Machine “Learning”—An Entertainment [Industry] Edition

Yes, “Machine ‘Learning’,” too, has been one of my “research” interests for some time by now. … Machine learning, esp. ANN (Artificial Neural Networks), esp. Deep Learning. …

Yesterday, I wrote a comment about it at iMechanica. Though it was made in a certain technical context, today I thought that the comment could, perhaps, make sense to many of my general readers, too, if I supply a bit of context to it. So, let me report it here (after a bit of editing). But before coming to my comment, let me first give you the context in which it was made:


Context for my iMechanica comment:

It all began with a fellow iMechanician, one Mingchuan Wang, writing a post of the title “Is machine learning a research priority now in mechanics?” at iMechanica [^]. Biswajit Banerjee responded by pointing out that

“Machine learning includes a large set of techniques that can be summarized as curve fitting in high dimensional spaces. [snip] The usefulness of the new techniques [in machine learning] should not be underestimated.” [Emphasis mine.]

Then Biswajit had pointed out an arXiv paper [^] in which machine learning was reported as having produced some good DFT-like simulations for quantum mechanical simulations, too.

A word about DFT for those who (still) don’t know about it:

DFT, i.e. Density Functional Theory, is “formally exact description of a many-body quantum system through the density alone. In practice, approximations are necessary” [^]. DFT thus is a computational technique; it is used for simulating the electronic structure in quantum mechanical systems involving several hundreds of electrons (i.e. hundreds of atoms). Here is the obligatory link to the Wiki [^], though a better introduction perhaps appears here [(.PDF) ^]. Here is a StackExchange on its limitations [^].

Trivia: Kohn and Sham received a Physics Nobel for inventing DFT. It was a very, very rare instance of a Physics Nobel being awarded for an invention—not a discovery. But the Nobel committee, once again, turned out to have put old Nobel’s money in the right place. Even if the work itself was only an invention, it did directly led to a lot of discoveries in condensed matter physics! That was because DFT was fast—it was fast enough that it could bring the physics of the larger quantum systems within the scope of (any) study at all!

And now, it seems, Machine Learning has advanced enough to be able to produce results that are similar to DFT, but without using any QM theory at all! The computer does have to “learn” its “art” (i.e. “skill”), but it does so from the results of previous DFT-based simulations, not from the theory at the base of DFT. But once the computer does that—“learning”—and the paper shows that it is possible for computer to do that—it is able to compute very similar-looking simulations much, much faster than even the rather fast technique of DFT itself.

OK. Context over. Now here in the next section is my yesterday’s comment at iMechanica. (Also note that the previous exchange on this thread at iMechanica had occurred almost a year ago.) Since it has been edited quite a bit, I will not format it using a quotation block.


[An edited version of my comment begins]

A very late comment, but still, just because something struck me only this late… May as well share it….

I think that, as Biswajit points out, it’s a question of matching a technique to an application area where it is likely to be of “good enough” a fit.

I mean to say, consider fluid dynamics, and contrast it to QM.

In (C)FD, the nonlinearity present in the advective term is a major headache. As far as I can gather, this nonlinearity has all but been “proved” as the basic cause behind the phenomenon of turbulence. If so, using machine learning in CFD would be, by the simple-minded “analysis”, a basically hopeless endeavour. The very idea of using a potential presupposes differential linearity. Therefore, machine learning may be thought as viable in computational Quantum Mechanics (viz. DFT), but not in the more mundane, classical mechanical, CFD.

But then, consider the role of the BCs and the ICs in any simulation. It is true that if you don’t handle nonlinearities right, then as the simulation time progresses, errors are soon enough going to multiply (sort of), and lead to a blowup—or at least a dramatic departure from a realistic simulation.

But then, also notice that there still is some small but nonzero interval of time which has to pass before a really bad amplification of the errors actually begins to occur. Now what if a new “BC-IC” gets imposed right within that time-interval—the one which does show “good enough” an accuracy? In this case, you can expect the simulation to remain “sufficiently” realistic-looking for a long, very long time!

Something like that seems to have been the line of thought implicit in the results reported by this paper: [(.PDF) ^].

Machine learning seems to work even in CFD, because in an interactive session, a new “modified BC-IC” is every now and then is manually being introduced by none other than the end-user himself! And, the location of the modification is precisely the region from where the flow in the rest of the domain would get most dominantly affected during the subsequent, small, time evolution.

It’s somewhat like an electron rushing through a cloud chamber. By the uncertainty principle, the electron “path” sure begins to get hazy immediately after it is “measured” (i.e. absorbed and re-emitted) by a vapor molecule at a definite point in space. The uncertainty in the position grows quite rapidly. However, what actually happens in a cloud chamber is that, before this cone of haziness becomes too big, comes along another vapor molecule, and “zaps” i.e. “measures” the electron back on to a classical position. … After a rapid succession of such going-hazy-getting-zapped process, the end result turns out to be a very, very classical-looking (line-like) path—as if the electron always were only a particle, never a wave.

