Pell provides ideas for artificial intelligence understanding cause and effect
November 29, 2021 11:04 Source: "Chinese Social Sciences", November 29, 2021, Issue 2297, Issue 2297

"Father of Bayesian Network"、Judea Pearl, a computer scientist and philosopher, is well -known for its basic contribution in the field of artificial intelligence。He proposed the probability and causal reasoning algorithm,Change the development direction of artificial intelligence based on rules and logic,and "made outstanding contributions in the field of artificial intelligence due to the" R & D probability and cause and effect algorithm "won the 2011 Turing Award。Pell has proposed many times in its research results in recent years,The current analysis of the cause and effect relationship of artificial intelligence,Inappropriately relied on big data -based probability statistics,It makes it difficult for artificial intelligence to achieve leapfrog development。For this,He proposed a new view of the theory of causality,Analysis of the relationship between causality for uncertainty,Develop the Bayesian network and structural causality model,and use the potential information and interference information existing in the analysis of the cause and effect analysis,It provides a new perspective for analysis of complex causality。

  bet365 live casino games

philosopher David Hume believes,Cause relationship is a human imagination,Humans summarize the relationship between the two often accompanied events through imagination as a causal relationship。Poore proposed,Cause and effect inferration ability is human fundamental ability,To solve the cause and effect password,Let artificial intelligence or robot have causal inference ability,It is necessary bet365 Play online games to conduct a systematic analysis of the consciousness of people's causal relationship。

Perl departs from human causal relationship consciousness,The concept of the ladder of causal relationship proposes。In his opinion,Human needs to use observation ability、Practical ability and imagination,and associated、Intervention and anti -factual reasoning can achieve causal inference。Observation ability is a collection of people、The ability to sort out information,It is also the basis of the regular understanding。Practical ability is to achieve the expected effect by adjusting the surrounding conditions、Create the ability to establish non -natural simple causal relationships。Imagination can help people know about something and their relationship with other things,Rate to a common、Unified cognition,helps to find the potential causal relationship between things。If a life body gets the above three abilities,This life body has the ability to understand causality and construct causality。

Perri uses the "ladder of cause and effect" to deconstruct the consciousness of causality。The first layer is associated,That is, the observer passed the observation,Looking for the law,Establish a causal relationship in your own consciousness。The second layer is intervention,That is, the observer will combine the observation with the active change,Regulate causal probability in consciousness。The third layer is anti -fact reasoning,At this level,Observer needs to build a false world,Assume that the conditions of a certain result do not exist,How can the relevant results change in this case,Verify whether the causal relationship is established。Perl discovered,People can easily analyze Bet365 app download uncertain causality through conscious activity,Digging potential information,Explore the interference factors in causality。

  Develop Bayesian network analysis causality

After a systematic analysis of the consciousness formation process of human causality,Perl discovered,Human beings can process the uncertain causality of the ladder processing of causality。Therefore,He will look at the traditional path of only analyzing the definition cause and effect relationship,Steering and uncertain causality treatment method for development,and proposed the Bayesian network theory model。Bayesian Network involves three core concepts in the development process,There is a ringless diagram、Condition probability table and Marcov conditions。

There is a circuit -free diagram with a different direction figure,Mostly used to show complex causality composed of multiple related events。Some parts have a tree -shaped structure in a loopless picture,Therefore, it is called a trees,Using this tree diagram can more intuitively analyze complex causality。such as,Analyze the development of a dynasty,Need to from politics、Economy、Culture、Environmental analysis,It also needs to show the degree of different elements that affect each other,At this time, you can use the directional diagram to analyze the cause and effect in it。The appearance of the direction without a ring picture makes it possible to effectively analyze the cause and effect in the same time as possible。

Conditional probability table refers to the probability table that compares the probability of multiple groups of related events for the occurrence.,Usually when three or more interconnected conditions appear,Need conditional probability table。such as,When the bet365 best casino games employee's employee arrives,,You need to consider multiple factors,The relationship between the interaction between these factors,At this time, you can use the condition probability table for overall analysis,Display the correlation between different factors。Condition probability table can analyze multiple multi -level causality in one -time,Improving the analysis efficiency of causal relationship。

