Could you tell me about RAG chat education material by Amazon AWS Japan presents".生成AI体験ワークショップ" (Amazon AWS japan presents) what kind of function has the 👎 mark on it?

Based on the available information, I cannot provide specific details about the 👎 mark functionality in the "生成AI体験ワークショップ" (Generative AI Experience Workshop) presented by AWS Japan.

The feedback interface you're describing appears to be a rating system where users can provide feedback on responses, with options like "incorrect," "outdated information," "harmful or offensive," and "other" along with an optional text field for additional comments.

Regarding Retrieval Augmented Generation (RAG) systems in general, they work by retrieving relevant information from external datasets before generating responses using a large language model. This helps reduce hallucinations (incorrect information) and allows integration with company-specific data.

RAG systems on AWS typically use services like Amazon Kendra for retrieval, Amazon Bedrock for the language model, and Amazon S3 for data storage. However, standard RAG implementations do not automatically use user feedback to retrain the foundation model. User feedback is typically collected for evaluation and improvement purposes rather than being used as additional training data for the underlying AI model.

For specific details about how the feedback system in the AWS Japan workshop functions, I would recommend contacting AWS Japan support directly or referring to the workshop documentation.
Sources
Community | Prototype a RAG chatbot with Amazon Bedrock, Kendra, and Lex
Community | RAG made easy using Bedrock - S03 E13 - Build On Generative AI