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.)]

 

Miscellaneous: my job situation, the Tatas, and taking a break…

The Diwali is here, already!

This year’s Diwali isn’t going great for me. I am still jobless—without reason or rhyme. It is difficult to enjoy Diwali against that backdrop.


As you know, engineering colleges affiliated to the Savitribai Phule Pune University (SPPU for short) have been telling me that my Metallurgy+Mechanical background isn’t acceptable, even though the rules have changed to the contrary, and say that I now qualify (in my interpretation).

Recently I attended an interview, and it seems like I may be able to obtain a clear-cut answer on my eligibility (i.e. the equivalence of Metallurgy and Mechanical) from SPPU.

The thing is, SPPU has been having no Dean for its Engineering faculty for about a year or more by now, because the Maharashtra state government hasn’t so far undertaken the procedure to elect (or select) the next Dean.

This recent interview which I mentioned above, was for a Principal’s post, and I was short-listed. As is the common practice here, the short-listed candidates were all invited at the same time, and thus, I had an opportunity to interact with these other, senior-level professors.

These senior professors (some of them already active as Principals at other colleges) told me that it isn’t just SPPU, but all the universities in Maharashtra. They all are currently having only an in-charge or acting Dean for their engineering faculties, because the procedure to appoint the next set of Deans, which was due to occur this month (October) has once again been postponed by yet another year.

Policy decisions such as the Metallurgy and Mechanical equivalence at SPPU have been pending, they told me, because the acting Dean can easily say that he has no powers to do that. Though the other universities are clear that I would qualify, if a genius running an engineering college under SPPU thinks that I don’t, then the matter normally goes to the Dean. If the Dean is not official, if he is only acting, he doesn’t want to take “risk,” so he takes no decision at all. Not just the equivalence issues, there are certain other policy decisions too, which have been pending, they told me. The in-charge Deans have been processing only the routine work, and not taking any policy decisions. The next set of Deans were expected to get appointed by June 2016, and then, after postponement, by October 2016. (“achhchhe din!”)

Now that the appointments have been officially postponed by one whole year (“achhchhe din,” again!), the colleges themselves have begun going to the universities for obtaining the professor’s approvals, arguing that faculty approvals is a routine matter, and that they cannot properly function without having approved faculty.

Thus, the university (SPPU) has begun appointing panels for faculty interviews. There has been a spate of faculty recruitment ads after the current semester got going (“achhchhe din!”).

The particular interview which I attended, these other candidates informed me, was with a University-appointed panel—i.e., of the kind which allows approvals. (Otherwise, the appointments are made by the affiliating colleges on their own, but only on a temporary, ad-hoc basis, and therefore, for a limited time.)

Please note, all the above is what I gathered from their talk. I do not know what the situation is exactly like. (Comments concerning “achhchhe din!,” however, are strictly mine.)

But yes, it did turn out that the interview panel here was from the university. Being a senior post (Principal), the panel included both the immediately past Dean (Prof. G. K. Kharate) and the new, in-charge Dean (Prof. Dr. Nerkar, of PVG College, Pune).

During my interview, if the manner in which Prof. Kharate (the past Dean) now said things is any indication, it means that I should now qualify even in the SPPU. This would be according to the new GR about which I had written a few months ago, here [^]. Essentially, Prof. Kharate wondered aloud as to why there was any more confusion because the government had already clarified the situation with the new rules.

I took that to mean that I qualify.

Of course, these SPPU geniuses are what they are, and therefore, they—these same two SPPU Deans—could very well say, in future, that I don’t qualify. After all, I didn’t ask them the unambiguous question “With my Metallurgy background, do I qualify for a Mechanical Engineering (full) Professor’s job or not? Yes, or no?;”  and they didn’t then answer in yes or no terms.

Of course, right in the middle of an on-going job interview couldn’t possibly have been the best time and place to get them to positively confirm that I do qualify. (Their informal indications, however, were clearly along the lines that I do qualify.)

Now that the Diwali break has arrived, the colleges are closed, and so, I would be able to approach Prof. Dr. Nerkar (the currently acting/in-charge Dean) only after a week or so. I intend to do that and have him pin down the issue in clear-cut terms.

At the conclusion of my interview, I told the interview committee exactly the same thing which I told you at the beginning of this post, viz., that this Diwali means darkness to me.

But yes, we can hope that Prof. Dr. Nerkar would issue the clarification at least after the Diwali. If not, I intend to approach Prof. Dr. Gade, the Vice-Chancellor of SPPU. … I could easily do that. I am very social, that way.

