3# NAVI EDU 演讲稿 / Presentation Script
中文： 各位朋友，还记得那首诗吗？"Do not go gentle into that good night"——不要温和地走进那个良夜。今天我想和大家分享一个关于质疑、思考与创新的故事。
我是Anna Wang，曾经是个"坏学生"，现在站在这里，代表NAVI EDU申请GIGA千兆加速器项目。我们要做的，就是弥补传统教育的不足，在学习与创造之间找到那个黄金平衡点。
English: My friends, do you remember that poem? "Do not go gentle into that good night." Today I want to share with you a story about doubting, reflecting, and innovating..
I'm Anna Wang, once a "bad student," now standing here representing NAVI EDU for the GIGA Gigabit Accelerator Program. What we're doing is making up for the shortcomings of traditional education, finding that golden balance between learning and creation.
中文： 在我开始之前，让我问你们一个问题。你们觉得我们在做什么产品？
是教育软件？是游戏？还是金融产品？其实...我们做的是梦想。我们做的是可能性。
English: Before I begin, let me ask you a question. What do you think our product is?
Educational software? A game? A financial product? Actually... we're building dreams. We're building possibilities.
中文： 让我告诉你们我是怎么走到今天的。想象一下，一个在高考制度下挣扎的孩子，每天被分数绑架，被固定思维折磨。那就是曾经的我。
但是你知道吗？有时候，正是这种痛苦让你看到了问题的本质。我从CHU到NTU，再到成为AI产品经理，最后研究Calabi-Yau流形——每一步都在思考同一个问题：教育到底应该是什么样的？
English: Let me tell you how I got here. Imagine a child struggling under the exam system, kidnapped by scores every day, tortured by fixed mindset. That was me.
But you know what? Sometimes it's exactly that pain that lets you see the essence of the problem. From CHU to NTU, then becoming an AI Product Manager, finally researching Calabi-Yau manifolds—every step was thinking about the same question: what should education really look like?
中文： 这就是我想和大家探讨的核心问题：在我们现在的教育体系里，那些爱问"为什么"的孩子，那些有奇思妙想的孩子，他们怎么生存？他们的创造力去哪儿了？
我见过太多聪明的孩子被磨成了考试机器。这不是教育，这是谋杀——谋杀好奇心，谋杀创造力。
English: This is the core issue I want to explore with you: In our current education system, what happens to those children who love asking "why"? Those children with wild imaginations? How do they survive? Where does their creativity go?
I've seen too many brilliant children ground down into exam machines. This isn't education—it's murder. Murder of curiosity, murder of creativity.
中文： 但是我相信，每个孩子都是一颗种子。你给他合适的土壤，充足的阳光，他就会长成参天大树。
我们的解决方案很简单：让学生自主学习，让他们互相连接。你知道为什么学校基础设施有差异？因为我们还在用工业时代的思维做信息时代的教育。孩子们是流水线上的商品。
我们要做的就是打破这个格局。
English: But I believe every child is a seed. Give them the right soil, sufficient sunlight, and they'll grow into towering trees.
Our solution is simple: let students learn autonomously, let them connect with each other. You know why school infrastructure varies? Because we're still using industrial-age thinking for information-age education.Children are like products being mass-produced on a production line.
What we're doing is breaking that pattern.
中文： 现在我要告诉你们一个令人震惊的发现。我们都知道，我们的起源源于遗传物质的划分和分化，因此我们从数学的角度对整个过程进行了建模。在细胞层面上，使用了螺旋函数，而在组织层面上，则实施了一种博弈收敛设计来确保分化。 实验结果是可喜的，这也成为了我们整个项目设计的数学基础。
English: Now I'm going to tell you about a shocking discovery. We all know that we originated from the division and differentiation of genetic material, so we modeled this entire process from a mathematical perspective. On the cellular level, spiral functions were utilized, while on the tissue level, a game convergence design was implemented to ensure differentiation. The results of the experiment were gratifying and served as the mathematical foundation for the overall design of our project.
