First principles of doing a PhD and taking up an industrial jobs are quite different, which this article sidesteps. I am talking from the perspective of someone who did a PhD, postdoc and migrated to be a founder/CEO.
A PhD system trains you to think about unsolved problems in an given domain deeply with a larger time runway. The end goal is not a tangible product that reaches millions of people, but rather a set of ideas that can take a crack at the unsolved problems in your field in a novel way. A good work should inspire others in the field, and eventually a larger audience to pick them up and expand and build on top of it. To give a small example, a majority of the fundamentals of machine learning was charted out by many, many PhD works over the last 40 years. Implementing a linear classifier is 2 lines of code in 2018, but many Bothans died to bring us this information :-) .
The goals of industry are more immediate. Expect for a privileged few research labs in industry, your work is expected to be monetized, and rightly so. The goal is for you, if you run the business, else your management team to first figure out a problem of high relevance and monetary value. Build products/solutions for that problem, that can be used by someone who is less versed/ambivalent of your technical solutions. Efficacy of solving that particular problem often defines the merit of your contribution.
The fundamental of choosing the PhD or industry should be taking stock of what kind of contribution you want to make as an individual. If it is a few set of ideas to science, which on a later date might become something fundamental in our understanding of the world, then PhD is a good path. If it is a set of contributions towards a product/solution that eases the pain of many users then go into the industry first.
Author here. Just to clarify, motive of this work is to ease curation, with the massive amount of content being created; but by no means an attempt at creativity or originality.
The work is not at all contradictory to Adorno, especially in the sense that it is explicitly trying to as non-reductionist as possible, and assuming notion of aesthetics is a dynamic entity .
There is a finite pattern in the dataset; more interesting, it has its interesting share of subtleties ( for example, as opposed to a image classification problems), and the technological question is whether we can capture these.
But there is another interesting data question. For our work, we curated our training set with the help of expert curators. But the dataset itself is a metamorphising entity; i.e. it is subject to revision ( it is a continuous process for us at the moment), but more interestingly it is a chance for open debate between our curators. In some sense, technology allow to codify and challenge our notion of aesthetics ( especially with the evolution in our training sets) at a given point of time.
Thanks for the followup. I enjoyed reading the article and it is a very interesting project and a great effort. I did read at the end that it was about how to amplify the efforts of the human curators -- a great problem to tackle.
The author here. I used the term "understanding", not as in machines understanding the images, but more as scientific attempt in understanding aesthetics. ( <snippet from the text>"empowering me to develop systems for understanding images from a computational and scientific perspective"</snippet ends> ).
I agree with you completely about "bad industry-focused research". It serves no end. My question is that, is it just a reflection of mediocracy and gaming/dishonesty being everywhere, including academica ?
In this case, please trying to take shortcuts towards their goal of improving the amount of publications and grant approvals. There is reward in the system for this kind of behaviour. You being a reviewer/jury position unfortunately do not have the luxury of a filter.
Is there a way to early catch this , by looking at past trends ?
Slight detour is that this is one of my rationale for spending time reviewing papers for journals/conferences. In average, only 10 to 20% of papers I review really stands out or appeal to me, which is correlated to acceptance rate of a top journal/conference.
I worked in industry research for a long time and almost every intern or new hire came in with laughably wrong assumptions, but it wasn't due to mediocracy, gaming, or dishonesty. They just didn't have anywhere to learn from. Their assumptions were copied from earlier "best paper" winners in the field whose assumptions were copied from a random previous paper whose assumptions were probably mostly speculation.
One of the authors here. Since it is a optimization, the difficulty can be controlled as a parameter ( the lambda parameter in Eq.1 in the paper ).
But you are right, some of the puzzles can be super-hard ( for example, the Seurat puzzle ) that we used to joke between ourself to name our paper "taking the fun out of puzzles".
Personally, what was fascinating for me is the shape of the puzzle curve it produced. Most of the common puzzles are grid based (i.e. four neighbours - up , down, left, down ). But in this scheme, there can be strange neighborhood pieces, with even stranger shapes.
