Artificial_Intelligence_of_Things (AIoT)
Artificial Intelligence of Things (AIoT)
The Ouline
What’s AIoT?
AIoT(人工智能物联网)=AI(人工智能)+IoT(物联网)
The Artificial Intelligence of Things (AIoT) is the combination of artificial intelligence (AI) technologies with the Internet of Things (IoT) infrastructure to achieve more efficient IoT operations, improve human-machine interactions and enhance data management and analytics.
AIoT Hardware Architecture
This is the architecture of AIoT hardware. Here is a example of the specific equipment:SI-61S-AI Industrial AI Computer
The SI-61S-AI is a highly scalable player with an artificial intelligence calculation analysis system developed for multi-screen video wall applications.
It is equipped with Intel’s latest 7th generation desktop processor and supports multiple players via one Matrox or AMD PCI-E (x8) graphics card, and perform artificial intelligence analysis through another NVIDIA GPU card for AI analysis.
High recognition rate from AI analysis will provide on-site personnel with accurate data for better management and judgment.
IoT and Data
- Extension of IoT Applications(Consumer, industries, smart home, etc.)
- Data Explosion(爆炸)
- Artificial Intelligence of Things(IoT —> Big Data —> AI )
Data-driven AIoT
- Data Sources
- Data Processing and management
- Application-based analytics
- Application & sources
Data Processing and management
1. Data Collection
- Filtering
- Pre-processing
- Storing
- Ingestion(摄取)
2.1 Data Discovery
- Data Fusion(数据融合,集成多个数据源)
- Enrichment(使用模糊逻辑辅助搜索;获得相关的源地址信息;纠正拼写错误等)
- Modeling(建模)
- Indexing(标引)
- Annotation(标注)
2.2 Data Analytics
- Data Semantic(数据语义)
- Meta-data Management(元数据管理)
- Reasoning Contextual Data(上下文数据推理)
3. Data Protection
- Anonymize(匿名化)
- Data Masking(数据屏蔽,创建结构相似但不真实的数据版本,用以保护实际数据)
- Data Transformation(数据转换,从一种格式或结构转换为另一种格式或结构的过程)
- Sharing Control(共享控制)
###4. Data Publication
- Publication(发表)
- Open Data Sharing(开放的共享控制)
- Personal Data Sharing(个人数据共享)
- Visualization and Report(可视化报告)
AI-Based Deep Analysis
Decision Management(决策管理)
Discovery and Exploration(发现和探索)
Reporting and Analytics(报告和分析)
Predictive Analytics(预测分析)and Modelling
AI algorithms used in the field of data analysis
// TODO
Linear regression(线性回归)
举例:y = B0 + B1 * x 给定输入x,我们将预测y,线性回归学习算法的目标是找到系数B0和B1的值Logistic Regression(逻辑回归)
The prediction of the output is worth transforming using nonlinear functions called logical functionsBayes Theorem(朴素贝叶斯)
KNN
Support Vector Machine
Random Forest
…
Trustworthiness in AIoT
// TODO,此处关于区块链,后续文章补充
Secure Decentralized Artificial Intelligence of Things (AI + Blockchain + IoT)
![](https://image-host-toky.oss-cn-shanghai.aliyuncs.com/Screen Shot 2019-12-27 at 10.47.00 AM.png)