Before we come to my New Year’s Resolutions for the last year and then for the new year, do take a moment to check out the poll I’ve posted on Twitter, and see if you wish to respond to it. I would be grateful to you if you do. The tweet in question is the following one:
The poll gets over within a day: On 01 January 2021 (all time-zones in the world) + a few hours.
1. A quick review of my blog posts in the year 2020:
Not counting the present post, I made in all 26 posts overall during the year 2020. So, on an average, there was one post every fortnight.
This statistic is somewhat misleading. My posts are always much longer than those of your average blogger (i.e. among those few left who still blog!) My posts are almost always > 1.5 words in length, often > 3 k words, and many times >= 5k words. Also, my last year’s NYRs had actually included making even fewer posts!
Let me pick out the more important posts I made this year:
05 February 2020: Equations in the matrix form for implementing simple artificial neural networks [^]
Don’t underestimate the amount of effort which has gone into writing the document mentioned in this post. The PDF document itself was uploaded at my GitHub account, here [^]. (Let me now make it available also from this blog, see here [^]). In this document, I have worked out each and every equation in a consistent, matrix-based format. Although the title says: “simple” neural networks, that word is somewhat misleading. Even the backward passes have been covered, in 100% detail, for all the fundamentally important layers. These same layers are used even in the most modern Deep Learning networks.
13 April 2020: Status: 99.78 % accuracy on the MNIST dataset (World Rank # 5, using a single CPU-only laptop) [^]
The Covid lockdowns had begun already. I had been on the lookout for jobs for more than a year by then. But now I knew that due to the lockdowns, no further interview calls are going to come in—not for me anyway. So, I had a lot of time at hand. I had just written the document on the equations of ANNs/DL. So, the knowledge was fresh. So, I decided to put to use some of my research ideas, with the foremost benchmark of Deep Learning. I broke through in World’s Top 10! (No Indian had ever been in top 20, perhaps even in top 50. That includes Indians from IITs and those gone abroad for higher studies/research/jobs/startups.)
11 May 2020: Status update on my trials for the MNIST dataset [^]
Continuation of the above work. I raised my performance by a tiny but significant 0.01%. I also briefly mention the kind of tricks I tried.
10 June 2020: Python scripts for simulating QM, part 2: Vectorized code for the H atom in a 1D/2D/3D box. [Also, un-lockdown and covid in India.] [^]
This post also mentioned some of the cutting-edge work I had done earlier, in software engineering—like, in writing a yacc-like tool, given just abstract grammar of a computer language.
18 July 2020: On the Bhagavad-Geetaa, ch. 2, v. 47—part 1: कर्मण्येवाधिकारस्ते (“karmaNyevaadhikaaraste”) [^]
This post still remains rather rough. It could do with some good editing. But the real meaning of the Sanskrit phrases aren’t going to change. Check it out if you think you know this famous verse from the Gita.
31 July 2020: “Simulating quantum `time travel’ disproves butterfly effect in quantum realm”—not! [^]
An unplanned post. I never meant to write anything on QM from this angle, but a new paper came up, and it was hyped rather too much. And that was a relative statement. It was hyped even more than the super-high type of hype which has always been an old normal in these fields of Foundations of QM and Quantum Computers
09 August 2020: Sound separation, voice separation from a song, the cocktail party effect, etc., and AI/ML [^]
This was my personal, opinionated, take on some AI/ML-based products/services.
23 August 2020: Talking of my attitudes… [^]
My answers to a questionnaire. This post should be of interest to those who don’t want all the details of my new approach, but just want to know how I view the various issues and interpretations concerning QM
17 October 2020: Update: Almost quitting QM… [^]
I try to be the worst possible critic of my new approach to QM. So, naturally, doubts like these can easily come up. It’s just that once I notice such things, I deal with them.
08 November 2020: Updates: RSI. QM tunnelling time. [^]
Another unplanned post. It covers a very impressive line of experimentation in QM, its coverage in other blogs, as also my own comments. In particular, I thought that this piece of work is likely to be nominated for a physics Nobel too (and gave my reasons for the same).
09 December 2020: Some comments on QM and CM—Part 1: Coleman’s talk. Necessity of ontologies. [^]
19 December 2020: Some comments on QM and CM—Part 2: Without ontologies, “classical” mechanics can get very unintuitive too. (Also, a short update.) [^]
Another couple of unplanned posts. I took this opportunity to present something on my (revised) views of ontologies in physics.
The above posts pretty much capture the two issues which kept me pre-occupied during 2020—Data Science, and Foundations of QM.
Other posts were relatively more topical (like updates), or not so important (though I might have made some good points in them, e.g., this 16 June 2020 post: The singularities closest to you [^].