And, the other reason is, at the university next door—the Shivaji University—they did answer my email asking them to clarify these branch-equivalence issues. The SPPU is the worst university among the three in the Western Maharashtra region (the other two being, the University of Mumbai and the Shivaji University Kolhapur). [I want to teach in Pune only because it’s my home-town, and thus convenient to me and my family, not because SPPU’s standards are high.]

Anyway, I now do have something in hand to show Prof. Dr. Gade when I see him—the letter from the Shivaji University staff. … At the Shivaji University, I didn’t have to go and see anyone in person there—not even the administrative staff let alone the acting Dean or the Vice-Chancellor. The matter got clarified just via a routine email. There is a simple lesson that SPPU may learn from the Shivaji and Mumbai universities, and under Prof. Dr. Gade, I hope they do.

… Of course, I do also hope that I don’t have to see Prof. Dr. Gade (the Vice-Chancellor). I do hope that meeting just Prof. Dr. Nerkar (the in-charge Dean) should be sufficient.

If they refuse me an appointment, I will get even more social than my usual self—I will approach certain eminent retired people from Pune such as Dr. Bhatkar (the founder of C-DAC) or Dr. Mashelkar (the former Director General of CSIR, India).

Here is a hoping that I don’t have to turn into a social butterfly, and that soon after Diwali, the matters would get moving smoothly. Let’s hope so.

And with that hope in my heart, let me wish you all a very happy and prosperous Diwali. … As to me, I will try to make as much good of a bad situation that I can.


Still, I don’t find myself to be too enthusiastic. I don’t feel like doing much anything. [In a way, I feel tired.] Therefore, I am going to take a break from blogging.

I have managed to write something more on the concept of space. I found that I should be able to finish this series now. I had begun it in 2013; see here [^].

Concepts like space and time are very deep matters, and I still have to get enough clarity on a few issues, though all such remaining issues are relatively quite minor. I should be able to get through them in almost no time.

From the new material which I have written recently, I guess it would be enough to write just one or two posts, and then the series would get over. What then will remain would be mostly polemics, and that part can be taken on the fly whenever the need to do so arises.

I may also think of giving some indications on the concept of time, but, as I said, I find myself too lacking in enthusiasm these days. Being jobless—despite having the kind of resume I have—does have a way of generating a certain amount of boredom in you, a certain degree of disintegration at least to your energy and enthusiasm, even if not to your soul.


So, let’s see. Let the Diwali vacations get over, and I should come back and resume my blogging, telling you what all transpired in my meeting/interaction with the in-charge Dean, and the related matters.


Since I am not going to be blogging for some time, let me note a couple of notable things.

One, the US Presidential elections. I am not at all interested in that. So let me leave it aside.

Two, the Tata Sons issue. It does interest me a bit, so let me write down a bit on it.


I was not as surprised as some of the newspaper editorials and columns say they were. The days of JRD are long gone. The Tatas already were a changed company when Cyrus Mistry took over from Ratan Tata.

Once I returned from the USA in 2001, despite my resume, I never got a chance with the new Tatas (either at TRDDC or at TCS). Such a thing would have been unthinkable during JRD’s times. … Even keeping it aside, what all I observed about the Tatas over the past 1.5 decades was enough for me not to be at all surprised by something like the current fiasco.

No, Prof. Pratap Bhanu Mehta, reading things from where I sit, the Tata fiasco doesn’t do any significant harm to the social legitimacy of Capitalism in India. People—common people—have long ago observed and concluded what had to be. If what the common people think were to be caricatured, it would look like the position you ascribe to the “cynics”. But no, IMO, this position isn’t cynical. To carry realistic impressions about hallowed icons is not quite the same as being a cynic.

Yes, as Harsh Goenka astutely pointed out in his comment in today’s ToI, Ratan Tata’s tenure coincided with the semi-liberalization era: 1991–2012. Whenever you come to compare Ratan Tata with Cyrus Mistry, you cannot overlook that broad context.

I have always thought that JRD left too big shoes for any one to fill in. But, with due respect to Ratan Tata, I still would have to say that no one could possibly entertain thinking in similar terms, when it comes to Ratan Tata’s retirement.

Looking at the facts and figures reported this week, I don’t think Mistry was doing a lousy job. Reading through his letter, I in fact marvel at how well he understood his job—and for this reason, I speculate that he must have been doing his job pretty well. …

Realize, the letter was written within a day or two after an unceremonious removal from the top post of a 100+ years old Indian icon, a $100 billion behemoth. Seen against this backdrop, the letter is extraordinarily restrained; it shows an unusual level of maturity. To expect any more “restraint” is to actually confess ignorance of such basic things as human nature and character. (Sadhus, let me remind you, are known to kill each other in their fights at the Kumbh Mela, just for the priority in taking the Shahi Snaan. Keep that in mind the next time you utter something on nobility of character and culture.)

