Oops about judge dear critically

The old saying that I say: “that which you hold very dear you judge very critically.”

But I just watched the first two volumes of Disney Family Sing-Along, and there’s not any Chinese artists here either… and Disney is certainly unarguably the definitive entity for story-telling, thought-inspiring, and in 4 years, multi-centennial, money-making media giga-conglomerate.

I cannot count the number of times I wrote about Star Trek not having any Chinese resembling Asian people in the future.

The one Chinese character they do have has misspelled name, her family name should be Hua not Fa. I wonder where Disney gets their history? From Hun-speak???!!

Star Trek still has another 50 years to exceed this, and I am confident it will!

The Hesitation

Suppose I want my chatbot to be conservative in its learning, what are the ways we control that? There is dropout, and weight decay, and normalization. One idea was to find a way for it to learn and then sleep for a while before learning again. If we look at the gradient of such a function, it would look like 1+ cos(x) :

1 + cos(x)

This gradient is motivated by the want of rest period, when the input passes a certain periodic magnitude the progress of gradient based optimizer is gradually slowed down(but still pointed towards the same direction.) After the sleep, the function wants to progress fast to make up for the time spent sleeping, so the steps are bigger. A bit of manipulation of that expression produces the function with that gradient function x+sin(x)

x + sin(x)

This addition of a hesitation can be used as a foreactivation (defined in previous FAM entries). Or in fact the hesitation layer can be placed any where a dropout is normally used.

A certain amount of experimentation is required to inject useful amount of randomization. This layer is particularly easy to instantiate after units that have known scaling, such as sigmoidal activations, softmax, batchnorm, and others. One formula in particular is amplitude. For \alpha \in (0, 1] The hesitation layer H_\alpha(t) = t + \alpha sin (2 \pi t)/{2\pi} makes it possible to adjust the flat part of the sleep cycle. Here is \alpha = 0.5 in blue next to the original:

H_{\alpha=0.5}(t) = t + \alpha sin (2 \pi t)/{2\pi}

More work is needed to establish the precise effect this layer has different types of optimizers. For optimizers like Adam, the added variability would normally increase the noise of the gradients passing through the layer thereby effecting a reduction the learning rate. For \alpha>1 the hesitation layer becomes non-monotonic, but since Adam accumulates gradients, the gradient in the “wrong” direction will slow the progress of optimization more significantly than smaller \alpha’s. It will not necessarily produce an unrecoverable valley of local minimum. For \alpha \in (0, 1) this layer will not introduce new univariate local minimums or saddles to the optimization it is being added to. With randomization, the units will sleep at different times giving other paths of gradient that are not sleeping a chance to explore their potential to improve.

Another direction to explore is learnable sleep cycles in the form of H_{\alpha\beta}=x + \alpha sin(2 \pi \beta x)/{2 \pi \beta} . Where \alpha \beta are either a single scalar or properly sized tensor for element-wise application. Generally the adjustment of \alpha \beta will be for quicker progress towards the direction of underlying gradients.


Ps one can work out the gradient H_{\alpha}(f(x))= f(x) + \alpha sin(f(x)) . Take the derivative wrt x. \partial H_{\alpha}(f(x)) / \partial x= (1+\alpha sin(f(x))\partial f(x)/ \partial x. So you see the gradient varies between [1-\alpha, 1+\alpha] times of f(x)’s gradients.

Transactional UBI

Maybe we should be paid not for “being nice citizens” but in quantifiable ways. For example, the government should pay us for the act of voting. This act divulges our preferences and intentions in a most intimate way, and it seems there is permanent record of our votes as well. would it not make sense to pay citizens to vote ?

This is similar to Andrew Yang’s idea of giving everyone money($100) to donate to political campaigns. The idea makes sense. The idea also enhances our Capitalism, it strengthens capitalism by making it even more integral part of our political system.

The idea can be extended further so that the entirety of the UBI is not a dividend for being a good citizen, it can be a payment for our efforts to perform political thoughts and actions: MATH, make American think about their politicians more and make them act, for money.

Of course, there could be problems. One may argue that if two candidates are running and they both suck! Everyone may just write their own children in for their choice. But of course, any good red blooded American will recognize that as a Socialist anti-free-market uncompetitive argument: if there are two sucky candidates, a third will rise and triumph over them in America: in a free world, you cannot be stuck with two really really bad candidates. (Can it happen? For several election cycles? Consecutively ?)

