Fuzzy logic usage


I want to summarize a little bit: “What the Fuzzy logic is about”, so that you can get some feeling out of it.

Fuzzy logic includes 0 and 1 as extreme cases of truth but also includes the various states of truth in between so that, for example, the result of a comparison between two things could be not “tall” or “short” but “0.38 of tallness.”

Fuzzy logic is almost synonymous with the theory of fuzzy sets, a theory which relates to classes of objects with unsharp boundaries in which membership is a matter of degree.

There is a high number of things/objects with unsharp boundaries. For example think about the surface area of Pacific Ocean. Wikipedia states that the value is 165,250,000 km2, but we can immediately see that it is rounded and contains a lot of imprecision. How would you express the value to make it more precise:

  • What about tides (water fluctuation, Moon’s movement, Earth rotation, solar gravitational effects, …)? The high of the ocean differs throughout day and year. Should we have a function with time parameter?
  • What about icebergs melting?
  • What about shore lines constant change?
  • What precision should we except: km2, m2, mm2?

It is very useful to understand that it is merely impossible to get the most accurate value which correctness can not be improved any more. This is stressing out how importatn it is to learn working with imperfection and how to get practical maximum out of it.

A trend that is growing relates to the use of fuzzy logic in combination with neurocomputing and genetic algorithms. More generally, fuzzy logic, neurocomputing, and genetic algorithms may be viewed as the principal constituents of what might be called soft computing. Unlike the traditional, hard computing, soft computing accommodates the imprecision of the real world.

Fuzzy logic is essential to the development of human-like capabilities for AI. Living organisms aggregate data and form a number of partial truths which are being aggregated further into higher truths which in turn, when certain thresholds are exceeded, cause certain further results such as motor reaction. A similar kind of process is used in:

Lotfi Zadeh’s research (credited with fuzzy logic’s formulation) focused on implementing methods that allowed computers to understand human language.

References / Quotation sources

List of references, from which I’ve quote + taken expressions:


Here is the underlay github repo: https://github.com/underlay/
I can’t find any active forums for the project however, looks to be in stealth mode still.


“in fact all AI is is a search problem”

I knew it! Thank you.


@dirvine , what is your take on the potential of AI with the known tech out there across the span of the forseeable future? What is your sense of the likely upper and lower bound for the forseeable future? And if it is helpful in paticular I am as always interested in the difference for the end user, citizen an consumer.

If all AI is a search problem its as if it was the future of Google all along. We know a little about what Google has been doing with a couple acquisitions and its Tensor Flow and custom asics and its attempts to marry q compute and neural nets with vast data sets. Musk thinks Google is the one to watch. So maybe its a guess about how far Google is likely to go over the next 20 years or a span as far forward as that which would take us backwards to Google’s rise to prominence. And apropos to SAFE as Google is a force that might affect the path SAFE takes.


I think it will move from search within bounds to open ended AI, i.e. continuous learning.

Upper bound I think it will stand a great chance of finding flaws in our thinking, i.e. math, science and research (papers) as well as improve medicine at least the reading scans and xrays parts as well as genome medicine match, I see this as a lower bound. Upper bound, well that it much more difficult as I think as we gain knowledge then the upper bound will be shifting sands. I suspect there will be huge resistance to truth matching capabilities as AI shows us probabilities of who has acted in which way on a world stage. This is where I think humankind will find a great battle with truth and consequently AI. Knowledge will win, eventually though.


There are some good talks on AI on these youtube channels - EDIT: added Lex Fridman’s channel - this is a really great one for technical AI discussion:


Thank you. Truth matching. It would be wonderful and exactly what I’ve wanted but I do understand it will be quite a culture shock.