跳转至

Real Agents

Abstract

目前还在写论文阶段，暂未开源

Real Agents是一个结合传统人工智能方法和大语言模型（Large Language Models，LLM），适用于生成式人工智能代理（Generative AI Agents）的规划框架,包含一个Unity插件和演示项目。

本项目的开发事由为完成我的本科毕业论文（设计）以及个人对游戏AI的兴趣。

项目链接

https://github.com/AkiKurisu/Real-Agents

示例项目

特点

  1. 结合目标导向的行为规划方法(Goal Oriented Action Planning,GOAP)
  2. 使用自然语言操控NPC执行和规划游戏中定义的行为集合
  3. 代理可以学习并优化决策

说明

由于运行时使用OpenAI API，游戏不具稳定性，仅用作实验用途

依赖

  1. AkiAI https://github.com/AkiKurisu/AkiAI
  2. AkiBT https://github.com/AkiKurisu/AkiBT
  3. AkiGOAP https://github.com/AkiKurisu/AkiGOAP
  4. Next Gen Dialogue https://github.com/AkiKurisu/Next-Gen-Dialogue

下载Demo源码或EXE

见Release页面

Demo部署

如果选用词向量嵌入则需要额外本地部署LangChain Server

命令行输入pip install requirement.txt安装依赖库

原理

先根据GOAP的数据用LLM标注Action和Goal。

运行时让Agent和Planner根据世界状态（WorldStates）生成和搜索Plan。

比对Plan，LLM反思并迭代Memory。

之后可关闭Planner，由Agent生成Plan。

流程图

流程图

步骤1：预处理

首先，开发者定义了一个可执行行为集合(ActionSet)和目标集合(GoalSet)。

GOAP规则下的行为集合

每个行为都基于GOAP规则下，具有先决条件Precondition和效果Effect。 

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
public class Eat : DescriptiveTask
{
        protected sealed override void SetupDerived()
        {
            Preconditions[States.HasFood] = true;
            Preconditions[States.IsHungry] = true;
        }
        protected sealed override void SetupEffects()
        {
            Effects[States.IsHungry] = false;
        }
}

其次，向大语言模型输入Action，将其总结为自然语言作为代理对行为的初始印象(InitialImpression)和长期记忆(Summary),并将短期记忆(Comments)置空。 

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
"Name": "Cook",
"InitialImpression": "To Cook, you need to have energy and ingredients, and as a result, you will have food.",
"Summary": "Latest summary",
"Comments": [
  "",
  "",
  "",
  "",
  ""
]
向大语言模型输入Goal，让LLM生成自然语言作为使用该Goal的解释(Explanation)。并将其组合获得该集合G的解释。

Real Agents中你可以直接在RealAgentSet的编辑器中点击Self Description生成上述数据。

自我描述

步骤2：学习

调整代理模式（AgentMode），设置为Training。

此时运行的Plan由Planner搜索，Agent同时生成一份Plan。如果存在差异的行为，由LLM对其原因进行反思，并生成Comment作为短期记忆注入数据库。

Agent在短期记忆达到阈值时根据初始印象、评论和当前长期记忆对其进行总结归纳，覆盖长期记忆，以此进行迭代。

步骤3：实时规划

根据外界输入的目标，代理在虚拟空间中实时推理。

演示项目Agent操作界面

LLM根据自身对每个行为的记忆找到合适的Plan，由于没有Planner对比，此时仅判断Plan是否具有可行性，如不具备，则由LLM生成（Comment）注入数据库。

Agent生成并评估Pla

Agent生成并评估Plan

Agent执行Plan

Agent执行Plan

推理后端

实验以OpenAI的ChatGPT3.5为推理模型，可选用本地部署的ChatGLM3-6B，不过经过测试可能难以完成路径搜索任务。

拓展

本项目为行为AI，你可以将其接入对话插件中，这样玩家可以根据对话来改变AI行为。

如果你对和NPC对话感兴趣可以参考对话插件：VirtualHuman-Unity

许可证

MIT

Demo许可证引用

VRM 模型请参阅各个作者的许可说明。

Polygon Fantasy Kingdom is a paid asset.

https://assetstore.unity.com/packages/3d/environments/fantasy/polygon-fantasy-kingdom-low-poly-3d-art-by-synty-164532

Example scene is optimized by Scene Optimizer created by [Procedural Worlds]

https://assetstore.unity.com/publishers/15277

Lowpoly Environment can be downloaded in assetstore for free.

https://assetstore.unity.com/packages/3d/environments/lowpoly-environment-nature-free-medieval-fantasy-series-187052

Heat - Complete Modern UI is paid asset

https://assetstore.unity.com/packages/2d/gui/heat-complete-modern-ui-264857

GUI-CasualFantasy is paid asset

https://assetstore.unity.com/packages/2d/gui/gui-casual-fantasy-265651

Unity Chan animation is owned by Unity Technology Japan and under UC2 license.

Medieval Animations Mega Pack is paid asset.

https://assetstore.unity.com/packages/3d/animations/medieval-animations-mega-pack-12141

Unity Starter Asset is under the Unity Companion License and can be downloaded in assetstore for free.

https://assetstore.unity.com/packages/essentials/starter-assets-thirdperson-updates-in-new-charactercontroller-pa-196526

UniVRM is under MIT license.

https://github.com/vrm-c/UniVRM

AkiFramework is under MIT license.

https://github.com/AkiKurisu/AkiFramework

Demo中使用的其他动画、音乐、ui、字体资源均为网络资源，请勿用于商业用途。

引用本仓库

1
2
3
4
5
6
7
8
9
@misc{realagents,
  author = {YiFei Feng},
  title = {Real Agents: An planning framework for generative artificial intelligence agents},
  year = {2024},
  publisher = {GitHub},
  journal = {https://github.com/AkiKurisu/Real-Agents},
  school = {East China University of Political Science and Law},
  location = {Shanghai, China}
}

参考文献

  1. Steve Rabin, Game AI Pro 3: Collected Wisdom of Game AI Professionals,International Standard Book.

  2. [Orkin 06] Orkin, J. 2006. 3 states and a plan: The AI of F.E.A.R., Game Developers Conference, San Francisco, CA.

  3. Joon Sung Park, Joseph C. O’Brien, Carrie J. Cai, Meredith Ringel Morris,Percy Liang, and Michael S. Bernstein. 2023. Generative Agents: Interactive Simulacra of Human Behavior.
  4. Wang, Zihao et al. “Describe, Explain, Plan and Select: Interactive Planning with Large Language Models Enables Open-World Multi-Task Agents.” ArXiv abs/2302.01560 (2023): n. pag.
  5. Lin J, Zhao H, Zhang A, et al. Agentsims: An open-source sandbox for large language model evaluation[J]. arXiv preprint arXiv:2308.04026, 2023.
  6. Xi, Z., Chen, W., Guo, X., He, W., Ding, Y., Hong, B., Zhang, M., Wang, J., Jin, S., Zhou, E., Zheng, R., Fan, X., Wang, X., Xiong, L., Liu, Q., Zhou, Y., Wang, W., Jiang, C., Zou, Y., Liu, X., Yin, Z., Dou, S., Weng, R., Cheng, W., Zhang, Q., Qin, W., Zheng, Y., Qiu, X., Huan, X., & Gui, T. (2023). The Rise and Potential of Large Language Model Based Agents: A Survey. ArXiv, abs/2309.07864.