It comes down to a redefinition of our relationship with our government, what, exactly, did we agree to in our social contract ? what are the axioms of our constitution? Can we perform object-oriented analysis of the constitution? Can we restate it using domain driven design? Can we apply what we know about organizational behaviors, modern psychology and cognitive psychology, can we put our knowledge of these to an analysis to revamp our constitution?! Can we use our knowledge of information, and again here I advocate my idea of formally recognizing information property as a different type of property and use property law to regulate it. Can we put that in the constitution using our newly acquired knowledge about the science of information? Can we write our awareness of our environment into our constitution? Can the land of the free and home of the brave also by definition preserve the land and home?

And as all blog entries must discuss in 2020, what if an AI can write this formalized and rigorously designed constitution better than us? What then? Can we demand that a test for suitability of an AI for us is that our natural actions largely follow the new agreement?

So in summation, I feel that we should rewrite our constitution to inject new knowledge and remove old biases. We can implement UBI in a transactional sense, that citizens are paid for their political services such as: organizing or attending political rallies, public participation in political discourse, supporting political candidates formally in campaigns and in votes. For these service, the government can pay its citizens money.

P.s. and honestly, the governments do need to have a bit more of perspective. A huge fine on a tax payment 12 minutes past midnight because their stupid website didn’t load, this is just not right. How about we give the government a bit of their own medicine and fine it every time mail arrives late or roads have pot holes or police fails to protect citizen? Some times, people who work for the people in the government acts more like they work against the people. Case in point, when traffic patrol increases during periods when police departments have funding shortage, it is a great demonstration of ingenuity to keep something great going, but it also betrays the trust of citizens that the government serve the people. UBI is a way to ascertain that we establish a canon of reverse financial flow available to all equally.

I should also mention another idea promulgated here by the FAMx blog. We need to establish a fine on the government for intentional acts that violate our civil liberties. Suppose we eventually arrive at the agreement that privacy is a fundamental human right, and that the government chose to violate a certain person’s privacy for a purported higher purpose, and suppose that it is found the purpose was not legal or effective, that the violation of privacy right is automatically remedied by cash payments through the same channel that pays UBI. If FBI reads my private email for suspicion of some crime, and the FBI becomes aware that I didn’t commit it, then I should be paid money—automatically. This type of penalty on the government will ultimately encourage it to operate efficiently and accurately. Systematic payments, as opposed to incidental ones where some body finds the wherewithal to sue the government for violations and succeed in being awarded remedy, will help those who choose to be servants of the public to do it well. And that is fair. It is fine grained check and balance for our country, for our world, for our posterities, we must advance.

Permanent record

I would rate this FAMX.1, I’m sure I will achieve that in the limit…

Been blasting through Snowden’s book on audible. I have to say that my previous misgivings about lack of opportunity to achieve greatness like Snowden is completely wrong. The first two or three chapters reads like my childhood, minus the link to the Mayflower and familial connection to the military. I mean, down to the publications and games that I were exposed to growing up in Pennsylvania, just north of Maryland where he grew up. I can find my own stupid posts on the internet made from a dial-up BBS too! Many this book speaks to me in so many tones, and some of it sound like myself explaining things.

Anyways, one revelation is particularly interesting. It seems that almost all recruiting agencies are run by American spoooks! This explains so many well timed emails that can float above all the spam, that one came when your fat finger accidentally hit the wrong button and traded 10 TSLA options the wrong direction instead of 1(can a finger really be that fat?) twice, or that LinkedIn alert that pop just as you find the stupidest and most embarrassing bug in your own side hustle code(or main project at work), or that email offering of a good package when you have a particularly intolerable rash without insurance, that email that come when you experience domestic strife at the worst!

THIS is WHY! They, the recruiters, have special access to “the system” the likes of post-911 Snowden built! Of course they know exactly when you are vulnerable! That’s what the spy’s computer systems were made for, finding people’s weaknesses. The fact that these consultants take advantage of the system to target you when you are most vulnerable so they can make commission is a little bit annoying—mainly you kind of are now perpetually embarrassed at everything you do, because you know they are watching! Peed your pants? Hit your child? Hit another car and drove away? Omg, the things that can happen when no one is looking.

Of course, I shouldn’t complain, almost all my jobs in the last few decades were from recruiters. Kept me busy. But for people who knows my work history, the profitability of those jobs were not as much as the jobs that did not come from recruiter. But I don’t want to sound ungrateful. They do serve a purpose and especially in your time of need. One has to get over the way it all happens and who now can see everything you do.

So true

You know, for all the reasons Americans hate president Trump, he does give some of the best public deliveries. I saw this snippet from a Quora article about math education

Sooo true

And I saw Trump say “this is sooo true…” in that voice of his. “sooo true, so much play, what a great country!”

😂, well, and then 😭,and😭.

