jiaxiaogang / He4o
Licence: gpl-3.0
和（he for objective-c） —— “信息熵减机系统”
Stars: ✭ 284
Labels
Projects that are alternatives of or similar to He4o
Flappy Es
Flappy Bird AI using Evolution Strategies
Stars: ✭ 140 (-50.7%)
Mutual labels: artificial-intelligence, reinforcement-learning, unsupervised-learning
Complete Life Cycle Of A Data Science Project
Complete-Life-Cycle-of-a-Data-Science-Project
Stars: ✭ 140 (-50.7%)
Mutual labels: reinforcement-learning, unsupervised-learning, transfer-learning
Free Ai Resources
🚀 FREE AI Resources - 🎓 Courses, 👷 Jobs, 📝 Blogs, 🔬 AI Research, and many more - for everyone!
Stars: ✭ 192 (-32.39%)
Mutual labels: artificial-intelligence, reinforcement-learning, unsupervised-learning
Awesome Artificial Intelligence
A curated list of Artificial Intelligence (AI) courses, books, video lectures and papers.
Stars: ✭ 6,516 (+2194.37%)
Mutual labels: artificial-intelligence, reinforcement-learning, unsupervised-learning
L2c
Learning to Cluster. A deep clustering strategy.
Stars: ✭ 262 (-7.75%)
Mutual labels: artificial-intelligence, unsupervised-learning, transfer-learning
Transfer Learning Suite
Transfer Learning Suite in Keras. Perform transfer learning using any built-in Keras image classification model easily!
Stars: ✭ 212 (-25.35%)
Mutual labels: artificial-intelligence, transfer-learning
Evostra
A fast Evolution Strategy implementation in Python
Stars: ✭ 227 (-20.07%)
Mutual labels: artificial-intelligence, reinforcement-learning
Deeppicar
Deep Learning Autonomous Car based on Raspberry Pi, SunFounder PiCar-V Kit, TensorFlow, and Google's EdgeTPU Co-Processor
Stars: ✭ 242 (-14.79%)
Mutual labels: artificial-intelligence, transfer-learning
temporal-ssl
Video Representation Learning by Recognizing Temporal Transformations. In ECCV, 2020.
Stars: ✭ 46 (-83.8%)
Mutual labels: transfer-learning, unsupervised-learning
Dm control
DeepMind's software stack for physics-based simulation and Reinforcement Learning environments, using MuJoCo.
Stars: ✭ 2,592 (+812.68%)
Mutual labels: artificial-intelligence, reinforcement-learning
Change Detection Review
A review of change detection methods, including codes and open data sets for deep learning. From paper: change detection based on artificial intelligence: state-of-the-art and challenges.
Stars: ✭ 248 (-12.68%)
Mutual labels: artificial-intelligence, unsupervised-learning
Revisiting-Contrastive-SSL
Revisiting Contrastive Methods for Unsupervised Learning of Visual Representations. [NeurIPS 2021]
Stars: ✭ 81 (-71.48%)
Mutual labels: transfer-learning, unsupervised-learning
Gan Sandbox
Vanilla GAN implemented on top of keras/tensorflow enabling rapid experimentation & research. Branches correspond to implementations of stable GAN variations (i.e. ACGan, InfoGAN) and other promising variations of GANs like conditional and Wasserstein.
Stars: ✭ 210 (-26.06%)
Mutual labels: artificial-intelligence, unsupervised-learning
Face.evolve.pytorch
🔥🔥High-Performance Face Recognition Library on PaddlePaddle & PyTorch🔥🔥
Stars: ✭ 2,719 (+857.39%)
Mutual labels: artificial-intelligence, transfer-learning
Applied Reinforcement Learning
Reinforcement Learning and Decision Making tutorials explained at an intuitive level and with Jupyter Notebooks
Stars: ✭ 229 (-19.37%)
Mutual labels: artificial-intelligence, reinforcement-learning
Amazing Machine Learning Opensource 2019
Amazing Machine Learning Open Source Tools and Projects for the Past Year (v.2019)
Stars: ✭ 198 (-30.28%)
Mutual labels: artificial-intelligence, reinforcement-learning
TA3N
[ICCV 2019 Oral] TA3N: https://github.com/cmhungsteve/TA3N (Most updated repo)
Stars: ✭ 45 (-84.15%)
Mutual labels: transfer-learning, unsupervised-learning
KD3A
Here is the official implementation of the model KD3A in paper "KD3A: Unsupervised Multi-Source Decentralized Domain Adaptation via Knowledge Distillation".