Exactly what you said, this solution doesn't make super-hard puzzles, it makes not-fun puzzles. The obvious next step that I would challenge you to do would be to optimizes for puzzle cuts so that they do not follow the main color lines of the image as much as possible. This would result in pieces that are just as interesting, but produce puzzles that are radically more fun to actually put together.
A good puzzle will not bore the player. Boring is that there are many blue pieces that all look exactly alike. Trying each one over and over is not fun. But with a tweak to your technique you could make sure that every blue piece that can contain another color would, thus increasing the variance of unique pieces which is not boring for players. Finding a blue piece with this little bit of brown and a spec of purple means the player can get to remember that and hunt for where it could go which is the fun part of puzzles. Do that and you can approach puzzle companies to license them a tool to make more fun puzzles and they will buy it.
Pushing away from color lines is very easy. For negative values of lambda in Eq.1 of the paper, (i.e. reverse the cost for color lines ), the optimization tries to push the puzzle shapes away from the color lines. We had tried a few in this configuration, but personally my co-authors and I liked the puzzles generated by the scheme of adhering to the color lines.
Hardness/fun factor in puzzles is a matter of personal taste. Hence the ability to personalize is very interesting. Not everyone like to solve 10000+ piece puzzles , nor color line adherence, but people invest time and effort in solving them.
This is really nice, from the scientific side, and it creates awesome puzzles as well. Any chance that we can see this software as a service, or maybe even for download?
The site has an interesting history. The former Stadtschloss suffered serious destruction during WWII ( https://en.wikipedia.org/wiki/City_Palace,_Berlin ) and the Communist East tore it completely down to build the Palace of the Republic https://en.wikipedia.org/wiki/Palace_of_the_Republic,_Berlin and acted as the hub of DDR government. Once the wall fell, and DDR disintegrated, in 2008 DDR's Palace of Republic was almost completed demolished, and work on new Stadtschloss which very much resembles the original commenced.
I this is a bit of legacy ( like the Stanford Bunny ). One of the seminal work of lighting was Photon mapping by Henrik Jensen , and he used Ludwig Mies van der Rohe's structure as an example ( http://graphics.ucsd.edu/~henrik/animations/jensen-the_light... ). Keep in mind this was done 14 years ago! It was a genuine wow moment for me back then!
I do empathize with the original article a lot. I used to have/still have a strong fear of failing, especially in intellectual tasks. According to my own introspection this is primary angst that caused/causes me to procrastinate. There are two major references I often go back when I find myself paralyzed.
One being an advise I got from one of my PhD advisors: All creative tasks might appear that it requires enormous amount of courage and effort. But usually it is more like a kitchen sink heaped with a lot of unwashed dishes. Chances are that once you wash one dish, you will end up cleaning the full lot; and you often get a strange form of pleasure while you are performing the task.
The other one is this essay http://www-rohan.sdsu.edu/~psargent/Mills_Intell_Craft.pdf on intellectual craftsmanship by Wright Mills. I do now a days actively collect memories of pure immersion and pleasure I experienced while my craft got exposed and exploited to its potential. The thought of me improving as a craftsman, coupled with these memories is a powerful self motivator to me. The shit feeling I gets when I waste my time is another reference. One of the potent lessons was also that craft can be improved only by dedicating time ( which is pleasurable); and by disassociating the end result and fears. The toughest part is to replay this logic while I find myself slipping into vortex of non productivity, but that is something I can work on and probably in my control.
I just have to say, I love washing dishes. I was convinced by Leo (the author) to start trying to enjoy myself while washing dishes, and now I love it. I don't even use the dishwasher anymore.
The way we approach and do science has evolved drastically ( http://en.wikipedia.org/wiki/History_of_scientific_method ). For example empirical falsifiability which is one of the primary tenant of modern science is less than 100 years old, but forms an essential part on how we do science now a days.
