参考笔记:https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E7%9B%AE%E5%BD%95.md
参考文档:https://github.com/jiaxiaogang/SMG_NOTE
第一梯队:1950年图灵提出"可思考的机器"和"图灵测试",他说:"放眼不远的将来,我们就有很多工作要做";
第二梯队:1956达特矛斯会议后,明斯基和麦卡锡等人穷其一生心血,虽未落地,却为AGI奠定了很多基础;
第三梯队:随着bigData,云计算等成熟,AI迎来DL热,但DL并不智能,故小僧希望通过此项目将AGI落地;
- 抽象
- 结构
- 无黑盒问题;
- 通用
- 可运行于单机终端
- 可解决当下NLP的瓶颈
- 不依赖巨量数据
- 不过度依赖算法
- Mind温和友善;
- 先天noData,需要像人类一样较长时间的学习与成长期;
- 意识并不神秘,只是这个"词汇"含盖了大量的含义;所以这个词汇是泛化了很多含义;要分解分布到系统中...
- 意识是开发出来的,不是电影和小说里的觉醒那么玄幻;
- "意识"的分解
- 意识流(AIAwarenessModel)
- 意识层(logThink)
- 意识行为
- 第6感
- 潜意识
- 点击查看相关笔记
-
Mind引擎
-
数据:
- 数据在Input时,从SMG架构图底层流向高层;
- 理解的过程其实就是数据的处理过程;
- logThink和noLogThink
- AIMemory AIAwareness AIMK AIMindValue AIMind...
- "性能塔"及"唯一性的性能优化"方式;
- 性能塔:(见数据流)底层流向高层;
- "唯一性的性能优化":避免了重复;性能良好;基于:类比、归纳与统计的理解系统:(反转替代优先原则)
- 唯一类比方式:
- 从可获取到的特征对比;(如夜间,根据声音)
- 从最显著的特征开始对比;(如杯子的颜色)
- 基于理解
只有基于理解的NLP才有可能达到人类同等水平;
- 基于:类比、归纳与统计的理解系统;
- 先天noData;
- 基于能力开发;
- 后天学习
- 灵活性
- 动态继承
- 动态属性
- 动态接口等
- 基于data的方式使其灵活,通用;
- 神经网络
- AILine(Stong等)
- AIPointer
- LightArea
- 纵向点亮
- 横向点亮
- 想像力
想像力:组合颜色,材质等,铺到模型上,GAN;
- 创造力
参考:(见笔记:AI/框架/创造力)
- 预测
基于常识CommonSence
- 知识表示
数据拆分(在Memory->AIAwareness的过程中,自然而然的实现拆分与抽象;)
参考:
- 重绘了新版架构图;
- AIFoundation
- 颠覆现有的编程方式
- 知识表示(已解决)
- 三维架构(参考笔记/AI/框架)
- BrainTree(参考N3P7,N3P8)
- ThinkTask(参考N3P11,Awareness->Demand->ThinkTask)
注:
1. 大家有交流需求或问题请直接Issues我;
2. 在Issues,我知无不言,言无不尽,回复及时,用词流利!