Stars: ✭ 63 (-77.82%)
Mutual labels: transfer-learning, unsupervised-learning
awesome-contrastive-self-supervised-learning
A comprehensive list of awesome contrastive self-supervised learning papers.
Stars: ✭ 748 (+163.38%)
Mutual labels: transfer-learning, unsupervised-learning
Polyaxon
Machine Learning Platform for Kubernetes (MLOps tools for experimentation and automation)
Stars: ✭ 2,966 (+944.37%)
Mutual labels: artificial-intelligence, reinforcement-learning
he4o系统
he4o是一个信息熵减机系统。
关键字:
- 支持迁移学习为主，强化学习为辅。
- 完善的知识表征：
稀疏码、特征、概念、时序、价值。- 神经网络支持：动态、模糊,终身动态学习，知识网络遗传迭代。
- 支持智能体自发动态常识获取问题（定义问题）。
- 无论宏观框架还是微观细节设计,都依从相对与循环转化。
- 支持感性与理性的递归决策。
- 认知学习、决策行为、反思反省。
- 数理：集合论。
- 计算：使用最简单的bool运算：
类比和评价。手稿：https://github.com/jiaxiaogang/HELIX_THEORY
网站：http://www.jiaxiaogang.cn
1. -------------引言-------------
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，明斯基和麦卡锡等等许多前辈穷其一生心血，虽然符号主义AI在面对不确定性环境下鲁棒性差，但却为AGI奠定了很多基础。
第三梯队:随着大数据,云计算等成熟,AI迎来DL热,但DL也并非全功能型智能体。
综上：近70年历史，人工智能研究经历跌宕起伏，但终是方兴未艾，he4o旨在以熵减机方式解决这一问题。
2. -------------目录-------------
3. -------------（一）熵减机-------------
熵减机从2017年2月正式开始研究至2018年2月成熟，历时一年。
 |
|---|
| 熵减机：其包含三大要素，分别为： 定义、相对和循环。 |
| https://github.com/jiaxiaogang/HELIX_THEORY#%E7%86%B5%E5%87%8F%E6%9C%BA |
4. -------------（二）信息熵减机模型-------------
信息熵减机理论模型在18年3月成熟，直至今天此模型仍在不断细化中。
 |
|---|
| 1. 从外到内,从内到外的双向,分别为:从动到静,从静到动（客观角度）。 |
| 2. 每外一个模块,与内所有模块之和相对循环 (如神经网络与思维,智能体与现实世界) |
| 注: 一切都是从无到有,相对与循环; |
注: he4o认为自己活着 源于循环; |
5. -------------（三）he4o系统实践-------------
V1.0《初版》：
2017年2月立项 －2018年10月21日正式落地发布V1.0版本。
V2.0《小鸟生存演示》：
2018年11月－至今开发完成，第三轮测试训练中...
