- 新しい順
- 投票が多い順
- コメントが多い順
Based on the information available, I cannot provide specific details about the "生成AI体験ワークショップ" (Generative AI Experience Workshop) by AWS Japan as it's not mentioned in the provided sources.
However, I can share some general information about RAG (Retrieval Augmented Generation) implementations on AWS:
Regarding your questions about chat history storage:
-
In typical RAG implementations on AWS, chat history can be stored in various ways depending on the architecture. Some implementations use Amazon Lex to manage session state for conversation history. There's no specific confirmation that chat communications are automatically saved to S3 in all RAG implementations.
-
If chat history is being stored, access would depend on how the specific workshop implementation was designed and what permissions were configured for your AWS account.
-
Without knowing the specific implementation details of the workshop you're referring to, I cannot provide exact steps for retrieving chat history.
For a typical RAG chatbot implementation on AWS, conversation history might be managed through:
- Session state in Amazon Lex
- Custom storage solutions that could include S3, databases, or other storage services
- Knowledge bases in Amazon Bedrock
If you're working with a specific AWS Japan workshop, I would recommend:
- Checking the workshop documentation for details about data storage
- Reviewing any IAM permissions you have in your account
- Contacting AWS Japan support or the workshop facilitators for implementation-specific questions
For general RAG implementations, you might explore the Amazon SageMaker notebook instances that are sometimes used to interact with RAG systems, as they could contain conversation logs or access to stored data.
Sources
Community | Prototype a RAG chatbot with Amazon Bedrock, Kendra, and Lex
Deploy a RAG use case on AWS by using Terraform and Amazon Bedrock - AWS Prescriptive Guidance