Parent comment's point being; while we may be trying to understand the same principle/phenomena, not only the data available to thinkers that time was very sparse compared to the present; but also the level of rigour applied was of significantly lower standards. While there might be scattered scientific truth in the vedas ( or any other religious document) ; it is insolent to believe that it is good reference manual for scientific knowledge.
As i see, the current science is more rigorous because people are producing lot of crap. So, we made it to be like "if it can't be verified/repeatable its not science". But do we really know for sure ? How many discoveries are being overridden by new discoveries coming from future ?
The amount of data accessible to the people in the past is a lot more when compared to current. Thats why there able to explain things that can't be experienced by our senses. To share such things in the current time, the "can be verified by our senses by a independent vendor ?" rule rejects. So, very few people experience them and bring it down to such a level that every human gets benefited from it.
The division between religion/science is very small, when both are approached using similar thought-process. Its just that some rules reject others. As human we need to approach and find truth for oneself without being biased.
> As i see, the current science is more rigorous because people are producing lot of crap
Good, grief. No!!! It is a way of managing uncertainty and saying something with a precision that is available at a given point of time.
> How many discoveries are being overridden by new discoveries coming from future ?
This is beauty/and USP of science. Every scientific proof is always open for scrutiny and revision in light of new data or discovery ( tenants of falsifiability kick in here). That is, it tries hard NOT to be dogmatic by being provisional. For example, science says that we are confident Higgs Boson exists "accounting for one-in-a-million chance on the contrary" ( 5-sigma).
Let me flip your argument on the converse; success rate at which we could make ground breaking theories [ like evolution, theory of relativity , uncertainty principle ] ( which is standing the test of time for extended period of time) using the scientific method is sheer staggering and amazing. The methodology has accelerated our progress and understanding by leaps and bounds which no alternate system has managed to do so, so far!
> The amount of data accessible to the people in the past is a lot more when compared to current.
I lost you completely here. Can you please elaborate and the rest of the paragraph. ( My belief: If you take 20 random guesses; one of them turned out to be true; it is more likely to be a coincidence than a mystical insight. If on the contrary, the Monte Carlo filter I routinely simulate might just be the most insightfully entity I have encountered ).
> The division between religion/science is very small
Epistemologically they are apples and oranges! Falsifiability is not applicable to religion nor is it is provisional and routinely advocates absolute (and imho dogmatic) reasoning!
> science says that we are confident Higgs Boson exists ( 5-sigma).
Agreed, Science comes from our experience/understanding of things around us by our senses. Try to explain the above Higgs Boson to a blind person who has never seen anything in their life. As long as science explains stuff that can be experienced by the senses, everybody else with similar senses get them.
> Epistemologically they are apples and oranges!
Its all in our thought process. Everything came from our thinking/undertsanding of things around us. It just happened to be that we are closer to prove somethings easily vs others.
> ... using the scientific method is sheer staggering and amazing
> The amount of data accessible to the people in the past is a lot more when compared to current.
Appreciated the hardwork done by all these determined people. How did only few people have access to such knowledge ? In order to find the truth we should not be biased. The reason why people in the older generations might not have shared such knowledge is to prevent mis-use of it, for better of mankind. While we take pride in such innovations.
It vastly depends on the actual science which standard of proof is accepted.
For maths, with a 5-sigma result you can maybe get a mention in the "curiosa" section if it's weird enough. It is certainly not considered a valid mathematical result.
For biology, a 1 sigma result is considered pretty good. And due to experimental restrictions, this is actually more strict than medicine requires.
Many science disciplines work with known-wrong theories. Civil engineering for example, works with pre-Newtonian mechanics (not even "turtle mechanics" : in the best simulations a building stands on ground, which stands on a plate which is magically suspended in a "downward" gravity field, not on a planet).
The idea of "this is the standard of proof for 'science'" is a nice one, but it doesn't exist in any reasonable sense. Only the utilitarian definition sticks : we have 100 standards of proof, and if the theory works (or gets enough money if your cynical) we'll find the standard of proof that allows us to call it science.