| 架构图 |  |
|---|---|
| 性能要求 | 可运行于单机终端 |
| 编程思想 | DOP (面向动态编程) |
| 架构设计 | 由熵减机理论展开成信息熵减机模型,再由信息熵减机模型展开为系统架构 碰巧与大脑多有相似之处^_^
|
| 代码占比 | 内核代码中神经网络占30％,思维控制器占50%,其它（输入、输出等）共占20%; |
| 神经网络 | 神经网络的模型十字总结:横向组与分,纵向抽具象; |
| 思维控制器 | 由向性规则决定，每一种方向操作代表一种思维操作。 |
6. -------------时间线-------------
2021.03.12
至今
- v2.0八测与训练：
防撞训练、R-模式测试
2021.03.03
耗时8天
- R-决策模式更理性迭代：
弄巧成拙BUG：静默成功
2021.01.30
耗时4天
- R-决策模式V3迭代、反向反馈外类比
2021.01.23
耗时35天
- v2.0七测与训练
防撞训练、R-模式测试
2021.01.15
耗时8天
- In反省类比迭代、R-决策模式V2迭代
迭代触发机制: 生物钟触发器
2020.12.24
耗时20天
- v2.0六测与训练
多向飞行正常
2020.12.07
耗时1个月
- AIScore评价器整理完善：
时序理性评价：FRS、稀疏码理性评价：VRS
2020.11.07
耗时1个月
- v2.0五测与训练
2020.10.21
耗时15天
- TIR_Alg支持多识别
2020.09.01
耗时1个月
- v2.0四测与训练
2020.08.12
耗时27天
- Out反省类比迭代 (DiffAnalogy)、生物钟(AITime)、PM理性评价迭代v2
2020.06.28
5天
- 决策迭代：PM理性评价
2020.06.06
耗时2个月
- v2.0三测与训练
2020.05.15
耗时20天
- 决策迭代：（根据
输出期短时记忆使决策递归与外循环更好协作）
2020.04.21
耗时1个月
- 决策迭代：（根据
输入期短时记忆使决策支持四模式）
2020.03.31
耗时1个月
- 迭代外类比: 新增反向反馈类比 (In反省类比) (构建SP正负时序、应用SP于决策的MC中、迭代反思)
2020.02.20
耗时18天
- 稀疏码模糊匹配
2019.12.27
持续3个月
- v2.0二测与规划性训练－－回归小鸟训练
2019.11.22
耗时1个月
- 理性思维——反思评价
2019.09.30
耗时2个月
- 理性思维——TOR迭代 (行为化架构迭代、支持瞬时网络)
2019.08.25
耗时1个月
- 理性思维——TIR迭代 (时序识别、时序预测、价值预判)
2019.06.20
耗时2个月
- v2.0版本基础测试改BUG 与 训练
2019.06.05
写完耗时15天,调至可用性达到标准至45天
- v2.0一测－－小鸟训练——神经网络可视化v2.0
2019.05.01
耗时1个月
- 优化性能——
XGWedis异步持久化和短时内存网络
2019.03.01
耗时2个月
- 内类比 （与外类比相对）
2019.01.21
耗时40天
- 迭代决策循环 （行为化等）
2018.11.28
耗时2个月
- 迭代神经网络 (区分动态时序与静态概念)
2018.11.05
规划耗时20天
- 势 (小鸟生存演示) (v2.0开始开发)
2018.10.21
耗时0天
- v1.0.0 (he4o内核发布)
2018.10.20
耗时0天
- 信息熵减机 (产生智能的环境)
2018.08.29
耗时2个月
- MOL
2018.08.01
耗时1个月
- MIL & MOL (重构中层动循环)
2018.07.01
耗时1个月
- HELIX (定义、相对和循环呈现的螺旋型)
2018.06.01
耗时1个月
- 三层循环大改版 (mv循环,思维网络循环,智能体与现实世界循环)
2018.05.01
耗时1个月
- 相对 (he4o实现定义,横向相对,纵向相对)
2018.02.01
耗时3个月
- 宏微 (前身是拆分与整合,宏微一体)
2017.12.09
耗时2个月
- 定义 (从0到1)
2017.11.10
耗时1个月
- 规则 (最简)
2017.09.20
耗时50天
- DOP_面向数据编程
- GNOP_动态构建网络
2017.08.23
耗时1个月
- 神经网络 (算法,抽具象网络)
2017.08.02
耗时20天
- MindValue（价值）
2017.07.10
耗时20天
- 树BrainTree(参考N3P7,N3P8)
2017.06.01
耗时40天
- 三维架构(参考笔记/AI/框架)
2017.05.22
耗时10天
- OOP编程思想->数据语言 (OOP2DataLanguage)
2017.05.21
耗时1天
- 重绘了新版架构图; (AIFoundation)
2017.04.21
耗时1个月
- 金字塔架构
2017.03.21
耗时1个月
- 分层架构
2017.02.21
耗时1个月
- 流程架构
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at [email protected].