[Fis] Is Dataism the end of classical hypothesis-driven research and the beginning of data-correlation-driven research?

Sat Mar 10 21:05:45 CET 2018

Dear Plamen and Colleagues,
If it can be feasible, I would very much welcome what you propose. Yes, it 
would be great developing a general articulation amongst all our exchanges. 
Roughly, I feel that a fundamental nucleous of neatly conceptualized 
information is still evading us, but outside that nucleous, and somehow 
emanating from it, there are different branches and sub-branches in quite 
different elaboration degrees and massively crisscrossing and intermingling 
their contents. A six-pointed star, for instance, radiating from its inner 
fusion the computational, physical, biological, neuronal, social, and 
economic. The six big branches in perfect periferic colussion and confusion. 
Could a blockchain, along its full develpment in time, represent a 
fundamental cartography of the originating fusion nucleous?
About dataism enchantment, well, too many times we have been said "look, 
finally
this is the great, definitive scientific approach"--behaviorism, artificial 
intelleigence, artifficial catastrophe & complexity theory, and so on. Let 
us wait and see. Welcome in the extent to which it really responds to 
unanswered questions. And let us be aware of the technocratic lore it seems 
to drag.
This was my second cent for the week.
best--Pedro

On Fri, 9 Mar 2018 10:30:01 +0100 "Dr. Plamen L. Simeonov"  wrote:
>These are wise words, Pedro.
>What I was meaning with my previous posting on FIS was that there is a foundational emerging technology - blockchain - that could give us, scientists organized in fora like FIS, IB, IS4IS etc. to become a valuable currency of the future. I am speaking not about finances or resources like petrol, gold, water, etc. What we are doing all the time with the exchange of ideas online are in fact transactions, often with huge potential. Why do not