Furthermore, there are several inconsistencies in the science underpinning, for example, the Higgs boson discovery. We do not actually have rigorous proofs for constructing even natural numbers by the standards of first-order logic. And second order logic has paradoxes that stand unresolved (there is a lot of research to find something "more flexible" than first-order logic, but stricter than second-order that works, but this research has been going on for more than a century and there are no really good candidates, only really bad ones like the famous failure of the Choice axiom)
The standard model doesn't even contain gravity, so if you're being pedantic you could drop a pen from your desk and claim, correctly, that you've just falsified the entire standard model, or at least proven it's incompleteness.
Less pedantically in the physics itself there is the massive open question. The Higgs field only causes inertia, not gravity. Yet the measure of interaction with the Higgs field of any object we've ever measured matches exactly the value we've got for that same object's gravitic interactions. Does anyone believe this to be a coincidence ? Major open hole there.
Falsification and Incompleteness are two different things. Since we reason about physicals system using the language of mathematics/logic; it has be based on certain axiom which cannot be proved or disproved ( Godel's incompleteness theorem ). Though this renders certain statements inside physical theorem non-provable ; it certainly does translate to every claim made by a proposed theory. Further many aspects of physicals systems can be disproven experimentally. ( It is still in active debate if Mathematics should treated as science per se : http://en.wikipedia.org/wiki/Mathematics#Mathematics_as_scie... )
While the pen falling from a desk do point out to the incompleteness ( non-Godel sense) of the standard model, which is widely accepted ( http://home.web.cern.ch/about/physics/standard-model : last paragraph ), it does not falsify it. Science is full of open holes, and no one knows ( my bet is against) that it will be completely patched up; but it is the best form of reasoning we have in understanding things, and its ongoing goal is to seek explanations that with the least amount of uncertainty possible.
I think we're largely making the same point : that science is largely based on a utilitarian definition of truth. A somewhat more direct way to state that is that scientific truth is simply
"What works for me"
And nothing more.
I do disagree on one point though. The standard model doesn't just "lack" gravity. It describes a world without gravity. Therefore that gravity exists must mean that the standard model is wrong. It describes a universe that is most certainly not the one we live in. I therefore find it hard to describe that theory as true. It is more akin to "currently the best-known least-wrong theory". Even best-known has to be in there since, for example, relativity theory was known long before Einstein got his ball rolling, and Newton's equation was known before the apple fell. So we do likely know about better theories than the standard model, we just currently have no way to distinguish them from either the standard model, or (more likely) the better theories are just failing to get enough attention from well-publicized physicists. Of course, when you don't know exactly which theories are in fact better, their existence doesn't matter.
A PhD system trains you to think about unsolved problems in an given domain deeply with a larger time runway. The end goal is not a tangible product that reaches millions of people, but rather a set of ideas that can take a crack at the unsolved problems in your field in a novel way. A good work should inspire others in the field, and eventually a larger audience to pick them up and expand and build on top of it. To give a small example, a majority of the fundamentals of machine learning was charted out by many, many PhD works over the last 40 years. Implementing a linear classifier is 2 lines of code in 2018, but many Bothans died to bring us this information :-) .
The goals of industry are more immediate. Expect for a privileged few research labs in industry, your work is expected to be monetized, and rightly so. The goal is for you, if you run the business, else your management team to first figure out a problem of high relevance and monetary value. Build products/solutions for that problem, that can be used by someone who is less versed/ambivalent of your technical solutions. Efficacy of solving that particular problem often defines the merit of your contribution.
The fundamental of choosing the PhD or industry should be taking stock of what kind of contribution you want to make as an individual. If it is a few set of ideas to science, which on a later date might become something fundamental in our understanding of the world, then PhD is a good path. If it is a set of contributions towards a product/solution that eases the pain of many users then go into the industry first.