3. (QQ/微信七天回复一次)
Reference notes:https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E7%9B%AE%E5%BD%95.md
Reference documentation:https://github.com/jiaxiaogang/SMG_NOTE
# I have a communication needs or questions please directly I am, in the Issues, I know nothing, words endless, reply in time, words fluent! (QQ / micro letter seven days to reply once)
The first echelon: 1950 Turing put forward the "think of the machine" and "Turing test", he said: "look forward to the near future, we have a lot of work to do";
Second echelon: 1956 Dartmouth meeting, Minsky and McCarthy and others poor life of their own efforts, although not landing, but for the AGI laid a lot of foundation;
The third echelon: With bigData, cloud computing and other mature, AI ushered in the DL heat, but DL is not smart, so the monk hopes to AGI landing through this project;
- abstract
- structure
- no black box problem;
- Universal
- can run on a stand-alone terminal
- can solve the bottleneck of the current NLP
- do not rely on huge amounts of data
- no overridden algorithm
- gentle and friendly;
- [LOP programming ideas] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0/Note4. md # n4p13loplayer-oriented-programming170803)
- [logThink and noLogThink] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%A1%86%E6%9E%B6/%E8%87%AA%E6%88%91.md # Two operations)
- [based on understanding NLP] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/Light%E5%8F%8A%E5%BA%94%E7%94%A8/Light.md#nlp language Be sure to describe the meaning of the core to achieve 100 correct rate)
- [deep understanding and shallow understanding] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%A1%86%E6%9E%B6/%E6%B3%A8%E6%84%8F % E5% 8A% 9B.md)
- [three-dimensional abstract architecture] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%A1%86%E6%9E%B6/3%E7%BB%B4%E6%8A%BD% E8% B1% A1% E6% A6% 82% E8% BF% B0.md)
- [tree knowledge representation] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0/Note3 .md # n3p7 data tree theory)
- [analogy, induction and statistics] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0/ Note5.md # n5p2 intelligent neural network)
- abstraction and structure
- No reliance on algorithms and large data
- congenital noData
- and many more
- naive noData, need to be as long as human learning and growth period;
- consciousness is not mysterious, but this "vocabulary" contains a lot of meaning; so the vocabulary is generalized a lot of meaning; to decompose distributed to the system ...
- consciousness is developed, not the film and the awakening of the novel so fantasy;
- the decomposition of "consciousness"
- A stream of consciousness (AIAwarenessModel)
- consciousness layer (logThink)
- Conscious behavior
- the first six sense
- Subconscious
- [click to view related notes] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0/Note4 .md)
-
Mind engine
-
Data:
- data in the Input, from the bottom of the SMG structure to the high level;
- the process of understanding is actually the process of data processing;
- logThink and noLogThink
- AIMemory AIAwareness AIMK AIMindValue AIMind ...
- "performance tower" and "unique performance optimization" approach;
- performance tower: (see data flow) the bottom of the flow to the top;
- "Unique performance optimization": avoid duplication; good performance; based on: analogy, induction and statistical understanding system: (reverse substitution priority principle)
- the only analogy:
- from the available feature contrast; (such as night, according to the sound)
- from the most notable features began to contrast; (such as the color of the cup)
- Based on understanding
Only NLP based on understanding is possible to achieve the same level of human beings;
- Based on: analogy, induction and statistical understanding system;
- naive noData;
- Capacity-based development;
- acquired learning
- Flexibility
- Dynamic inheritance
- Dynamic properties
- dynamic interface and so on
- based on the data to make it flexible, generic;
- Neural network
- AILine (Stong et al)
- AIPointer
- [LightArea] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0/Note4.md # n4p17lightarea)
- vertical light
- Lights up
Imagination
Imagination: combination of color, material, etc., spread to the model, GAN;
- Creativity
Reference: (see note: AI / frame / creativity)
-
Forecast Based common sense
-
Knowledge representation
Data split (in the Memory-> AIAwareness process, the natural implementation of the split and abstract;) Reference: [https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%A1%86%E6%9E%B6/%E7%9F%A5%E8%AF%86%E8%A1%A8 % E7% A4% BA.md] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%A1%86%E6%9E%B6/%E7%9F%A5%E8%AF% 86% E8% A1% A8% E7% A4% BA.md)
- redraw the new version of the composition;
- AIFoundation
- subversion of existing programming methods
- knowledge representation (resolved)
- 3D architecture (reference notes / AI / framework)
- BrainTree (see N3P7, N3P8)
- ThinkTask (refer to [N3P11] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0/ Note3.md # n3p11awareness-demand-thinktask --- or old info demand), [Awareness-> Demand-> ThinkTask] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%A1%86 % E6% 9E% B6 / Understand.md # awareness-demand-thinktask task))
- [awareness] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0/Note3.md # n3p15 awareness 170801)
- [thinking process finishing] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0/Note4 .md # n4p2 thinking evolution of the essence of ai thinking 170714)
- [LightArea-area lit] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0 /Note4.md#n4p17lightarea)
- [GeneralFeel - general feeling] (https://github.com/jiaxiaogang/SMG_NOTE/blob/master/%E6%89%8B%E5%86%99%E7%AC%94%E8%AE%B0 /Note4.md#n4p18 general feeling)
Neural network
- [AILine] (https: // gi