Re: The nature of Heuristics. D Lenat (Reader)

I mistyped Alan’s family name, which is: Alan Collins. His work includes:

The logic of plausible reasoning: A core theory, 1989, see: https://www.sciencedirect.com/science/article/abs/pii/0364021389900104

> The theory consists of three parts: a formal representation of plausible inference patterns; such as deductions, inductions, and analogies, that are frequently employed in answering everyday questions; a set of parameters, such as conditional likelihood, typicality, and similarity, that affect the certainty of people's answers to such questions; and a system relating the different plausible inference patterns and the different certainty parameters.


Collins also worked with Dedre Gentner on analogies and qualitative reasoning for mental simulations of physical processes, see https://www.academia.edu/1037745/How_People_Construct_Mental_Models

A much earlier work from 1954 is by G. Polya on induction and analogy in mathematics. He notes that

> Demonstrative reasoning is safe, beyond controversy and final. Plausible reasoning is hazardous, controversial and provisional. … Anything new we learn about the world involves plausible reasoning, which is the only kind of reasoning for which we care in everyday affairs.

In essence, new ideas in science and mathematics start with plausible reasoning before being formalised.

Further work is needed on declarative means to express strategies and tactics for argumentation, along with associated work on machine learning. It is probably going to be easier and more scalable to work on neural networks that learn and reason like we do if the goal is to construct tools to help with the kinds of applications envisaged by Lenat and others.

See: https://www.academia.edu/42170473/Polya_1954_Mathematics_and_Plausible_Reasoning_I_Induction_and_Analogy_in_Mathematics


> On 8 Sep 2024, at 11:50, Paola Di Maio <paoladimaio10@gmail.com> wrote:
> 
> Thank you D
> You may have posted about plausible reasoning before, perhaps, if you have a spare neuron or two,
> sometime do a slide or two on comparing/contrasting heuristics with plausible R
> in the context of AI KR, could be nice.
> 
> 
> 
> On Sun, Sep 8, 2024 at 12:47 PM Dave Raggett <dsr@w3.org <mailto:dsr@w3.org>> wrote:
>> Lenat says:
>> 
>>> Heuristics are compiled hindsight, and draw their power from the various kinds of regularity and continuity in the world; they arise through specialization, generalization, and —surprisingly often— analogy. 
>> 
>> 
>> You could argue that generative AI is mostly about heuristics as winning memes in the parameter space. Whilst this lacks transparency it works very effectively in respect to text, images, video and sound by exploiting statistical regularities at different levels of abstraction.  Heuristics expressed in symbolic form are by comparison transparent and weaker in their applicability, especially when they lack the metadata needed for plausible reasoning with fuzzy concepts. Lenat was probably unaware of the work being done by Alan Clark and others on plausible reasoning as a means to model human problem solving.
>> 
>>> On 8 Sep 2024, at 10:02, Paola Di Maio <paola.dimaio@gmail.com <mailto:paola.dimaio@gmail.com>> wrote:
>>> 
>>> Hope everybody had a decent summer
>>> 
>>> As we weigh human vs artificial intelligence, the question of whether it is possible to automate
>>> heuristic reasoning comes up. I have the pleasure of remembering  Doug Lenat with the attached paper written almost half a century ago and still relevant today
>> 
>> UK summer was pretty wet and cool apart from a few warm days. My colleagues in Europe have had to put up with excess heat.
>> 
>> Dave Raggett <dsr@w3.org <mailto:dsr@w3.org>>
>> 
>> 
>> 

Dave Raggett <dsr@w3.org>