Research On Dreyfus's Philosophy Of Artificial Intelligence

In 1960 s,the early Symbolic Artificial Intelligence,as the representative AI,has made a rapid development as well as a series of spectacular achievements.Phenomenologist Dreyfus questioned the Artificial Intelligence research at that time.From the unique perspective of phenomenology,the early AI was analyzed and criticized.Based on this,Dreyfus' s philosophy of AI was forming in the progress.This thesis systematically investigates the formation process,the basic characteristics,and the phenomenon ideas of Dreyfus' s AI philosophy.It focuses on his critique of traditional Artificial Intelligence,and reveals its theoretical value in the research field of artificial intelligence and philosophy and practical significance.First of all,phenomenologist Dreyfus revealed that there are two insurmountable obstacles to AI: commonsense problem and knowledge representation.Besides,he found that the Symbolic AI which is actually based on the theory of three unproven philosophical assumptions.Therefore,Dreyfus predicted that the Symbolic AI would certainly fail.At the same time,he put forward that the real AI should be embedded-embodied.In other words,intelligence is closely related to the body.Secondly,in order to solve the knowledge problem of the early AI,Dreyfus made a creative interpretation of phenomenology and introduced it into the study of AI.He mainly misappropriated Husserl's theory of intentionality,Heidegger's existentialism and Merleru-Ponty's embodied theory,demonstrating the outstanding advantages of phenomenology in solving the problem of representation.Finally,after Dreyfus' s criticism and reflection on the early AI,he tried to put forward the idea of the seven stage skill acquisition model to solve the knowledge dilemma of AI.The philosophy of AI created by Dreyfus has a great influence on the study at that time.On one hand,the researchers of the Symbolic AI have gradually realized the direction of their research and the fatal loopholes on the way to reduce the loss.On the other hand,the solution of skill acquisition model brings inspiration for subsequent researchers.Scientists began to pay attention to the connection between intelligence and body and emphasized the understanding and grasp of intelligence to the whole or context.When contemporary philosophy and science getting far away,Dreyfus has successfully established the deep connection between AI and phenomenology by exploring the “crossover” of AI.Under his influence,AI researchers began to reflect on the philosophical basis of AI theory,and philosophers began to focus on the development of AI technology.Unfortunately,Dreyfus' s AI research also has weaknesses.Firstly,Dreyfus has a pessimistic view of science and technology,so as to the search for Connectionism of specific problems.However,the Connectionism represented by machine learning is obviously developing well now.Here,Dreyfus seems to confuse the path of Symbolism and Connectionism.The former is a top-down way of general intelligence,and the latter is the path of specific intelligence from bottom to top.Similarly,he underestimated the rapid development of computer technology and neurobiology.The deep learning based on neural network simulation has made amazing achievements in a specific field.Secondly,the philosophy of AI created by Dreyfus still has a flaw in Husserl's embodied theory.He thought Husserl had a vague attitude towards the embodied problem.But in Husserl's “return to the real thing itself”,he clearly emphasized the intuitive feeling that I came back to the world,and the body was present in every perceptual experience.He ignored Merleru-Ponty's inheritance and development of Husserl's view of the body.Finally,in Dreyfus' s model of Skill Acquisition,senior technical experts not only require learning beyond the context of the intuitive body reaction of the unconscious,but also emphasizes that the learners need to deepen skills in the overall environment,thus resulting in hermeneutic circulation problems.