Connectionists: Scientific Integrity, the 2021 Turing Lecture, etc.

[Randall O'Reilly]

I'm sure everyone agrees that scientific integrity is essential at all levels, but I hope we can avoid a kind of simplistic, sanctimonious treatment of these issues -- there are lots of complex dynamics at play in this or any scientific field.  Here's few additional thoughts / reactions:

* Outside of a paper specifically on the history of a field, does it really make sense to "require" everyone to cite obscure old papers that you can't even get a PDF of on google scholar?  Who does that help?  Certainly not someone who might want to actually read a useful treatment of foundational ideas.  I generally cite papers that I actually think other people should read if they want to learn more about a topic -- those tend to be written by people who write clearly and compellingly.  Those who are obsessed about historical precedents should write papers on such things, but don't get bent out of shape if other people really don't care that much about that stuff and really just care about the ideas and moving *forward*.

* Should Newton be cited instead of Rumelhart et al, for backprop, as Steve suggested?  Seriously, most of the math powering today's models is just calculus and the chain rule.  Furthermore, the idea that gradients passed through many multiplicative steps of the chain rule tend to dissipate exponentially is pretty basic at a mathematical level, and I'm sure some obscure (or even famous) mathematician from the 1800's or even earlier has pointed this out in some context or another.  For example, Lyopunov's work from the late 1800's is directly relevant in terms of iterative systems and the need to have an exponent of 1 for stability.  So at some level all of deep learning and LSTM is just derivative of this earlier work (pun intended!).

* More generally, each individual scientist is constantly absorbing ideas from others, synthesizing them in their own internal neural networks, and essentially "reinventing" the insights and implications of these ideas in their own mind.  We all only have our own individual subjective lens onto the world, and each have to generate our own internal conceptual structures for ourselves.  Thus, reinvention is rampant, and we each feel a distinct sense of ownership over the powerful ideas that we have forged in our own minds.  Some people are lucky enough to be at the right place and the right time to share truly new ideas in an effective way with a large number of other people, but everyone who grasps those ideas can cherish the fact that so many people are out there in the world working tirelessly to share all of these great ideas!

* To support what Pierre Baldi said: People are strongly biased to form in-group affiliations and put others into less respected (or worse) out-groups -- the power of this instinct is behind most of the evil in the world today and throughout history, and science is certainly not immune to its effects.  Thus it is important to explicitly promote diversity of all forms in scientific organizations, and work against what clearly are strong "cliques" in the field, who hold longstanding and disproportionate control over important organizations.