中文： 在复杂的数学建模之外，形成了我们全新的教育认知。毕竟谁会比人体更擅长上万年的学习和创造呢？
你们有没有想过，人生其实就是一个巨大的游戏？
在这个游戏里，你不断地学习，不断地创造，不断地成长。你的每个技能，每个经历，每个思考，最终都会汇聚在一起，让你成为独一无二的自己。
这就是我们产品的核心理念。
English: Beyond the complexities of mathematical modeling, a completely new understanding of education has emerged. After all, who could possibly surpass the human body in terms of its ability to learn and create over thousands of years?
Have you ever thought that life is actually one massive game?
In this game, you constantly learn, constantly create, constantly grow. Every skill, every experience, every thought you have will eventually converge to make you uniquely yourself.
This is the core philosophy of our product.
中文： 所以我们做了什么？人类最成熟的博弈论诞生了金融工具。所以我们把学习变成了交易，把知识变成了货币，把创造变成了投资。
听起来疯狂？不，这才是创造！我们用模拟金融的方式，让孩子们在游戏中学习，在学习中创造，在创造中成长。
English: So what did we do?The most mature form of game theory in human history has given rise to financial instruments. So we turned learning into trading, knowledge into currency, creation into investment.
Sounds crazy? No, this is creation! We use simulated finance to let children learn through gaming, create through learning, grow through creating.
中文： 你们看到的这个界面，不是普通的交易平台。这是知识的华尔街，智慧的纳斯达克！
e代币在这里不是虚拟货币，是学习成果的量化体现。每一次交易，都是一次思维的碰撞；每一次涨跌，都是知识价值的重新评估。
English: What you're seeing isn't an ordinary trading platform. This is the Wall Street of knowledge, the NASDAQ of wisdom!
e tokens here aren't virtual currency—they're quantified learning achievements. Every trade is a collision of minds; every rise and fall is a revaluation of knowledge.
中文： 你们知道为什么我们选择1.618这个黄金比例吗？
因为这个数字藏着宇宙的秘密！从向日葵的花瓣到贝壳的螺纹，从人体的比例到星系的结构，这个比例无处不在。
我们的教育哲学也基于这个比例：应用知识、拓展思维、减轻焦虑。这不是巧合，这是自然的法则！
English: Do you know why we chose 1.618, the golden ratio?
Because this number holds the universe's secret! From sunflower petals to shell spirals, from human proportions to galaxy structures, this ratio is everywhere.
Our educational philosophy is also based on this ratio: applying knowledge, expanding thinking, reducing anxiety. This isn't coincidence—it's natural law!
中文： 看看我们的内容创作中心！量子计算、GPT训练、Web3分析、创新思维...
这不是简单的课程列表，这是未来的知识地图！每一个模块都是一个世界，每一篇内容都是一次探险。
我们在建造的不是学校，是探索未知的太空站！
English: Look at our content creation center! Quantum computing, GPT training, Web3 analysis, innovative thinking...
This isn't just a course catalog—it's a map of future knowledge! Every module is a world, every piece of content is an adventure.
We're not building schools—we're building space stations for exploring the unknown!
中文： 这些数字会说话！127.5小时学习，78.8小时创作，15篇发布内容，3个NFT收藏...
你看到的不只是数据，是一个学习者的成长轨迹！这个1.618的黄金比例追踪器告诉我们，这个学生已经找到了学习和创造的完美平衡。
English: These numbers speak! 127.5 hours learning, 78.8 hours creating, 15 published contents, 3 NFT collections...
You're not just seeing data—you're seeing a learner's growth trajectory! This 1.618 golden ratio tracker tells us this student has found the perfect balance between learning and creation.
中文： 但是光有梦想还不够，我们还需要可持续的经济模型。
你担心通胀？我们有e币系统！它像血液一样在各个项目间流动，保持生态的平衡。这个设计和我们的螺旋函数理论完美吻合——一切都在收敛，一切都在平衡。
English: But dreams alone aren't enough—we need a sustainable economic model.
Worried about inflation? We have the e coin system! It flows like blood between projects, maintaining ecosystem balance. This design perfectly aligns with our spiral function theory—everything converges, everything balances.
中文： 这就是我们的交易引擎！现在e代币1.25美元，你可以实时买入。
但这不是投机，这是投资未来！每一次交易背后，都是一个学习者的努力，一个创造者的梦想。
English: This is our trading engine! e tokens at $1.25 right now, you can buy in real-time.
