Amazon SageMaker AI 多轮强化学习实战：以 Wordle 为例跑通完整流程
Multi-Turn Reinforcement Learning on Amazon SageMaker AI: A hands-on guide
精选理由
AWS 发的实战教程，用 Wordle 教你在 SageMaker 上跑多轮强化学习，从 Bedrock AgentCore 部署到训练评估都覆盖了。
这篇指南介绍如何在 Amazon SageMaker AI 上做多轮强化学习（MTRL），用整个多步任务的结果作为奖励，而不是依赖逐步标注。文章以 Wordle 游戏为例，演示将 agent 和奖励函数部署到 Amazon Bedrock AgentCore，再完成数据集准备、训练与评估的完整流程。训练后的模型按服务时的设置进行评估。
原文 · AWS Machine Learning Blog
Multi-Turn Reinforcement Learning on Amazon SageMaker AI: A hands-on guide
Multi-Turn Reinforcement Learning (MTRL) on Amazon SageMaker AI trains models on the outcome of a whole multi-step task rather than labeled steps. This hands-on guide walks the full pipeline using Wordle: deploying the agent and reward on Amazon Bedrock AgentCore, preparing the dataset, training, and evaluating at serving settings.