3

Human beings are capable of deciding upon rules based on intuitions and observations their neurons presumably provide (certainly metaphysical presumptuous). According to WP, this is inductive reasoning:

Inductive reasoning is a method of reasoning in which a general principle is derived from a body of observations. It consists of making broad generalizations based on specific observations.

Computers are capable of deciding upon rules based on weighted connection models of computation based on data sets where connection models are modeled on actual neurons and data sets might be data from visual processing systems designed around cameras. These systems are called rule-based machine learning systems:

Rule-based machine learning (RBML) is a term in computer science intended to encompass any machine learning method that identifies, learns, or evolves 'rules' to store, manipulate or apply... rule-based machine learning applies some form of learning algorithm to automatically identify useful rules, rather than a human needing to apply prior domain knowledge to manually construct rules and curate a rule set.

The question is simple. Does rule-based machine learning qualify as inductive reasoning?

J D
  • 19,541
  • 3
  • 18
  • 83
  • 1
    The issue with "modeled on actual neuron" is that the model is ... just a model, may be very crude, may not entirely reflect what a real neuron does in a brain, or could also be somewhat wrong. The model may not be accurate at all. I regularly see articles from neuroscience about new and important mechanisms in real neurons that are nowhere near modeled in artificial neurons... – Frank May 16 '23 at 18:26
  • @Frank Indeed. Sounds like you might advocate [all models are wrong](https://en.wikipedia.org/wiki/All_models_are_wrong). – J D May 16 '23 at 18:49
  • 2
    I think it's something to keep in mind before getting into whether the machines are conscious or other similar "fascinating" topics. – Frank May 16 '23 at 18:53
  • 3
    in terms of answering the question itself, the whole ML, which is just statistics with better computers is "inductive". – Frank May 16 '23 at 20:03
  • @Frank You're an ML guy, so I place some faith in that. But why the [scare quotes](https://en.wikipedia.org/wiki/Scare_quotes)? – J D May 16 '23 at 20:18
  • I was not sure if "inductive" as in "based on data" would completely correspond go "inductive logic" even with your second paragraph. – Frank May 16 '23 at 20:37
  • Yes; based on the set od data provided, the machine applies an algorithm (that from the point of view of the machine is a sort of *a priori*) to formulate and test rules that act as general laws. – Mauro ALLEGRANZA May 17 '23 at 05:41
  • There are many ways ta handle a storm and we're in what someone I knew once described as *the coming* **AI storm**. Landfall made early 2020s I suppose. I like the crispiness of the query - sharp!! Makes me wanna dive, but all I see is non-blue things. – Agent Smith May 17 '23 at 07:12
  • What is meant when a machine "observes" something and is this different than a human observation? What is a "broad generalization" in terms of machine learning? The problem with AI is that it really is nothing like the human brain so applying philosophical arguments that compare human behavior may not be appropriate. –  May 17 '23 at 14:54
  • @StevanV.Saban Nothing alike? It seems to me they are both made out of matter and electricity flows through them, and that they both meet the condition of [physical computer as per philosophers (SEP)](https://plato.stanford.edu/entries/computation-physicalsystems/). That doesn't seem like nothing to me. – J D May 17 '23 at 14:56
  • AI is an assembled machine that works solely on electrical impulses. The human brain is a bio-electrochemical system that is grown and not assembled. It learns as it grows. AI learns after it's assembled. Just adding a chemical component to the system adds degrees of freedom that is not possible in a simple electrical system. AI and the brain may perform similar functions but the mechanism behind those functions is very different. –  May 17 '23 at 15:07
  • As I recall, there is a method of creating a 'computer' using cups and paper clips that learns the winning strategy for a simple game. – Scott Rowe May 17 '23 at 23:00

0 Answers0