Machine Learning Artificial Intelligence Natural Language Processing Semantic Web Ontology Knowledge Information Processing Digital Transformation Probabilistic Generative Model Deep Learning Navigation of this blog
Free will and soft determinism
The logician Kazuhisa Todayama, in his book ‘What is Meaning (1): An Introduction to Philosophy’, also discusses ‘freedom’ in the last chapter of his book ‘An Introduction to Philosophy’.
In it, he says: ‘Stones and frogs do not have freedom, but humans do. And that freedom is valuable. Thus we believe and do not doubt. And so, in order to gain more freedom, to fight against those who would take it away, we go around at night in school buildings, smashing windows and shooting invisible guns. But in the process of doing so, you suddenly realise that what you are looking for is ‘invisible freedom’, and you start to cry out ‘what the hell is freedom’ (sorry for the retrospective mode). Are we really free, and why is freedom valuable?’ He stated.
Christianity, the Bible and Related Writings’ also states: ’The whole course of the world is predetermined by God, who created it. Above all, it has already been decided who will go to heaven and who will not.’ This theological determinism is called,
In the world of mechanics, starting with Newton, who found the familiar Newton’s Law in machine learning, which is also described in ‘Overview of Newton’s Law and its algorithm and implementation’, everything in this world is also considered to be a deterministic world that operates according to certain laws.
Instead of this total determinism, in which the whole world is predetermined, there is also ‘mechanistic determinism’ or ‘machine theory’, in which the output is determined only by environmental inputs and internal mechanisms, as in a computer, limited to a certain area.’ as described in ‘Invitation to Cognitive science ‘Reading notes’, is based on this ‘mechanism determinism’ approach and states that ‘I am a mechanism that moves based on stimuli from the external world and my internal state. What appears to be ‘what I did’ is really just ‘what happened to me’.’ This is defined as.
Three schools of thought exist on this determinism
- Libertarianism: ‘Determinism is wrong, at least with regard to human action. We are not mere physical systems or computational mechanisms. We are more than that and therefore have freedom’.
- Hard determinism: ‘The existence of freedom is wrong. We are a deterministic system and therefore have no freedom. Free will is an illusion.’ This is called.
- Compatibilism (or soft determinism): ‘Determinism and the existence of freedom are compatible. We are both deterministic systems and bearers of freedom’.The book that compiles papers on these arguments is Free Will: Skinner, Dennett and Ribet.
Among these are the classic work of B. F. Skinner, who made challenging arguments against the concept of ‘freedom’ from the standpoint of behaviourist psychology, and R. M. . who set out the fundamental composition of the consideration as a philosophical problem. Chisam’s article, the seminal work of the outstanding contemporary compatibilist D. C. Dennett, and the two controversies that B. Ribet had with P. Churchland and J. Searle over his famous experiment are described.
In Dennett’s compatibilism in these, it is stated that being ‘determined’ means that the event is not ‘inevitable’, but that it ‘can be done otherwise’ and is the ‘freedom’ it deserves to have. The ability to reconfigure programming to make such free decisions is stated to be a freedom that humans have.
When considering these ideas in terms of the use of artificial intelligence technology, it is stated that a machine should be able to derive options that ‘can be done otherwise’ beyond the possible options of a human being, and among these, not simply algorithms that can be realised by a machine, but algorithms that are based on ‘causal reasoning and consideration towards realising a strong AI’, as stated in the section ‘Considerations towards realising causal reasoning and a strong AI’. If problems can be solved by algorithms based on deep imagination and models based on that imagination, then humans will be able to play a role that machines cannot play.
The abilities required of humans in the age of artificial intelligence are imagination and logical thinking based on that imagination, which can be acquired by gaining a lot of experience.
Zhuangzi’s theory of freedom and compatibility
‘The Thought of Zhuangzi: How can the mind be free?’, Zhuangzi, an ancient Chinese philosopher, developed his own perspective and philosophy on this freedom as follows
- Relativism and pluralism: Zhuangzi’s thought developed from the standpoint of relativism and pluralism, and considered that all things are interrelated and relative. He recognised that reality and values differ from individual to individual and culture to culture and the diversity of the world as seen from different perspectives and standpoints.
- Relationship between nature and man: Zhuangzi also provides important insights into the relationship between man and nature. He argues that nature has the capacity for self-regulation and that it is important to leave it in its natural state. He also believed that humans are part of nature and are expected to live according to the natural order.
- Freedom of mind and enlightenment: Zhuangzi pursued freedom of the human mind. He taught that it is important to realise one’s true self through free thought and action, unbound by external constraints and norms. He also believed that through enlightenment and rushing thoughts (kuken jisang), human beings could open their wisdom and discover the truth.
- Emphasis on play and enjoyment: Zhuangzi stressed the importance of play and enjoyment. He argued that humans should play, enjoy and pursue a rich life in their natural state. Play means to be free from bondage and toil, and to gain comfort and satisfaction through free expression and creation in accordance with one’s own nature.
At first glance, Zhuangzi’s thought seems to be a form of libertarianism, but as stated in ‘Human Creativity and Symbiosis with AI – The Unconscious and Memory’, there are various mechanisms in the back of human thought that are in line with memory, and humans cannot make completely free choices.
Even if we were to leave this to some kind of random choice, it would be difficult to create complete randomness, and we would have to create predictable randomness based on various probability distribution models, such as those described in ‘Various probability distributions used in probabilistic generative models’, which is a far cry from free choice. This is a far cry from free choice.
Zhuangzi freedom can also be thought of as based on soft determinism, which is not bound by external constraints or norms, but rather re-programming free decision-making.
This kind of thinking is in the Taoist tradition and leads to the Zen philosophy described in ‘Zen, metacognition and AI’.
コメント