And yes, I also had come to think that the Nano project was doomed—I just didn’t have the sales and profitability figures, which got reported only today. My reasons were simple; they were purely from an ordinary consumer’s point of view. If you are selling the Nano at around Rs. 2.5 lakhs, just think of the alternatives that the consumer has today: you could get a used car in a “good enough” condition, not just Maruti Alto but even a somewhat more used Toyota Innova, at roughly the same price.

Anyway, I don’t understand these corporate matters much, so let me shut up.


But, yes, knowing the house of Tatas and their brand managers, I can predict right away that in the near future, you are going to see the Tatas announce a product like “Tata Quantum Dot,” or “Tata Silicon Dot,” or something like that. … Why do I think so?

I started writing on quantum mechanics, and roughly around the same time, the cable-less Internet, based on the electromagnetic waves (mobile, Wi-Fi) was getting going in India. So, the Tatas came out with the Tata Photon. Yes, “Photon”. The Tata Photon. … It meant nothing more than the usual Internet dongle (2G, and then 3G) that everybody else was already supplying anyway. (And the Tata Photon never worked too well in areas other than in the Mumbai city.)

Then, the USA was abuzz with the catch-words like nano-technology, and the Tata brand managers decided to do something with that name, and thus came the Tata “Nano.” By now, every one knows what it means.

Today, the USA and other countries are abuzz with words like “Quantum Supremacy” and things like that. You can only expect some Tata brand managers to latch on to this buzzword, and launch a product like, say, Tata Quantum Dot or Tata Silicon Dot—or both!

Tata Silicon Dot, I predict, would signal the arrival of the house of Tatas into the business of supplying the sand required for civil engineering construction.

Tata Quantum Dot, on the other hand, would mean that the house of Tatas had taken an entry into the business of plastic dart toys. Or, the business of the “bindi”s that ladies wear. That is what the house of Tatas would mean by the name Tata Quantum Dot.

And here our policy analysts think that something happening to the house of Tatas is going to affect the credibility or social legitimacy of Capitalism itself in India! Oh wow!!

Ummm…. Does any policy research center in India have any data on the proportion of the private business in the overall Indian economy (including both the organized and the unorganized sectors) over the years, say starting from 1930s? Also, the quantum of the government expenditure in the Indian economy, and its proportion in the national GDP over the same period? Would they care to share it, please? Or is it that they don’t have to look at such data for their policy research purposes? … As to me, I have been on the lookout for data like that for quite some time now, but never could see it compiled anywhere. That’s why the request. Please drop me a line if you spot a reliable source.

OK, bye for now.


A Song I Like:

Since I won’t be blogging for a while, let me give away the “other” song right away, I mean the song which had somehow happened to strike me as being similar to the song “too laali hai savere waali”; see the Song I Like section here [^]. This other song is:

(Hindi) “bhigee bhigee raaton mein…”
Music: R. D. Burman
Singers: Kishore Kumar, Lata Mangeshkar
Lyrics: Anand Bakshi

I take the “raaga” of the earlier song (“too laali hai”) as “pahaaDee”—or at least that’s what I got from an Internet search. The “raaga” of the current song (“bhigee…”) isn’t listed at any Web site. Assuming it’s not “pahaaDee” (or a variant on that), the question becomes, why the two songs might have struck at least somewhat similar to me—why, humming one song, I very naturally and casually happened to remember the other song.

It would be interesting to see if Data Science can be used to spot (and quantify) similarities in songs. The traditional music theory puts too much emphasis, IMO, on “raaga” alone. But there can be other bases for similarities, too. The sound patterns of musical pieces, I think, don’t get exhaustively (and at times not even essentially) characterized by the idea of the “raaga” alone. Talking of these two songs in particular, the similarity I caught might have been connected with certain ups and downs in notes with a somehow similarly sounding tempo. The style of the tunes sounds similar. Guess Data Science might be able to shed some light on things like that…. It would be interesting, to look into that, no? That’s what I had thought…

I mean, I had thought. … But then, these days, as I said, I am unable to work on this topic, too…  I just don’t have any enthusiasm left. Honest. I somehow finished this post, only because I won’t be posting for a while…

So, there. Bye for now, take care, and best!


[E&OE]