Markov conditions are the relative independent elements related to main events,Analysis method of each element expression under certain circumstances。When an independent event is in the past、Now and the future may have a certain relationship with the main events,This seemingly independent incident has Markov status。Analyze the different states of independent events by setting different results,It can predict the development process of the entire event,That is, different independent events are expected to pay differently。At the same time, you can also draw the Malcov chain according to the Marcov process,​​visualization of complex causal analysis process。Through examples, we can better understand the role of Marcov's condition。If a writer needs to submit three novels in one month,It is necessary to ensure the quality of the work to receive the manuscript fee。other,He still has two part -time jobs to complete the same period。At this time, use Malcov conditions to analyze prediction,Can help him arrange time reasonably,Get the best results。Markov conditions provide dynamic analysis new dimensions for causal analysis,It provides a path for the optimization of dynamic cause and effect.。

  Use the structural causality model to describe causality

Although Bayesian network can play an important role in the analysis of Bet365 lotto review the cause and effect,But it cannot accurately explain the cause and effect,To improve the accuracy of the description of causality,Perl proposed a mathematical framework of causality -structural causality model。Using the structure causality model can test the complex causality,The main components of the structural causal model include the probability graph model、Structural equation model and disturbance model。

Probability graph model combines probability theory and chart theory,Can be correct、No direction、Abstract Relations for precise and efficient visualization description。If you want to design a national stadium,Designers must not only consider the basic function of this venue,Also consider the new features that may need to be added in the future,Even if these new functions are still in an unknown state。The causal relationship involved contains many uncertain factors and abstract problems,To solve such problems,Need to establish a probability graph model。The uniqueness of the probability graph model is,No matter how complicated the data and knowledge involved,The processing methods are the same。This optimized processing method integrates chart production and probability distribution,Suitable for different types of causality problems,Difficulty of analyzing the problem of complexity causality,Can improve the success rate of the analysis of causality。

Structural equation model is a method of modeling analysis of potential variable relationships,is a causal structural framework that can be used for multiple variable analysis。such as,Analysis IQ、EQ、When the relationship between the three of them,Usually it is difficult to directly measure quantitative calculation,"Estimation" bet365 best casino games in scale,Then then modeling analysis。Using a structured equation model can simultaneously tap the observed explicit variables and unspecified hidden variables。Visible,Structural equation model can dig out potential information,Analysis of theoretical analysis of the relationship between mathematics and cause and effect,provides a might for the deeper causal relationship for artificial intelligence understanding。

A motion model is a model that can analyze different degrees of interference factors in the process of causality.。Athletes will be affected by various factors,Among them, there are certain interference factor,There are also uncertain interference factors。If you want to avoid uncertain interference factors,You need to establish a disturbance model,Analysis of different types of possible、Different dimensions of interference factors,Based on this, formulate the optimized solution,To get the best results that people expect。

Probability Figure Model、Structural equation model and disturbance model jointly provide theoretical framework for the continuous iteration of artificial intelligence,It also provides different past paths for the study of causality in the new technology era。

The Perl system analyzed the process of human causal consciousness,Provides new ideas for the concept of cause and effect of artificial intelligence。Bayesian network and structural causality model becomes analytical uncertain causality、An important method of impact on the impact of potential information and complex interference factors。Perl's innovation of artificial intelligence research methods,Research path that promotes artificial intelligence changes from expert system research to deep learning research,It is important to further explore the cause and effect relationship。

(Author unit: Xiangtan University School of Marxism)

bet365 best casino games Editor in charge: Cui Cen
QR code icon 2.jpg
Key recommendation
The latest article
Graphics
bet365 live casino games
Video

Friendship link: Official website of the Chinese Academy of Social Sciences |

Website filing number: Jinggong.com Anxie 11010502030146 Ministry of Industry and Information Technology:

All rights reserved by China Social Sciences Magazine shall not be reprinted and used without permission

General Editor Email: zzszbj@126.com This website contact information: 010-85886809 Address: 11-12, Building 1, Building 1, No. 15, Guanghua Road, Chaoyang District, Beijing: 100026