Uncommon Descent Serving The Intelligent Design Community

Science has outgrown the human mind? Now needs AI?

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

From molecular cancer biologist Ahmed Alkhateeb at Aeon:

Science is in the midst of a data crisis. Last year, there were more than 1.2 million new papers published in the biomedical sciences alone, bringing the total number of peer-reviewed biomedical papers to over 26 million. However, the average scientist reads only about 250 papers a year. Meanwhile, the quality of the scientific literature has been in decline. Some recent studies found that the majority of biomedical papers were irreproducible.

The twin challenges of too much quantity and too little quality are rooted in the finite neurological capacity of the human mind. Scientists are deriving hypotheses from a smaller and smaller fraction of our collective knowledge and consequently, more and more, asking the wrong questions, or asking ones that have already been answered. Also, human creativity seems to depend increasingly on the stochasticity of previous experiences – particular life events that allow a researcher to notice something others do not. Although chance has always been a factor in scientific discovery, it is currently playing a much larger role than it should.

One promising strategy to overcome the current crisis is to integrate machines and artificial intelligence in the scientific process. More.

No. The mountain of data is not the main problem. The main problem is perverse incentives. Artificial intelligence will only carry out the wishes of its programmers; it can’t be better than them.

Also, keep up to date with Retraction Watch

See also: Breaking: National Academy of Sciences notices research integrity problem.
Unfortunately, I (O’Leary for News) have been covering these “sweeping reviews” for probably fifteen years. Mostly, they amount to clucking noises followed by sweeping the problems under the carpet when all is quiet. The kinds of conversations people would need to have, in order to make a real difference, are conversations few are ready for.

Follow UD News at Twitter!