- Randy

> On Oct 29, 2021, at 4:13 AM, Anand Ramamoorthy <valvilraman at yahoo.co.in> wrote:
> 
> Hi All,
>                     Some remarks/thoughts:
> 
> 1. Juergen raises important points relevant not just to the ML folks but also the wider scientific community
> 
> 2. Setting aside broader aspects of the social quality of the scientific enterprise, let's take a look at a simpler thing; individual duty. Each scientist has a duty to science (as an intellectual discipline) and the scientific community, to uphold fundamental principles informing the conduct of science. Credit should be given wherever it is due - it is a matter of duty, not preference or "strategic vale" or boosting someone because they're a great populariser.  
> 
> 3. Crediting those who disseminate is fine and dandy, but should be for those precise contributions, AND the originators of an idea/method/body of work ought to be recognised - this is perhaps a bit difficult when the work is obscured by history, but not impossible. At any rate, if one has novel information of pertinence w.r.t original work, then the right action is crystal clear. 
> 
> 4. Academic science has loads of problems and I think there is some urgency w.r.t sorting them out. For three reasons; a) scientific duty b) for posterity and c) we now live in a world where anti-science sentiments are not limited to fringe elements and this does not bode well for humanity.  
> 
> Maybe dealing with proper credit assignment as pointed out by Juergen and others in the thread could be a start. 
> 
> Live Long and Prosper!
> 
> Best, 
> 
> Anand Ramamoorthy
> 
> 
> 
> On Friday, 29 October 2021, 08:39:14 BST, Jonathan D. Cohen <jdc at princeton.edu> wrote:
> 
> 
> The same incentive structures (and values they reflect) are not necessarily the same — nor should they necessary be — in commercial and academic environments.
> 
> jdc
> 
> 
> 
>> On Oct 28, 2021, at 12:03 PM, Marina Meila <mmp2 at uw.edu> wrote:
>> 
>> Since credit is a form of currency in academia, let's look at the "hard currency" rewards of invention. Who gets them?  the first company to create a new product usually fails.
>> However, the interesting thing is that society (by this I mean the society most of us we work in) has found it necessary to counteract this, and we have patent laws to protect the rights of the inventors.
>>  
>> The point is not whether patent laws are effective or not, it's the social norm they implement. That to protect invention one should  pay attention to rewarding the original inventors, whether we get the "product" directly from them or not. 
>>  
>> Best wishes,
>>  
>>        Marina
>>  
>> -- Marina Meila
>> Professor of Statistics
>> University of Washington
>>  
>>  
>> On 10/28/21, 5:59 AM, "Connectionists" <connectionists-bounces at mailman.srv.cs.cmu.edu> wrote:
>> 
>> As a friendly amendment to both Randy and Danko’s comments, it is also worth noting that science is an *intrinsically social* endeavor, and therefore communication is a fundamental factor.  This may help explain why the *last* person to invent or discover something is the one who gets the [social] credit.  That is, giving credit to those who disseminate may even have normative value.  After all, if a tree falls in the forrest… As for those who care more about discovery and invention than dissemination, well, for them credit assignment may not be as important ;^). 
>> jdc
>>  
>>  
>> On Oct 28, 2021, at 4:23 AM, Danko Nikolic <danko.nikolic at gmail.com> wrote:
>>  
>> Yes Randall, sadly so. I have seen similar examples in neuroscience and philosophy of mind. Often, (but not always), you have to be the one who popularizes the thing to get the credit. Sometimes, you can get away, you just do the hard conceptual work and others doing for you the (also hard) marketing work. The best bet is doing both by yourself. Still no guarantee.  
>> Danko
>>  
>>  
>>  
>>  
>> On Thu, 28 Oct 2021, 10:13 Randall O'Reilly <oreilly at ucdavis.edu> wrote:
>>  
>>  
>> I vaguely remember someone making an interesting case a while back that it is the *last* person to invent something that gets all the credit.  This is almost by definition: once it is sufficiently widely known, nobody can successfully reinvent it;  conversely, if it can be successfully reinvented, then the previous attempts failed for one reason or another (which may have nothing to do with the merit of the work in question).
>>  
>> For example, I remember being surprised how little Einstein added to what was already established by Lorentz and others, at the mathematical level, in the theory of special relativity.  But he put those equations into a conceptual framework that obviously changed our understanding of basic physical concepts.  Sometimes, it is not the basic equations etc that matter: it is the big picture vision.
>>  
>> Cheers,
>> - Randy
>>  
>> > On Oct 27, 2021, at 12:52 AM, Schmidhuber Juergen <juergen at idsia.ch> wrote:
>> > 
>> > Hi, fellow artificial neural network enthusiasts!
>> > 
>> > The connectionists mailing list is perhaps the oldest mailing list on ANNs, and many neural net pioneers are still subscribed to it. I am hoping that some of them - as well as their contemporaries - might be able to provide additional valuable insights into the history of the field.
>> > 
>> > Following the great success of massive open online peer review (MOOR) for my 2015 survey of deep learning (now the most cited article ever published in the journal Neural Networks), I've decided to put forward another piece for MOOR. I want to thank the many experts who have already provided me with comments on it. Please send additional relevant references and suggestions for improvements for the following draft directly to me at juergen at idsia.ch:
>> > 
>> > https://people.idsia.ch/~juergen/scientific-integrity-turing-award-deep-learning.html
>> > 
>> > The above is a point-for-point critique of factual errors in ACM's justification of the ACM A. M. Turing Award for deep learning and a critique of the Turing Lecture published by ACM in July 2021. This work can also be seen as a short history of deep learning, at least as far as ACM's errors and the Turing Lecture are concerned.
>> > 
>> > I know that some view this as a controversial topic. However, it is the very nature of science to resolve controversies through facts. Credit assignment is as core to scientific history as it is to machine learning. My aim is to ensure that the true history of our field is preserved for posterity.
>> > 
>> > Thank you all in advance for your help! 
>> > 
>> > Jürgen Schmidhuber
>> > 
>> > 
>> > 
>> > 
>> > 
>> > 
>> > 
>> > 
>