Conclusion? Be realistic about how smart the “dumb” “curve-fitting” involved in machine learning can at all get. Yet, at the same time, also remain open to all the application areas where it can be made it work—even including those areas where, “intuitively”, you wouldn’t expect it to have any chance to work!

[An edited version of my comment is over. Original here at iMechanica [^]]


 

“Boy, we seem to have covered a lot of STEM territory here… Mechanics, DFT, QM, CFD, nonlinearity. … But where is either the entertainment or the industry you had promised us in the title?”

You might be saying that….

Well, the CFD paper I cited above was about the entertainment industry. It was, in particular, about the computer games industry. Go check out SoHyeon Jeong’s Web site for more cool videos and graphics [^], all using machine learning.


And, here is another instance connected with entertainment, even though now I am going to make it (mostly) explanation-free.

Check out the following piece of art—a watercolor landscape of a monsoon-time but placid sea-side, in fact. Let me just say that a certain famous artist produced it; in any case, the style is plain unmistakable. … Can you name the artist simply by looking at it? See the picture below:

A sea beach in the monsoons. Watercolor.

If you are unable to name the artist, then check out this story here [^], and a previous story here [^].


A Song I Like:

And finally, to those who have always loved Beatles’ songs…

Here is one song which, I am sure, most of you had never heard before. In any case, it came to be distributed only recently. When and where was it recorded? For both the song and its recording details, check out this site: [^]. Here is another story about it: [^]. And, if you liked what you read (and heard), here is some more stuff of the same kind [^].


Endgame:

I am of the Opinion that 99% of the “modern” “artists” and “music composers” ought to be replaced by computers/robots/machines. Whaddya think?

[Credits: “Endgame” used to be the way Mukul Sharma would end his weekly Mindsport column in the yesteryears’ Sunday Times of India. (The column perhaps also used to appear in The Illustrated Weekly of India before ToI began running it; at least I have a vague recollection of something of that sort, though can’t be quite sure. … I would be a school-boy back then, when the Weekly perhaps ran it.)]

 

Trivia Like the M. F. Hussain Controversy and the Women’s Reservation Bill

0. With this post, I once again resume blogging…

First, I need to quickly get a few things out of my system before I am ready to write on some of the things I have wanted to write about. … So, here we go with the more trivial (but far more discussed) matters first…

(1.) About the M. F. Hussain Controversy…

There was a spike of discussions concerning this particular controversy about one/two weeks ago. So many interesting angles got thrown up that it would be impossible to even summarize them. I felt like jumping in, but instead, just kept on reading on the ‘net and otherwise, to get the “lay of the land” before I wrote. In a way, this turned out to be a good decision.

After all, I did find a very highly quotable position post which explains most of what I had wanted to say anyway. By that, I mean the post on the topic by Dr. Atanu Dey, here [^]. Please do read it. Highly recommended.

Not that I agree with every nuance of every point he states. Speaking in overall terms about his blogging about other matters too, I suspect that there might be a difference among us in that I might look at something from a moral/judgmental viewpoint whereas he wouldn’t, necessarily. That hardly matters here, though…

Here, I find his ability to think in principles, and the straightforward way in which he puts his thoughts, marvelous! And I completely agree with all the essential points of this post of his.

Just a couple of points I shall add to what Dey has already said.

(1.a) Dey says that “[he is] not much of a paintings person, anyway.” But I am, to a certain extent. And used to be one to a major extent about two-three decades ago. So, I can add a bit about this matter.

The question I very briefly address here is: how great is Hussain, as a painter (i.e. artist)?

Even a casual glance at his paintings would tell you that he has an extraordinary mastery over the line. He is an abstract painter—which, to my mind, generally speaking, doesn’t qualify as art to begin with. This applies as much to Hussain as also to Souza, or Gaitonde, or Anjali Menon, or even Sujata Bajaj, or anyone else of their kind—which means, about 99% of today’s painters: they, too, are not artists.

But keeping this aside for a moment, the next question is: Doesn’t he show at least some elements of great art in his work?

Here, I think, as a craftsman, his defining skill is not at all light and perspective, certainly not color, nor even subject, but it’s: his line. His painting unmistakably show that had he chosen higher goals, he would have made for a recognizably great artist—and, despite spending 95 years of his life, he still has not managed to even become an artist let alone a great one.

