当前位置:首页 > 报告详情

利用生成式人工智能和LLM代理加速游戏设计评审[重复].pdf

上传人: 明**** 编号:1013598 2025-12-21 17页 410.71KB

1、 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.I N D 3 9 5-RAccelerating Game Design Reviews with Generative AI and LLM AgentsChristina Defoor(she/her)Solutions ArchitectSam Patzer(he/him)Solutions Architect 2025

2、,Amazon Web Services,Inc.or its affiliates.All rights reserved.AgendaIntroducing the game analystWhat are we buildingUnderstanding the tools in our toolboxCode timeQA 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.2025,Amazon Web Services,Inc.or its affiliates.All rights reserved

3、.What is the role of a game analyst?2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.“The modern game analyst serves as both interpreter and storyteller.They interpret vast oceans of player data,transforming seemingly random actions into coherent narratives about player behavior.Th

4、rough this lens,they dont just see numbers-they see players struggling with a difficult level,engaging with new content,or finding unexpected ways to enjoy the game.This understanding becomes the foundation for informed design decisions.”-Ben Medler 2025,Amazon Web Services,Inc.or its affiliates.All

5、 rights reserved.Breaking down the core competenciesGame Analyst Quality AssuranceGame StrategyCore CompetenciesLore SpecialistGame Design 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.Time to TrainCurrent ChallengesSubjective FeedbackSlower Iteration 2025,Amazon Web Services,In

6、c.or its affiliates.All rights reserved.What we are buildingQA AgentStrategy AgentSpecialized AgentsData SourcesLore AgentGameplay AgentGame Design DocumentsStrategy DocumentsLive Game DataGame Analyst Orchestrator Agent 2025,Amazon Web Services,Inc.or its affiliates.All rights reserved.BenefitsAcce

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
根据报告的内容,全文主要内容概括如下: - **游戏分析师角色**:游戏分析师是数据解释者和故事讲述者,通过分析玩家数据,将玩家行为转化为设计决策的基础。 - **项目目标**:使用生成式AI和LLM代理加速游戏设计审查。 - **工具和代理**:构建QA代理、策略代理、专门代理等,利用数据源如游戏数据、策略文档等。 - **工具优势**:易于使用、可扩展、强大功能、快速开发。 - **Strands Agents**:开源Python SDK,用于构建高效AI代理。 - **Amazon Bedrock AgentCore**:简化代理开发,支持快速原型设计和迭代。 - **关键点**: - 加速创新周期。 - 提高协作智能。 - 实现无偏见的标准基准评估。 - 使用AgentCore Memory增强代理性能和用户体验。 - 通过GitHub和演示视频了解更多信息。
AI加速设计" AI如何助力游戏迭代" AI助手大显身手"
客服
商务合作
小程序
服务号
折叠