Apart from this blog, I also made a lot of tweets. Off-hand, these included: Comments on pop-sci articles and videos; comments on papers from QM; comments on developments in Data Science (e.g. the big news related to the Protein-Folding problem), etc. Also, longer twitter-threads (up to 9 or 10 tweets long, not longer) mentioning my thoughts on various topics, e.g., my ideas on relation of maths with reality and physics, finer points related to my ongoing development of my new approach, etc.
Comments at others’ blogs:
I’ve been making many comments at others’ blogs, and some of them have included my spontaneous/ live/ latest thoughts too. However, I don’t want to go through all of that at this point of time. Most of these times, I have saved these spontaneous responses, and I use them in my R&D too, especially of QM and Data Science.
2. A quick review of my last year’s resolutions (for the year 2020):
The resolutions I made last year, are here: My new year’s resolutions—2020 edition [^]
How did I fare on those? Let me jot down in brief:
2.1. Review of last year’s NYRs: Data Science:
2.1.1: A set of brief notes
I did write the “Equations in the matrix form” document; see the comment above.
I had actually planned to write a set of “of brief notes (in LaTeX) on topics from machine learning.” I didn’t. Two reasons:
- Reason 1: Looking at the way the IT industry people treated my applications through job sites, I came to realize that publishing such notes was likely to push the IT industry people in considering me only for the training jobs.
- Reason 2: I achieved 99.78% accuracy (see above). Writing notes, just to show my understanding, naturally took a back-seat. Now I was actually putting to use the knowledge in research too, not just compiling it in a neatly processed form (as in the Equations document mentioned above).
2.1.2. New ideas:
I was going to develop new ideas concerning ML/DL/AI, and perhaps publish a paper.
Well, I did develop my ideas, at least w.r.t. the image data problems, as noted above. However, I have decided to withhold publications. (I think that so far, I’ve been able to hold also the network-hacking/screen-grabbing and bitmap-shipping (remember Citrix?) folks at the bay. So, I think I should be able to publish some time later.
2.2. Review of last year’s NYRs: Quantum Mechanics:
Yes, I worked on all the points noted in the last year’s NYRs, and a lot more.
In particular, my development during the year threw some completely new conceptual issues to me, and I dealt with them. In the process, my understanding of QM became even better and deeper. However RSI struck right around the same time, which was yet another unexpected development. As a result, I could not implement in code all my new ideas and see them in action (and verify them!).
2.3. Review of last year’s NYRs: Health etc.:
I had resolved for going “for walks (30 minutes) on at least six days each month”, and to “do surya namaskars on at least two days a week”.
As to walking, it soon enough got ruled out due to Covid. As to “soorya namaskaara”s, I did try to keep up with “at least 12 namaskaar’s every day”, but I couldn’t. And, once there was a break, it soon became a complete break.
On the positive side, I didn’t have any alcoholic drink (not even wine) after 16th March 2020. … No, it wasn’t a religious thing. I just avoided going out, due to Covid, and soon later, there were the lockdowns. Then, even when the wine shops reopened, I just thought of continuing until (a) the year-end, or (b) a validation-through-simulations of my new ideas in QM, whichever came first. As a result, I didn’t.
I plan to have a bit tonight, as an exception. I also plan to have a celebration with drinks once I achieve the milestones noted for the new year (2021).
2.4. Review of last year’s NYRs: Blogging:
I had resolved to reduce blogging.
However, there were enough developments that I ended up doing 26 posts instead of the planned 12 to 20 posts.
2.5. Review of last year’s NYR’s: Translations:
I was going to attempt translating उपनिषद (Upanishad). Well, I did pursue this activity but mostly in the mind. OTOH, I did publish on what used to be arguably the most often quoted verse from Gitaa.
2.6. Review of last year’s NYR’s: Meditation, side-readings (on all topics—not just spirituality), etc.
I had said: “Satisfactory pace achieved already. No need to change it; certainly no need to make NYR’s about them as such.” Yes, my reading did continue, satisfactorily enough.
3. My new year’s resolutions for 2021:
3.1. Try to “un-become” “Bohmianism”:
Remember, I had resolved to be a Bohmian via my Diwali-time resolutions? See this 23 November 2020 post: “Now I am become Bohmianism” [^]
Now my NYR is to cancel it!
Reason: I won’t tell you right away. You should be able to figure it out anyway, esp. over the course of the new year!
Aside: However, you know, NYR’s are sooo sooo hard to keep. So, don’t be surprised if I end up saying something like “We the Bohmians…,” also in the new year!