But why do I say it’s the line which really defines his craftsmanship? Just look at the lines that define the contours of his horses, and the women he paints. His line is capable of bringing to life the sheer life power, the very unruly dynamic, of a horse. Just one apparently careless stroke of a brush in the right place while drawing the eye of a horse, and that raw, unruly energy of the horse begins to jump at you. Similarly, consider the fact that despite carrying the crudeness of the abstract technique, his straight lines still perfectly capture the contours of the feminine form, whenever he manages to slip-in to the remnants of the better elements of the technique he must have been taught at the JJ School of Arts.

So, here is a very curious phenomenon. You have a gifted craftsman—at the level of the line. But this same guy, then, refuses to use that gift to paint a picture—i.e. to create a work of art. Instead, he uses his more abstract powers to mangle the elements like the objects making up those lines, the color and the perspective etc, deliberately disorients them all, throws them together to deliberately create incoherence or even un-intelligibility in his work of “art.”

Consider its counterpart in other forms of art, for example, literature, for example, poetry. What Hussain’s approach would yield is not a poem but something like a poem. Of course it would be called a “free verse.” But the matter doesn’t end there—it gets worse. What Hussain would give you would be a collection of in-principle disconnected bunch of lines, some phrases of which being extraordinarily brilliant on counts such as drama, innovation of expression, metaphor, imagination, etc. Mind you, the brilliance would be restricted only to phrases, not even to lines—the mangling would begin right at that level. And, the lines, taken as sequence, would all be disjointed, hinting at something which, in principle, cannot at all be known, not in toto. The hints themselves could at times be grotesque, at other times sly, at other times profane (and this term is to be taken in its objective sense, not necessarily in connection with this religion or that)… You could, if you try, easily locate Hussain’s parallels in modern “poetry” too. The point isn’t that. The point is to convey what Hussain really is like, when taken as a painter. Namely, that he isn’t one.

(It would be an error to compare Hussain’s paintings with the strokes produced by a student studying at a school for the mentally retarded—the first has the ability to do better, the second doesn’t, and the deliberateness of the rebellion against integration is the crucial difference.)

(1.b) Another point that many people seem to have missed is this. I ran across a court judgment that did agree with the opponents of Hussain in all other points. However, it refused to try Hussain on a point of legal technicality. And, that brilliant piece of the legal technicality was supplied by the current Central Government of India minister Kapil Sibal. … The less I say, the better it will be to my health and life…

(1.c) Nevertheless, we must stop and ask ourselves one question. If merely brilliance in respect of an element like the line-work can be enough to qualify a guy to be counted as an artist, even when ample evidence from his art-work as well as his interviews exists that he has deliberately followed a policy of working against proper integration as required by a proper piece of art, then, following the same standards, why not also consider those millions of anonymous Indians whose “work” adorns the walls of all our public urinals to be artists in their own right, too? [And, I deliberately use the word “urinal” rather than “toilet” or “rest room,” because only the former can adequately convey the strength of the stink in question.] Why not decorate also them with those Padma awards?

Any answer, Delhi “intellectuals”? Rich Bombay “businessmen” patrons of Hussain’s “art”?

(2) About the Women’s Reservation Bill

First of all I want you to note that here I am going against many politicians I otherwise respect, first and foremost, Sharad Pawar. Also, many other politicians I fear. … The reader must excuse me here; there would be too many to name them to list them individually. …

The best commentary—and the only reasonable one—that I saw in print or on monitor, came from one Mr. Parsa Venkateshwar Rao, Jr., in a column he wrote for DNA, here [^]. The only other media/blog to highlight it (in my limited browsing) was “Churumuri,” here [^].

… As usual, at least one qualification. What Rao calls “politics of identity,” I would call such things as “politics of narrowness/of insularity/of divisiveness.”

And, here’s the extraordinarily brilliant part of Rao’s comment, expressed so tersely but so well:

…Women’s reservation bill too is supposed to promote gender equality but what it really does is create yet another special interest. And society is turned into a bureau of cubbyholes. And the power of the State is increased yet again. …

Thank you for saying it, Mr. Rao!

To Swamy of Times of India, regarding his today’s column. Nope, Swamy, you don’t get it right. Hmm…

Back to basics. There are three pillars of a nation state: (i) legislative (L for short), (ii) judicial (J for short), (iii) executive (E for short). In India, the mangling of the E branch began right with the original version of the Constitution (C for short)—it’s just for five years, no principled, i.e. unreserved respect/acknowledgment of the individual rights, etc. As such, the J, if pushed to the wall, would have been helpless, in principle. For the aforementioned reason (viz. the absence of an explicit ack. of the Individual Rights), the Constitution always had been sufficiently vague—i.e. weak–that if L grew, it could not only overpower the E but also effectively restrict the J in various indirect ways. Enter the mixed ideals of Nehruvian socialism. L had become powerful. In Indira’s semi-dictatorship, it changed C and systematically weakened L and then also J. With the fall of the Berlin wall and the collapse of USSR, there seemed to be a reversal of sorts, but it was necessarily doomed because C was weak to begin with anyway so there was poor theory, and, in fact, Indira’s years had weakened the level of the public discourse to such low levels. So much so that on this issue of reservation, all major parties—Congress, BJP, and the communists—they all agree. (Also all the rest of them: they simply fight for greater reservation—not less). Ok.