But this isn't speculation—this is investing in the future! Behind every trade is a learner's effort, a creator's dream.
中文： 最后，让我告诉你们我们最骄傲的创新：每个学习模块都有专门的知识图谱构建师，配合AI智能代理。
想象一下，每个学生都有一个专属的AI导师，它了解你的学习风格，知道你的兴趣点，能够为你量身定制学习路径。
这不是科幻小说，这就是我们正在做的事情！
English: Finally, let me tell you about our proudest innovation: every learning module has dedicated knowledge graph architects, working with AI agents.
Imagine every student having a personalized AI mentor that understands your learning style, knows your interests, and can customize learning paths just for you.
This isn't science fiction—this is what we're building right now!
中文： 我们的核心模块包括黄金比例引擎、智能内容系统和用户身份管理。技术栈配置采用React + TypeScript前端，Node.js + PostgreSQL后端，定制轻量级游戏引擎，以及差分隐私技术。
English: Our core modules include the Golden Ratio Engine, Intelligent Content System, and User Identity Management. Technology stack configuration uses React + TypeScript frontend, Node.js + PostgreSQL backend, custom lightweight gaming engine, and differential privacy technology.
中文： 那么，如何运行我们的项目呢？
English: So, how do we proceed with our project?
中文： 我们提供两种主要服务：网络服务的开源保障，以及游学项目的质量保证。
English: We provide two main services: open-source assurance for internet services, and quality assurance for study tours.
中文： Navigator游学项目致力于拓宽视野、培养兴趣、播种希望。通过利用网站的知识图谱，快速构建学生产业蓝图。同时，作为开源网站的线下质量保障，采用半合约链式增长模式，为当地社区创造就业机会和收入。
English: The Navigator Study Program is dedicated to broadening perspectives, cultivating interests, and sowing hope. Building a student industry blueprint quickly by leveraging the knowledge graph of the website. At the same time, as an offline quality assurance for an open-source website, the semi-contractual chain-like growth model is employed to create job opportunities and generate income for the local community.
中文： 我们的财务计划是什么？
English: What is our financial plan?
中文： 我们的内部收益率达到329.4%，回收期仅需3.6个月，极短的回收期意味着极低的风险。3年净现值为21.5万美元，具有显著的长期价值。投资倍数达到66.5倍，盈利能力极强。投标地方政府项目只需要一份工资。
English: Our Internal Rate of Return reaches 329.4%, with a payback period of only 3.6 months - an extremely short payback with very low risk. The 3-year NPV is $215,960, showing significant long-term value. Investment multiple reaches 66.5x, demonstrating extremely strong profitability. To bid on local government projects, all that is required is a salary.
中文： GIGA 5万美元资金使用计划：产品开发2万美元，试点扩展1.75万美元，团队成长1.25万美元。关键里程碑包括：1000名活跃学习者，50名教师培训，完成开源代码发布，5国试点部署。
English: GIGA $50,000 Funding Usage Plan: Product Development $20,000, Pilot Expansion $17,500, Team Growth $12,500. Key Milestones include: 1,000 active learners, 50 teachers trained, complete open source code release, 5-country pilot deployment.
中文： "新教育将基于对个体差异的尊重，让每个人沿着自己的热情探索世界。"
教育是释放人类潜能的关键。通过尊重个体差异和激发内在智慧，我们正在为新时代的教育体系奠定基础。
我们正在行动，这是我们在西安研学基地为铁一曲江小学生们进行研学的场景。我们希望有更多的机会可以向大家推广我们的项目。
English: "New education will be based on respect for individual differences, allowing each person to explore the world along their own passion."
Education is key to unlocking human potential. By respecting individual differences and inspiring inner wisdom, we're laying the foundation for a new era's educational system.
We are actively working on this project. This is a scene from our educational study program for primary school students from Tieyi Jujiang in the Xi'an study base. We hope for more opportunities to promote our project to a wider audience.