3.2. My new approach to quantum mechanics:
Conduct some basic simulations and write *some* preliminary documentation on the spinless electrons (up to 3 electron systems) by Q1-end. If the RSI severity goes up, by Q2-end.
QM with spin:
Conduct some basic simulations and write *some* preliminary documentation on the spinning electrons (up to 3 electron systems) by Q2-end. If the RSI strikes, by Q3-end or year-end.
Depending on the progress, revise my 2019-times Outline document on my new approach. Note, revising that document is optional.
Scope of this NYR:
The above resolutions do not cover the more advanced topics like: photons, detailed studies on the times taken by QM processes, detailed studies of the multi-scaling issues, etc.
However, the above resolutions do cover predicting the bonding energies of electrons in the helium atom (the 1D case), and the preliminary three-particle simulations.
3.3. Data science:
Work a bit on some projects I have in mind—at least two or three of them.
Note: Nothing big here. Just some small little projects of personal interest. Details will become apparent over the course of the year.
3.4. Health etc.:
As noted above, the last year’s NYRs failed. So: Adjust the resolution further, for this year. Accordingly, the NYR for this year is:
Commit to performing one soorya namaskaar every day.
If even this routine fails for whatever reasons (say, a genuine reason like travel etc., or even plain out of irregularity, forgetfulness, boredom, etc.), then it still would be fine. But the next time the issue crosses the mind, just resume the “at least one” routine back again. (Yes, such resumptions are an integral part of this very resolution itself!)
Also included in this resolution is this point: Publish a summary of my actual performance at the end of the next year. (So, keep a record!)
No resolutions concerning food, drink, etc., this year.
Start learning Sanskrit in a more formal manner, doing an online course (or more of them).
That’s it! Nothing in the miscellaneous department, this year. The rest of the routines are doing fine (like, e.g., meditation, studies of other topics, and all). No need to change anything about them; no need to make any resolution either.
Apart from it all, take care of yourself, and have a Happy, Productive and Prosperous New Year!
I will only return after I have progressed to a definite extent in my QM-related work, which might be a couple of weeks from now on.
A song I like:
(Instrumental, “fusion”): “The River”
Composer: Ananda Shankar
[This was one of the instrumental pieces I would most often listen to, back in my IIT Madras days (1985–87).
… By now, I’ve forgotten whether I had heard this piece first in IIT Madras or in Pune. I do distinctly remember buying the original cassette for this album (“Ananda Shankar: A Life in Music”) in Pune, in particular, from a certain shop in the “Wonderland” shopping complex, on the Main Street in the Pune Camp area, and listening to it often in Pune. So, some chances are that I listened to it first in Pune. OTOH, my nearest retrievable memory also says two things: (1) I would listen to it on a
Sony National Panasonic “Walkman” (having a 3-band equalizer), and (2) I had bought such a Walkman in the Burma Bazaar area of Madras (now Chennai), though I no longer remember whether I bought it while being a student at IIT M or some time later. … So, all in all, I am not sure when or where I listened to it the first time.
… In any case, I am sure that the song became a routine transitioning music on TV and radio in India only some time later (may be after a few years or so). At least I hadn’t listened to it first on TV/radio. …
This song was one of my all time top-most favorites for a long time during my youth. Frankly though, listening to it once again only last year, after a gap of some two+ decades, it sounded a slight bit different. … But yes, it still remains one of my most favorites. (It’s surprising that I happened not to have run it here so far.)
This piece is short (just about 3 minutes long), but it is absolutely innovative, fresh, and creates a wonderfully unhurried, placid, but not lackadaisical mood. … Just as if you were sitting by the side of a river on an unhurried evening, while resplendent colors slowly unfolded in the sky and also on the placid waters, until it became fully dark, completely unknown to you…. Or, as a small, lone wasp of a cloud loitered around the sky, became thinner, and somehow, the next time you looked at it, it was almost gone… An outstanding piece this one is, IMO, even when compared to other pieces by Ananda Shankar himself… This piece carries that unmistakable Indian touch even as the composition unfolds with a Western-sounding orchestration/trappings. (And it was always a bad idea, IMO, to use this piece for transitioning in between TV/radio programs… But then, that’s an entirely different matter!)
So, anyway, give it a listen and see if you like it too. … Back then, it sounded very fresh and innovative. But with a lot of music of the similar kind, not to mention the easy access to the World music these days, if you are listening to it for the first time, this piece may not sound all that extraordinary. But back then, it was, for me at least. … A good quality audio can be found here [^].
Alright, bye for now and take care…
— 2020.12.31 19:19 IST: First published
— 2020.12.31 20:26 IST: Very minor corrections and additions, all in the songs section.