Science as Antiscience

[Azamat Abdoullaev]

Today's science is downgrading to pseudoscience with all the prospects of degenerating to antiscience, the rejection of universal science and causal scientific method, and its replacement with statistics and ad hoc hypotheses, often for financial, commercial and political gains.

This all comes from the fragmentarity, ignorance, narrow specialization, reductionism, scientism, and extensive commercialization of today’s science, failing to pursue universal knowledge.

https://www.linkedin.com/pulse/todays-science-pseudo-anti-science-why-ai-takes-over-abdoullaev/?published=t

[John F Sowa]

Ed and Azamat,

 

EJB: Today’s science, as conducted and reported by scientists via reputable journals, is not ‘downgrading’ in any real sense.

 

I agree.  But the WWW with its endless amount of "social media" buries the respectable science in mindless garbage or deliberate lies..

 

AA:  Today's science is downgrading to pseudoscience with all the prospects of degenerating to antiscience...

 

No.  As Ed wrote, good science is still good science.  And good scientists know good science when they see it.  Unfortunately, the huge volume of garbage on the WWW is making it harder for the general public to distinguish the solid work from the fluff and the lies.

 

For example, when Betty White died, her cause of death was not immediately reported because she died in her sleep and nobody knew

what happened.

 

But the liars, who are deliberately killing people with their lies, immediately claimed that she died from taking a booster shot of Covid vaccine.   When the autopsy was done, it turned out that she died of a stroke.  But the damage was done.  Millions of people repeated the lie that a booster shot killed Betty White.

 

Fundamental problem:  Facts spread linearly, but lies spread exponentially.

 

John

[John F Sowa]

Azamat,

 

Your note supports my point:  The DNN technology is just one more tool in the AI toolkit.  Applying DNNs to the data abpit Covid that has been gathered in the past two years has not led to any major breakthroughs.

 

The real breakthroughs have come from traditional scientific methods in analyzing DNA and RNA and discovering new methods for creating vaccines and drugs.  That is a fantastic achievement of traditional scientific methodology.

 

As I have been saying, all those words about metaversing and super Ai are nothing but a sales pitch for companies that want to extract money from sources with deep pockets.

 

AA: But here are some facts supporting my thesis that today's science is badly ill, if not dead. 

 

• despite a loud call to arms and many willing participants, the ML community has had surprisingly little positive impact against Covid-19. One of the most popular problems - diagnosing coronavirus pathology from chest X-ray or chest computed tomography images using computer vision - has been a universal clinical failure.  

   

• A systematic review of all papers published in 2020 that reported using ML for diagnosis and prognostication of Covid-19 found that “none of the reviewed literature reaching the threshold of robustness and reproducibility essential to support utilization in clinical practice.” There were many methodological, dataset, and bias issues. 

• For example, 25% of papers used the same pneumonia control dataset to compare adult patients without mentioning that it consists of kids aged 1-5. 

[John F Sowa]

Azamat,

 

I agree with you about the nature of science.  But the people who are talking about some great singularity in 2030 are *not* scientists.  They are engineers who are working for companies that are trying to get money for engineering projects .As I keep repeating, the many versions of DNNs have no similarity to actual neurons in the brain.  They are just algorithms for doing certain kinds of computations.

 

 

JFS:  The real breakthroughs have come from traditional scientific methods in analyzing DNA and RNA and discovering new methods for creating vaccines and drugs.  That is a fantastic achievement of traditional scientific methodology, which is then applied to (a) making people healthy and (b) making drug companies wealthy.

 

AA:  The standard steps or principles of the scientific method include the following algorithm:

• Ask Why and How, What and Who, What and Who questions as a problem statement.
• Perform research.
• Establish your inductive hypothesis, to approve or reprove the null hypothesis.
• Deduce predictions from your narrow or broad conjectures, testing your hypothesis by conducting an observation, experiment or simulation or reasoning.
• Make an observation or experiment or simulation.
• Analyze the results and draw a conclusion.
• Present the findings, announce your discovery…or start everything anew.

AA:  It is the iterative interactions of DATA and THEORIES, data and facts or observations and experimental results with concepts, abstractions or generalizations, hypotheses and theories. Traditionally, scientific investigation is about the research methods and techniques to develop hypotheses, gather data, conduct experiments, analyze data, and draw conclusions.

AA:  In reality, any deep research is deep data analytics and causal patterns, laws or interrelationships, causal claims, causal discovery, causal models and causal inferences. This is SCIENCE, not today's science relying on functional relations and statistical correlations.

JFS:  On these issues, we agree.  But Google, Facebook, and other companies that have been developing AI technology have one goal about all others:  MAKE MONEY!  They tolerate people who are doing science, but only on one condition:  it must make money.