中文： 感谢大家的聆听！这就是我们NAVI EDU的愿景——在学习与创造之间找到理想的平衡。我们期待与GIGA千兆加速器项目合作，共同推动教育创新的未来。
English: Thanks for listening! This is NAVI EDU's vision - finding the ideal balance between learning and creation. We look forward to collaborating with the GIGA Gigabit Accelerator Program to drive the future of educational innovation together.
重大更新：黄金比例算法支持的多智能体架构与可编辑知识图谱，让项目变成强大的私人AI，可本地部署。 Major Update: The multi-agent architecture and editable knowledge graph supported by the golden ratio algorithm transform the project into a powerful private AI that can be deployed locally.
重大更新2： 1.产品优化：灵感源自于线上约会软件tinder与bumble，他们对于主动权的设计和侧滑筛选的设计令人记忆犹新。结合了本项目的模拟金融游戏属性。
识别用户需求，期望在前的先给出知识树，然后分点订正；要求在前的先给出要求内容 再形成知识树。
根据以上更新在知识交易所增加期货与现货机制
NAVI的双栏协作更新为分知识点左右侧滑（赞同/质疑）
2.全域营销计划：灵感源自于游戏机器人
为游戏化知识交易所增加促流动性机器人
Major Update 2:
- Product Optimization: Inspired by the online dating apps Tinder and Bumble, their designs for initiative and swiping selection are still fresh in our memory. These elements have been integrated with the simulated financial game nature of this project.
Identify user needs. If expectations are stated first, provide the knowledge tree first and then correct it point by point; if requirements are stated first, provide the required content first and then form the knowledge tree.
Based on the above update, add the futures and spot mechanisms to the knowledge exchange.
NAVI's dual-column collaborative update has been changed to a left-right sliding format by knowledge points (Agree/Question)
- Omnidirectional Marketing Plan: Inspired by Game Robots
Adding liquidity-promoting robots to the gamified knowledge exchange
我意识到使用开源方式并不能等同于展示财务计划，评估机构并不牢靠。而开源项目的一次MVP实践将是最好的证明。或许我们应该从此入手，签订一种全新的对赌协议。
I realize that adopting an open-source approach does not equate to presenting financial plans, and evaluation institutions are not reliable. A successful MVP practice of an open-source project would be the best proof. Maybe we should start from here and sign a brand-new type of wager agreement.
我们正在探索全新的开源融资方式
We are exploring a completely new open-source financing method.
重大更新3：
一花一世界，一叶一菩提 - 每个知识点都是一个完整的知识宇宙
Navi 是一个革命性的智能学习助手，通过多智能体协作架构与分形知识图谱，为用户提供真正个性化的学习体验。我们相信每个知识节点都蕴含着整个知识宇宙的奥秘。
| 智能体 | 职责 | 特色 |
|---|---|---|
| 🎓 学习辅导助手 | 系统化知识传授与学习指导 | 结构化教学 |
| 🤔 批判思考助手 | 培养批判性思维，提出深度问题 | 多角度质疑 |
| ⚖️ 协调平衡助手 | 整合观点，构建个人知识图谱 | 黄金比例决策 |
# 每个知识节点都是一个完整的分形宇宙
knowledge_node = {
"macro_view": "38.2% 抽象概览", # 黄金比例宏观
"meso_core": "核心知识内容", # 当前层级
"micro_details": "61.8% 深度细节" # 黄金比例微观
}- 专业智能体各司其职，协同提供全方位学习支持
- 动态权重调整，基于学习反馈优化智能体参与度
- 知识融合引擎，自动整合多元观点形成完整体系
- 黄金比例存储结构：每个节点按 0.618:1 比例分层
- 自相似性设计：任意子图都与整体结构相似
- 多分辨率访问：从宏观概览到微观细节的无缝切换
用户查询 → 黄金比例决策 → 分形检索 → 多层级响应
↑ ↓
个性化知识图谱 ← 置信度更新 ← 学习效果评估
import navi_learning
# 初始化Navi助手
assistant = navi_learning.NaviAssistant()
# 提出学习问题
response = assistant.ask("请解释量子力学的基本原理")
# 获得分层次回答
print(response.macro_overview) # 宏观理解
print(response.core_concept) # 核心知识
print(response.micro_details) # 深度细节# 开启分形探索模式
deep_dive = assistant.explore_fractal(
topic="神经网络",
resolution="micro" # 深入微观层面
)
# 获取知识宇宙的全景视图
cosmic_view = assistant.get_cosmic_pattern(
local_topic="卷积神经网络",
find_analogies=True # 发现跨领域自相似性
)- 🎯 个性化路径：基于分形置信度的自适应学习路线
- 🔍 深度探索：从任意知识点切入整个知识宇宙
- 💡 直觉理解：黄金比例优化的认知负荷分配
- 📊 洞察学习模式：通过分形维度分析学习行为
- 🎨 课程设计：基于自相似性的知识结构规划
- 🔄 持续优化：动态置信度调整的教学内容
- 自相似索引：基于局部预测全局的知识检索
