News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
The five major schools of AI machine learning
The five major schools of AI machine learning
The way is "Rome wasn't built in a day". The development of machine learning has gone through a long time. In the process, there are five schools. The five schools have their own characteristics.
1, Symbolism (Symbolists)
Name: Symbolism (Symbolists)
Origin: logic, philosophy
The core idea: cognition is calculation, and the result is predicted by the deduction and inverse deduction of symbols.
Problem: knowledge structure
Representative algorithm: inverse deductive algorithm (Inverse deduction)
Representation application: Knowledge Map
Representative figures: Tom Mitchell, Steve Muggleton, Ross Quinlan
2, the Bias group (Bayesians)
Name: Bayes (Bayesians)
Origin: Statistics
Core ideas: subjective probability estimation, occurrence probability correction, optimal decision
Problem: uncertainty
Representative algorithm: probability reasoning (Probabilistic inference)
Representative application: anti spam, probability prediction
Representative figures: David Heckerman, Judea Pearl, Michael Jordan
3, connectionism (Connectionist)
Name: connectionism (Connectionist)
Origin: Neuroscience
Core idea: Simulation of the brain
Problem: reliability distribution
Representative algorithms: back propagation algorithm (Backpropagation), depth learning (Deep learning)
Representative applications: machine vision, speech recognition
Representative figures: Yann LeCun, Geoff Hinton, Yoshua Bengio
4. Evolutionism (Evolutionaries)
Name: evolutionism (Evolutionaries)
Origin: evolutionary biology
Core ideas: Simulation of evolution, genetic algorithms and genetic programming
Problem: structural discovery
Representative algorithm: gene programming (Genetic programming)
Representative application: starfish robot
Representative figures: John Koda, John Holland, Hod Lipson
5, behavioral analogism (Analogizer)
Name: behavioral analogism (Analogizer)
Origin: Psychology
The core idea: the similarity between the old and the new knowledge
Problem: similarity
Representative algorithms: Kernel machines, near neighbor algorithm (Nearest Neightor)
Representative application: Netflix recommendation system
Representative figures: Peter Hart, Vladimir Vapnik, Douglas Hofstadter
1, Symbolism (Symbolists)
Name: Symbolism (Symbolists)
Origin: logic, philosophy
The core idea: cognition is calculation, and the result is predicted by the deduction and inverse deduction of symbols.
Problem: knowledge structure
Representative algorithm: inverse deductive algorithm (Inverse deduction)
Representation application: Knowledge Map
Representative figures: Tom Mitchell, Steve Muggleton, Ross Quinlan
2, the Bias group (Bayesians)
Name: Bayes (Bayesians)
Origin: Statistics
Core ideas: subjective probability estimation, occurrence probability correction, optimal decision
Problem: uncertainty
Representative algorithm: probability reasoning (Probabilistic inference)
Representative application: anti spam, probability prediction
Representative figures: David Heckerman, Judea Pearl, Michael Jordan
3, connectionism (Connectionist)
Name: connectionism (Connectionist)
Origin: Neuroscience
Core idea: Simulation of the brain
Problem: reliability distribution
Representative algorithms: back propagation algorithm (Backpropagation), depth learning (Deep learning)
Representative applications: machine vision, speech recognition
Representative figures: Yann LeCun, Geoff Hinton, Yoshua Bengio
4. Evolutionism (Evolutionaries)
Name: evolutionism (Evolutionaries)
Origin: evolutionary biology
Core ideas: Simulation of evolution, genetic algorithms and genetic programming
Problem: structural discovery
Representative algorithm: gene programming (Genetic programming)
Representative application: starfish robot
Representative figures: John Koda, John Holland, Hod Lipson
5, behavioral analogism (Analogizer)
Name: behavioral analogism (Analogizer)
Origin: Psychology
The core idea: the similarity between the old and the new knowledge
Problem: similarity
Representative algorithms: Kernel machines, near neighbor algorithm (Nearest Neightor)
Representative application: Netflix recommendation system
Representative figures: Peter Hart, Vladimir Vapnik, Douglas Hofstadter