-up.html) [ Amazon Q Developer in chat applications ](/amazonq/latest/qdeveloper-ug/q-in-IDE-setup.html) +Choose from multiple foundation models, customize them with your data, and build generative AI applications | [ Amazon Bedrock ](/bedrock/latest/userguide/getting-started.html) [ Amazon Nova ](/nova/latest/userguide/getting-started.html) [ Amazon Titan ](/bedrock/latest/userguide/titan-models.html) +Build, train, and deploy machine learning models at scale | [Amazon SageMaker AI ](/sagemaker/latest/dg/gs.html) +Maximize price-performance for foundation model training and inference | [AWS Trainium ](/dlami/latest/devguide/trainium.html) [ AWS Inferentia ](/eks/latest/userguide/inferentia-support.html) + +## Start building + +Now that we've covered the criteria you need to apply in choosing an AWS generative AI service, you can select which services are optimized for your needs and explore how you might get started using each of them. + +Amazon Q Business + +**[What is Amazon Q Business?](/amazonq/latest/qbusiness-ug/what-is.html) ** + +Get an overview of Amazon Q Business, with explanations of what it is, how it works, and how to get started using it. + +**[Create a sample Amazon Q Business application](/amazonq/latest/qbusiness-ug/quick-create-app.html) ** + +Learn how to create your first Amazon Q Business application in either the AWS Management Console or using the command line interface (CLI). + +**[ Combine Amazon Q Business and AWS IAM Identity Center to build generative AI ](https://aws.amazon.com/blogs/machine-learning/build-private-and-secure-enterprise-generative-ai-apps-with-amazon-q-business-and-aws-iam-identity-center/) ** + +Build private and secure enterprise generative AI apps with Amazon Q Business and AWS IAM Identity Center. + +Amazon Q Developer + +**[What is Amazon Q Developer?](/amazonq/latest/qdeveloper-ug/what-is.html) ** + +Get an overview of Amazon Q Developer, with explanations of what it is, how it works, and how to get started using it. + +**[Understanding tiers of service for Amazon Q Developer](/amazonq/latest/qdeveloper-ug/q-tiers.html) ** + +Review the following information to understand the tiers of service for Amazon Q Developer, including Amazon Q Developer Pro and Amazon Q Developer at the free tier. + +**[Working with Amazon Q Developer ](https://builder.aws.com/learn/topics/agentic-ai) ** + +Use the Amazon Q Developer Center for fast access to key Amazon Q Developer articles, blog posts, videos, and tips. + +Amazon Bedrock + +**[What is Amazon Bedrock?](/bedrock/latest/userguide/what-is-bedrock.html) ** + +Learn how to use this fully managed service to make foundation models (FMs) from Amazon and third parties available for your use through a unified API.