- 黄金比例缓存：0.618优化原则的内存管理
- 多尺度分析：从概念概览到技术细节的无缝缩放
# 黄金比例决策过程
decision_flow = {
"learning_input": 0.618, # 学习助手权重
"critical_review": 0.382, # 批判思考权重
"balanced_output": "黄金融合结果"
}- 概念深度理解：通过分形探索建立知识联系
- 跨学科学习：发现不同领域的自相似模式
- 研究创新：基于分形思维的原创性发现
- 技能树构建：黄金比例优化的学习路径
- 问题解决：多尺度分析复杂业务问题
- 创新思维：分形模式启发的新解决方案
我们正在构建一个真正理解知识本质的学习系统，其中：
- 🔄 每个学习会话都在丰富全球知识分形
- 🌍 每个用户都贡献于集体智能的进化
- 💫 每个问题都开启一段跨越尺度的探索之旅
开始您的分形学习之旅 - 在Navi的世界里，每个问题都是一个宇宙，每个答案都包含无限可能。
“真正的知识不是信息的堆积，而是模式的理解” - Navi 设计哲学
📖 [详细文档] | 🔧 [开发指南] | 🚀 [快速上手] | 💬 [社区交流]
A flower contains a universe, a leaf holds enlightenment - Every knowledge point is a complete cosmic ecosystem
Navi is a revolutionary intelligent learning assistant that delivers truly personalized learning experiences through multi-agent collaborative architecture and fractal knowledge graphs. We believe every knowledge node contains the mysteries of the entire knowledge universe.
| Agent | Responsibility | Specialization |
|---|---|---|
| 🎓 Learning Tutor | Systematic knowledge transfer & learning guidance | Structured teaching |
| 🤔 Critical Thinking | Cultivates critical thinking & deep questioning | Multi-perspective analysis |
| ⚖️ Balance Coordinator | Integrates perspectives & builds personal knowledge graphs | Golden ratio decision-making |
# Each knowledge node is a complete fractal universe
knowledge_node = {
"macro_view": "38.2% abstract overview", # Golden ratio macro
"meso_core": "Core knowledge content", # Current level
"micro_details": "61.8% deep details" # Golden ratio micro
}- Specialized agents with distinct roles working in harmony
- Dynamic weight adjustment optimizing agent participation based on feedback
- Knowledge fusion engine automatically integrating diverse perspectives
- Golden ratio storage structure: Each node layered by 0.618:1 ratio
- Self-similarity design: Any subgraph mirrors the overall structure
- Multi-resolution access: Seamless scaling from overview to details
User Query → Golden Ratio Decision → Fractal Retrieval → Multi-level Response
↑ ↓
Personal Knowledge Graph ← Confidence Update ← Learning Evaluation
import navi_learning
# Initialize Navi Assistant
assistant = navi_learning.NaviAssistant()
# Ask learning questions
response = assistant.ask("Explain the basic principles of quantum mechanics")
# Get layered responses
print(response.macro_overview) # Big picture understanding
print(response.core_concept) # Core knowledge
print(response.micro_details) # In-depth details# Enable fractal exploration mode
deep_dive = assistant.explore_fractal(
topic="Neural Networks",
resolution="micro" # Dive into microscopic level
)
# Get cosmic view of knowledge universe
cosmic_view = assistant.get_cosmic_pattern(
local_topic="Convolutional Neural Networks",
find_analogies=True # Discover cross-domain self-similarities
)- 🎯 Personalized Pathways: Adaptive learning routes based on fractal confidence
- 🔍 Deep Exploration: Access entire knowledge cosmos from any starting point
- 💡 Intuitive Understanding: Golden ratio optimized cognitive load distribution
- 📊 Learning Pattern Insights: Analyze learning behaviors through fractal dimensions
- 🎨 Curriculum Design: Knowledge structure planning based on self-similarity
- 🔄 Continuous Optimization: Dynamic confidence-adjusted teaching content