Comments
Dionisio: Yes, I've seen Amadeus. Fun show. Antonio Salieri as the bitterly envious antagonist-admirer, and Amadeus Mozart as the vulgar (but brilliant) clown. True to life in the latter case, I've heard.LocalMinimum
April 29, 2017
April
04
Apr
29
29
2017
07:04 AM
7
07
04
AM
PDT
LocalMinimum, You've touched a delicate nerve, but I'm glad you did it. The UD thread with the most suggestive title is "Mystery at the heart of life". Why do some folks understand something that other even more educated people don't. Did you watch the movie Amadeus which won several academy awards in the 1980s or 1990s?Dionisio
April 28, 2017
April
04
Apr
28
28
2017
09:17 PM
9
09
17
PM
PDT
Dionisio:
...some highly educated and academically influential folks claim with kind of ‘convincing’ verbosity* that the powerfully pleiotropic problem-solving capacity of Darwinism has been satisfactorily demonstrated long ago, but we just don’t understand evolution.
Well, surely, as we are here, able to have this conversation, SOMETHING must have made it possible; with our only other option being some mysterious Designer doing mysterious things that we can't even imagine (even as we are racing to do such things ourselves to prove no such thing need happen). Is this not sufficient demonstration? If only we could understand it; then we'd be able to unlock the general problem solving capabilities inherent to sunlight and soil with our laser etched and liquid helium thermos enclosed processors. Who needs Artificial Intelligence when the universe is permeated with Natural Intelligence? It would be a truly Green (Brown?) revolution!LocalMinimum
April 28, 2017
April
04
Apr
28
28
2017
03:23 PM
3
03
23
PM
PDT
LocalMinimum:
[...] the stack space of mutation + natural selection is zero – the result space is the only thing it can use as memory, which doesn’t work in practice AT ALL.
Very interesting way to put it.Dionisio
April 28, 2017
April
04
Apr
28
28
2017
01:37 PM
1
01
37
PM
PDT
LocalMinimum,
But the day when you can drop a (precise, correct, unfudged) definition of Darwinism into a general problem solver, and it churns out a either a solution set exclusively populated by extra-astronomical improbabilities, or the null set…that would be a very amusing day…when Darwinism becomes the new division by zero.
Wow! That's really funny. Thank you for making me laugh out loud. But let's get serious, because some highly educated and academically influential folks claim with kind of 'convincing' verbosity* that the powerfully pleiotropic problem-solving capacity of Darwinism has been satisfactorily demonstrated long ago, but we just don't understand evolution. (*) parole, parole, parole... :)Dionisio
April 28, 2017
April
04
Apr
28
28
2017
01:26 PM
1
01
26
PM
PDT
Well, actually, it's true. Already, even. We approach tensor problems that would take years to do by hand nowadays in engineering and science. When programming, I am acutely aware of my "stack space" when hammering out some recursively hierarchical object (and losing half of it when someone says "hi" to me). It's that "stack space" that is the limit of the sophistication of what I can build - which, note, the stack space of mutation + natural selection is zero - the result space is the only thing it can use as memory, which doesn't work in practice AT ALL. We need better calculators, and we'll eventually have calculators that can frame a problem as well as provide solutions to it. This will be very useful. However, the problems we'll have framed will still rely on the people ordering the framing, and you might eventually be able to produce lies which are prohibitively hard to approach by hand with the right software suite by employing a sufficiently deep/hidden kludge. This is why open source is awesome. But the day when you can drop a (precise, correct, unfudged) definition of Darwinism into a general problem solver, and it churns out a either a solution set exclusively populated by extra-astronomical improbabilities, or the null set...that would be a very amusing day...when Darwinism becomes the new division by zero.LocalMinimum
April 28, 2017
April
04
Apr
28
28
2017
07:17 AM
7
07
17
AM
PDT
In 2012, the University of Connecticut announced that it had concluded that Dipak K. Das, Ph.D., a professor in its Department of Surgery and director of the Cardiovascular Research Center, was guilty of 145 counts of fabrication and falsification of data and that the university had notified eleven journals about this problem [20]. In recent years, Das had gained attention for his reports on allegedly beneficial properties of resveratrol. As of March 2014, journals had retracted 20 of his papers, many of which were repeatedly cited by others [21]. Das died in 2013. Scientific journals notified following research misconduct investigation. UConn Today, Jan 12, 2012. 21.Oransky I. Late resveratrol researcher Dipak Das up to 20 retractions. Retraction Watch Web site, April 3, 2014.
http://www.quackwatch.org/01QuackeryRelatedTopics/DSH/resveratrol.html But who knows how reliable is the source of that news report itself?Dionisio
April 28, 2017
April
04
Apr
28
28
2017
04:16 AM
4
04
16
AM
PDT
"[...] the average scientist reads only about 250 papers a year."
Does that mean that the readers of this website UD are above the level of the average scientist in reading science-related papers? Just in two discussion threads alone there have been over 3000 biology-related papers referenced within around 3 years, which makes an average of a thousand referenced papers per year? Assuming that the anonymous readers of those two threads --which together have been visited over 8 thousand times without leaving "footprints" in the form of posted comments-- have read just a third of the referenced papers, then it's over 300 papers per year? No wonder some of the politely dissenting interlocutors are so upset about seeing the increasing number of biology-related papers referenced in UD lately. :)Dionisio
April 28, 2017
April
04
Apr
28
28
2017
03:40 AM
3
03
40
AM
PDT
The twin challenges of too much quantity and too little quality are rooted in the publish or perish culture. You must publish even if the results are trivial or questionable.aarceng
April 28, 2017
April
04
Apr
28
28
2017
01:18 AM
1
01
18
AM
PDT
Interesting points. Thanks. Would more honesty help too? Would more humility help too? Would more open-mindedness help too? Would more thinking out of wrongly preconceived boxes help too? Would top-down research approach help too? At least the words 'surprisingly' and 'unexpectedly' would appear less frequently in research papers. :)Dionisio
April 27, 2017
April
04
Apr
27
27
2017
09:49 PM
9
09
49
PM
PDT

Leave a Reply