 

I worked at IBM for thirty years, and I was very well aware of the pressure to do something that would make management happy -- i.e., make money.  But I also wanted to do serious research and publications.  I was able to juggle the issues by doing enough to make money so that I could spend more time on doing the kind of research I preferred.

 

Fundamental principle:  The work on artificial neural networks is pure engineering whose primary goal is to make money..  There are some analogies to what goes on in cognition, but they are very loose, very distant analogies.  Don't confuse that work with science or try to evaluate it as a contributions to science.

 

The people who run those companies have one and only one goal:  make money.  They fire people who put science ahead of money.  They also fire people who put ethics ahead of money.  After they're fired, many of those people testify against those companies.

 

Re Google:   I knew the people who were developing Google as a Stanford project.  It was a great innovation while it was still running on the Stanford computers.  But after they left Stanford, their technology kept degenerating.  Every innovation was designed to make money, not promote research.

 

John

 

PS:   I am not against the idea of making money.  But  it's important to be clear about the differences between science, engineering, and profits.  It's important to have some of the latter to support the former.  But don't confuse one for the other.

 

[John F Sowa]

Ed and Azamat,

 

The fundamental difference between science, engineering, and exploration is in the kinds of questions that are asked and answered.

 

1. An explorer or naturalist asks the questions who?, what?, when?, and where?.  The goal is a classification of the many kinds of things and events that have been found and their locations in time and space.

 

2. An engineer asks the question How?  The goal is to find some way of constructing something within the limits of available time, budget, and resources.  The process of answering the "How?" question would use the results of previous explorers, engineers, and scientists who asked other questions.  And it may have a side efffect of generating answers to the other kinds of questions.  But whoever is paying the budged is primarily interested in the results.  Very often, the people who control the funding would prefer to keep the other answers secret in order to give them a competitive edge.

 

3. A scientist asks the question Why?  The scientist usually begins with a vague hypothesis, which  does not fully answer "Why?"  And the process of answering the question may go far beyond the original  hypothesis.  It may, in fact, refute the hypothesis and replace it with a far more general theory.   But the answer to every "why?" question usually opens up far more questions than it answers.

 

EJB:  What makes science valuable in the long term is its additions to the pool of human knowledge; what makes it valuable to communities and businesses is its successful *application* to a problem they have.  Good science + applied science + engineering -> problem solutions.

JFS:  Engineers answer certain kinds of how-questions, but why-questions aren't what they are paid to answer.

AA:  In reality, any deep research is deep data analytics and causal patterns, laws or interrelationships, causal claims, causal discovery, causal models and causal inferences. This is SCIENCE, not today's science relying on functional relations and statistical correlations.

 

JFS: I agree.  But it's very hard to predict scientific breakthroughs.  It's also very difficult to get funding for pure science.  Many scientific breakthroughs are discovered in the process of solving engineering problems.  In my years at IBM, I saw a  lot of science that occurred as a side effect of doing the engineering.  But most of the funding came from people who wanted results, not theories.  From their point of view, theories are just  "fun facts" that you put in a research journal for added prestige.  The "how?" questions result in patents. which are much more profitable for the company.

EJB:  Finally, I would observe that ‘functional relations’ are a common way of stating accepted scientific ‘laws or interrelationships’. 

JFS:  All precise results of any kind are stated in some notation from some branch of mathematics.  Sometimes, the scientist or the engineer creates a new branch of math.   Formal logic is just a name for some version of "mathematical logic"..  And there are many versions that emphasize different kinds of relations.

John

[Avril Styrman]

John, here's another fundamental difference between science and engineering:

A physicist makes 10 tests and one succeeds: he/she gets invited to Stockholm.

An engineer builds 10 bridges and one falls: he/she gets invited to prison.

Cheers,

Avril

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/12946126d3494fc295143e857a71dc8a%40bestweb.net.

--

Kind regards,

Avril Styrman, PhD

+358 40 7000 589

DoubleVerify

Physics Foundations Society 

ResearchGate

Academia

YouTube

[John F Sowa]

Avril,

 

That point, although oversimplified, is a corollary of what I was saying:

 

 

AS: A physicist makes 10 tests and one succeeds: he/she gets invited to Stockholm.  An engineer builds 10 bridges and one falls: he/she gets invited to prison.

 

The goal of a scientist is to find answers to why-questions.   An answer to one great why-question is worth a Nobel prize.That is independent of how many partial answers or wrong answers that he or she claimed over a career.

 

But an engineer is paid by someone to find an answer to a how-question and to demonstrate that the answer indeed satisfy the goals of the contract.  If the answer fails to meet the goals, the engineer may be sued.  If the answer fails because of fraud or negligence, the engineer may be found guilty of some crime.

 

John