- Self-similar indexing: Knowledge retrieval predicting global from local patterns
- Golden ratio caching: Memory management optimized with 0.618 principle
- Multi-scale analysis: Seamless zooming from conceptual overview to technical details
# Golden ratio decision process
decision_flow = {
"learning_input": 0.618, # Learning agent weight
"critical_review": 0.382, # Critical thinking weight
"balanced_output": "Golden fused result"
}- Conceptual deep understanding: Building knowledge connections through fractal exploration
- Interdisciplinary learning: Discovering self-similar patterns across domains
- Research innovation: Original discoveries based on fractal thinking
- Skill tree construction: Golden ratio optimized learning paths
- Problem solving: Multi-scale analysis of complex business problems
- Innovative thinking: New solutions inspired by fractal patterns
We are building a learning system that truly understands the essence of knowledge, where:
- 🔄 Every learning session enriches the global knowledge fractal
- 🌍 Every user contributes to the evolution of collective intelligence
- 💫 Every question begins a cross-scale exploration journey
Begin Your Fractal Learning Journey - In Navi's world, every question is a universe, every answer contains infinite possibilities.
"True knowledge is not accumulation of information, but understanding of patterns" - Navi Design Philosophy
📖 [Detailed Documentation] | 🔧 [Development Guide] | 🚀 [Get Started] | 💬 [Community Discussion]
重大更新4：存储支持生态延伸
Tree-KG要解决的核心问题是：如何在科研、医疗、法律等知识密集型领域，快速构建高质量且可持续扩展的知识图谱？传统方法面临三大困境——知识复杂度高、人工标注成本巨大、知识更新速度快。实验数据显示，Tree-KG在多个数据集上的F1分数比第二名高出12-16%，在Text-Annotated数据集上达到0.81的F1分数，同时token使用量降低约40%。
项目名称：Tree-KG: An Expandable Knowledge Graph Construction Framework
开发团队：清华大学计算机系
第一作者：Songjie Niu, Kaisen Yang (共同一作)
通讯作者：Hongning Wang, Wenguang Chen
发表会议：ACL 2025 (Association for Computational Linguistics)
开源协议：MIT License
GitHub地址：https://github.com/thu-pacman/Tree-KG
论文链接https://aclanthology.org/2025.acl-long.907.pdf
技术支持：清华大学-鹏城实验室PAC-MAN团队
清华大学深圳国际研究生院
Major Update 4: Storage Support for Ecosystem Expansion
The core problem Tree-KG aims to solve is: How can we rapidly construct high-quality and sustainably scalable knowledge graphs in knowledge-intensive fields such as scientific research, healthcare, and law? Traditional methods face three major challenges: high complexity of knowledge, enormous costs of manual annotation, and the rapid pace of knowledge updates. Experimental results show that Tree-KG outperforms the second-best method by 12-16% in F1 score across multiple datasets, achieving an F1 score of 0.81 on the Text-Annotated dataset, while reducing token usage by approximately 40%.
Project Title: Tree-KG: An Expandable Knowledge Graph Construction Framework
Development Team: Department of Computer Science, Tsinghua University
First Authors: Songjie Niu, Kaisen Yang (Co-first authors)
Corresponding Authors: Hongning Wang, Wenguang Chen
Published Conference: ACL 2025 (Association for Computational Linguistics)
Open Source License: MIT License
GitHub Address: https://github.com/thu-pacman/Tree-KG
Paper Link: https://aclanthology.org/2025.acl-long.907.pdf
Technical Support: Tsinghua University - Peng Cheng Laboratory PAC-MAN Team
Shenzhen International Graduate School, Tsinghua University