try to elevate them to the level that they deserve? 
>
>I am not sure if the FIS forum members can follow me. Can you?
>
>All the best.
>
>Plamen
>
>
>____________________________________________________________
>
>
>On Thu, Mar 8, 2018 at 6:15 PM, PEDRO CLEMENTE MARIJUAN FERNANDEZ <pcmarijuan.iacs at aragon.es> wrote:
>>
>>head>Dear Alberto,
>> 
>>Many thanks for the kickoff text. I will try to produce acouple of direct comments.
>>You have reminded me of the early70's, when I first approached science. A few computers had made theirentrance in the university halls. During those years, and for somedecades to come, a new mantra was to be ensconced: modeling,simulations. Thanks to computers, we had a fascinating new tool; amathematical machine that was opening a new window to the world ofscience, equivalent to the telescope or the microscope in thescientific revolution. Now, almost 50 years
later, after havingprovoked their own "information revolution" it seems that 
computersare more than a new tool. Dataism coupled with 
artificialintelligence, deep learning and the other techniques, have taken 
themto the command post, so that they are becoming direct "agents" of 
thescientific progress. And this is strange. They have already 
defeatedmasters of chess, of go and of other contests... are they going 
todefeat scientists too? Are they the "necessary" new lords of allquarters 
of techno-social complexity?
>>You have depicted verycogently the new panorama of biomedical research, probably themainstream, and I wonder whether this is the most interestingdirection of advancement. In some sense, yes (or no!), as it is wherebig biomed companies, technological firms, and managementestablishment are pointing at. It is easy to complain that they areleaving aside the integrative vision, the
meaningful synthesis thatfacilitate our comprehension, the "soul" in the 
machine... But we havebeen complaining in this way at least during the last 
two decades. SoI really do not know. Fashions in science come and go: maybe 
all ofthis is a temporary illusion. Or a taste of the science of the future.
>>In any case, it was nice hearing from a biomedical researcher inthe wet lab.
>>Best wishes--Pedro
>> 
>>On Tue, 06Mar 2018 21:23:01 +0100 "Alberto J. Schuhmacher"  wrote:
>>blockquote>Dear FIS Colleagues,
>>I very much appreciate thisopportunity to discuss with all of you.
>>My mentors and scienceteachers taught me that Science had a method, rules and proceduresthat should be followed and pursued rigorously and with perseverance.The scientific research needed to be preceded by one or severalhypotheses that should be subjected to validation or refutationthrough experiments designed and
carried out in a laboratory. TheOxford Dictionaries Online defines the 
scientific method as "a methodor procedure that has characterized natural 
science since the 17thcentury, consisting in systematic observation, 
measurement, andexperiment, and the formulation, testing, and modification 
ofhypotheses". Experiments are a procedure designed to test 
hypotheses.Experiments are an important tool of the scientific method.
>>Inour case, molecular, personalized and precision medicine aims toanticipate the future development of diseases in a specific individualthrough molecular markers registered in the genome, variome,metagenome, metabolome or in any of the multiple "omes" that make upthe present "omics" language of current Biology.
>>Thepossibilities of applying these methodologies to the prevention andtreatment of diseases have increased exponentially with the rise of anew religion, Dataism,
whose foundations are inspired byscientific agnosticism, a way of thinking 
that seems classical butapplied to research, it hides a profound revolution.
>>Dataismarises from the recent human desire to collect and analyze data, dataand more data, data of everything and data for everything-from themost banal social issues to those that decide the rhythms of life anddeath. “Information flow” is one the “supreme values” of thisreligion. The next floods will be of data as we can see just lookingat any electronic window.
>>The recent development of giganticclinical and biological databases, and the concomitant progress of thecomputational capacity to handle and analyze these growing tides ofinformation represent the best substrate for the progress of Dataism,which in turn has managed to provide a solid content material to analways-evanescent scientific agnosticism.
>>On many occasions
theestablishment of correlative observations seems to be sufficient toinfer 
about the relevance of a certain factor in the development ofsome human 
pathologies. It seems that we are heading towards a path inwhich research, 
instead of being driven by hypotheses confirmedexperimentally, in the near 
future experimental hypotheses themselveswill arise from the observation of 
data of previously performedexperiments. Are we facing the end of the wet 
lab? Is Dataism the endof classical hypothesis-driven research (and the 
beginning ofdata-correlation-driven research)?
>>Deep learning is based onlearning data representations, as opposed to task-specific algorithms.Learning can be supervised, semi-supervised or unsupervised. Deeplearning models are loosely related to information processing andcommunication patterns in a biological nervous system, such as neuralcoding that attempts to define a
relationship between various stimuliand associated neuronal responses in the 
brain. Deep learningarchitectures such as deep neural networks, deep belief 
networks andrecurrent neural networks have been applied to fields 
includingcomputer vision, audio recognition, speech recognition, 
machinetranslation, natural language processing, social network 
filtering,bioinformatics and drug design, where they have produced 
resultscomparable to and in some cases superior to human experts. Will 
bedata-correlation-driven research the new scientific method forunsupervised 
deep learning machines? Will computers becamefundamentalists of Dataism?
>>Best regards,
>>AJ
>>p> ---
>>Alberto J. Schuhmacher,PhD.
>> Head, MolecularOncology Group
>> 
>> AragonHealth Research Institute (IIS Aragón)
>>Biomedical Research Center of Aragon (CIBA)
>> Avda. Juan Bosco 13, 50009 Zaragoza (Spain)br> email:
ajimenez at iisaragon.es <https://mailto:ajimenez@iisaragon.es>
>> Phone:(+34) 637939901 <unknown://tel:+34%20637%2093%2099%2001>  
>>
>> 
>>
>>_______________________________________________
>>Fis mailing list
>>Fis at listas.unizar.es
>>http://listas.unizar.es/cgi-bin/mailman/listinfo/fis
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listas.unizar.es/pipermail/fis/attachments/20180310/8788ba8c/attachment.html>