Relationship between AI, Machine Learning and Deep Learning
- Author: Damon Yuan
- Date: 2019-04-15
- Reference: 算法工程师眼中的AI岗位
AI
Artificial Intelligence Tripleism
Symbolicism
Knowledge Map
NLP
Evolutionism / Cyberneticsism
Reinforcement Learning
Automation Control
Connectionism
Neural Networks
2 Low tide or 3 high tide
- Perceptron, 1950s, High
- Linear perceptron function is limited, 1968, Low
- Back-propagation Algorithm, 1986, High
- SVM Algorithm, 1995, Low
- Multilayer neural network structure with AutoEncoder, 2006, High
Machine Learning = Algorithm + Neural Networks
Let the machine behave like the intelligent behavior that people show.
ID3 Decision tree algorithm, 1986
Regression & Clustering -> VC Dimension
SVM & RandomForest
Three Modes of Machine Learning
Supervised Learning
Unsupervised Learning
Reinforcement Learning
Machine Learning Directions
Machine Learning Scientist
Research Machine Learning Algorithm
Machine Learning Engineering
Apply Machine Learning Algorithm to current business.
Machine Learning Platform
Build and Maintain Distributed Machine Learning Platform
Deep Learning = AutoEncoder + Multilayer neural network structure
Data Mining vs Machine Learning
Recommended System and Advertisement System
Machine Learning -> CTR Prediction algorithm
Data Mining -> Mining new characteristics in data
Searching Engine
Machine Learning is relatively less important.
Crawler Sub-system
Information Index Sub-system
Association Rules
NLP, Computer Visual and Sound Recognization
Big Data vs Machine Learning
4V
Volume
Velocity
Variety
Value
3 Factors
Hadoop is created based on it, and Spark is based on it.
Google FS
MapReduce
Bigtable
Data Analysis
Focus in human's subjective initiative to analysis the data.
A funny history roadmap of Hadoop