Against this background what would this Bill do? There are certain implicit grounds for negotiating any kind of agreement in any free society. Due to a better past—and interactions with better countries like the USA and UK—despite the systematic abuse of the above sort and all the weaknesses of the Constitution and legal codes, the common implicit grounds in India actually tend to be better. This is the reason why gays (as much the “chhakke”s as the more hype urban ones) could at all live without having cases slapped against them. This is the reason why in Maharashtra, the ANS-sponsored Bill gets halted. (The reason I oppose it: What standards would permit an ordinary police officer to distinguish between proper private practice of religion and blind faith as prescribed by ANS?) This is the reason why business can at all in fact function even if enough legal codes exist that in theory it would be impossible to run a business without breaking some or the other legal code. That implicit ground is important.

In a country with as huge illiterate, semi-literate, and literate-but-uneducated population as India, a country where to run the elections you have to use symbols—not candidate’s names—it does matter a lot what kind of signals we project to all those people.

When reservations in jobs came into force, it actually did not matter to large parts of population: most of the labor is in agriculture or unorganized sector, and even in organized sector, job reservations applied only to government jobs, not private. It was bad, but it was limited in terms of impact. When the Constitution got mangled almost with each successive amendment (some of which being more deeply mangling than the others), it rather affected the upper echelons of the society—their effects on that implicit negotiating grounds that I alluded to above was at least initially minimal; in any case, their effects would have to slowly trickle and diffuse.

But when you introduce a Reservation Bill of this sort—whether on the caste basis, gender, or any other, it matters not in principle—what you do is that you not only mangle the L branch of the government out of its shape, but, since the common illiterate man, right since the Freedom Movement, has always been an active part of the political process, you also affirm to him that divisive agendas like that are alright so long as ratified by an overwhelming majority, as led by the likes of Sonia Gandhi and Sharad Pawar and Advani and others.

In other words, you affect that implicit understanding of what kind of state one lives in, for that common illiterate man. In essence, you tell him: It’s perfectly “sarkari” to be prejudiced against any innocent man. It is perfectly OK to be prejudiced. It is perfectly OK to be so even at the level of elections for law-makers. It is perfectly OK to follow the blind politics of special interest groups.

The first implications of this kind of a message has already emerged, in the form of the opponents to the Bill. … And, Sharad Pawar, and Sonia Gandhi, and Nitin Gadkari, and Brinda Karat and their lesser colleagues all find a cause to celebrate for. What a tragedy!!

. . . . .

[BTW, if someone from NCP or Indira Congress comes and asks me (which is very unlikely), rather than give them a lecture on principles and all, I am just going to be a bit smart and raise a few points in turn: (i) Why did “Sakal” stop carrying the news of new PhD awards precisely around the time I was awarded one—and why does, through other columns, it does sometimes (even if rarely) does cover the news of other PhDs… Is “Sakal” ashamed of the kind of work I had submitted for my PhD? (ii) Why did I not get that job in COEP—even after my PhD defence? [^]  (iii) Why did the IIT Bombay Conference ICCMS09 reject my paper (citing such flimsy grounds that I had used the grammatical first person while writing the abstract)? Who gave them the encouragement to behave thus anti-intellectually? (iv) Why did CERN reject my paper?. I think this might keep them busy for a while… We could discuss principles and all later on…]

– – – – –

Things I Wanted to Write About

Now that the trivia are out of my system, here is a word about what I have been wanting to write about for quite sometime, and may write in near future (not necessarily in the next post):

On the political side: The magnitude of the black money kept abroad by Indians, Why no Maharashtrian could become a PM thus far.

And, then, of course, Physics: A simple but important example illustrating how, in Physics, it is impossible to get rid of certain basic assumptions delineating the nature of your theory.

– – – – –

A Couple of Songs I Like

1. (Marathi) “kase kase, haasaayaache…”
Music: Hridaynath Mangeshkar
Singer: Asha Bhosale
Lyrics: Aarati Prabhu

2. (Hindi) “jaaye to jaaye kahaan…”
Singer: Talat Mahmood
Music: S. D. Burman
Lyrics: Sahir Ludhianwi

PS: As usual, I might edit/streamline this post a bit, later on…