A generation of greats

Covid-19 is forcing almost all parents to spend all day with their kids… in all human knowledge, it seems parents presence and attention is one of the best ways to guarantee a good upbringing.

I guess this is going to be one really kick ass generation of kids! We may even cal it the C-gen for covid-19 generation… these kids who benefit from 24 hours of biparental rearing are going to come out great! I can feel it! Humanity is greater and it’s future is brighter because of covid-19!

And no, snarky commentators, I’ve not gone insane from spending a lot of time with my kids.

Chinese Funnies

Americans may not understand why so many people wants to come to America. For those of you who, who fee like India felt when it banned several dozen Chinese apps, that al things Made in China are just evil, just imagine how Chinese people in China feels! Everything they buy are made in China!

So, imagine that and this following conversation ensued with Chinese-American parents recently:

P: the garbage bags you bought have disintegrated into small pieces about 1” in diameter. Did you buy fake garbage bags to save money?

S: uhhh, no! Those are bio-decomposable bags.

P: so they are not fake?

S: nope, you Americans all think alike, just because I was born in China doesn’t mean my stuff are all fake! In fact I’m so genuine, even when the stuff I prepare for you rot, they are still the real thing!

P: ~ (晕了)

Anyways, hopefully all this Chinese fear-hate-doubt (or fud) will wash away with the tide of time, may there be a happier time for us Chinese-Americans in the future.

Material Foundation of Intelligent Identity

There is a lot written regarding soul. There are also a lot written about the uniqueness of Human intelligence. Certainly I will start into fiction as I often do from Star Trek stories. Picard S1 ends with his body dying but the “substrate” of his mind transferred to an Android body. This is a far out idea, especially to happen to the star of a show whose Humanity is what we tune in for. Boldly we have fictionalized! FAMX0.25

So… it seems that in the Trekverse, there remains some idea that there is a non-replicable component of our intelligence, or “our being” when the get sentimental This essence can be transported but it cannot be replicated to make two of you.

It is certainly within our imagination that human uniqueness can be interpreted as an Pseudo-random number generator. Each of us can have our own PRNG seed that were set at behest of god or the Big Bang. It kind of determines who we are. At any time our beings have a state, it results from all of our history and determines how we think in the future. One can also imagine that perhaps the limit of our lives is at the limit of the PRNG cycle, when it repeats your initial state, the life ends. Even if it repeats, the essence of the existence is same as the first time it ran, the life has ended. So suppose our intelligence, or our biological/genetic process is a program that invokes this PRNG. And the ideas of free will is encapsulated as the PRNG as well, the freedom from deterministic behavior and stimulus response is the PRNG, it is our freedom to ultimately decide no matter what and where the Universe puts us.

Along those lines, one can imagine a kind of state that cannot easily be observed or replicated. So when they make a copy of you, at some level, some material, some black box machine, must be transported to the new copy, and thereby your identity within this body is destroyed.

This kind of give rise to ideas of reincarnation… of a particular PRNG has a very long cycle that somehow outlasts corporeal longevity of the computer using it, it may have to be transported to another computer to continue running.

Also, PRNG may be a retronym. In retrospect or in a frame of thinking where time is static, for example when considering the world as a 4-D volume, the information of this soul we consider may most easily be described using an PRNG, either in analogy or to model it. Ex ante, the soul is the soul, how it will play out is difficult to think about in the 3D representation that is the body and mind. For all intents and purposes, it is a true RNG, to me.

And of course, that I think it now seems to imply that what I use to think is not quite 3-dimensional either, or that it is but is a result of a suitable RNG.

Woohoo USS Zheng He

Oh hoooohooo!! My complaints into the inter web have been heard, the federation, in the 24th century, celebrates 14th century Chinese Navigator Zheng He by naming their state of the art flagship after him. Wow! And she leads an armada of ships… as the original Zheng He did.

The willingness to use a Chinese Naval general is awesome! There is so little Chinese here that anything is a huge boon! I mean, also, talk about getting all the bang for the diversity buck in using just two words… Zheng He is was born to a Muslim family in 14th century China. He was enslaved and castrated by the Han-Chinese Ming dynasty at a young age,… but to make a long story short, he overcame that kind of adversity and rose to become a general/Admiral and lead a series most successful naval missions known in all of Chinese history.

It’s like Dunkirk… so much in so few words

The season of Star Trek Picard has ended… in great haste. That 10th episode was at least two episodes in Best of Both Worlds… I don’t understand why everything feels so rushed. It may be just that the story and acting is so good that one wishes to linger longer in the Universe of Star Trek. Especially, now that we have regained our corrected Starfleet. In my younger youth, I remember being able to hold that excitement and memories of emotions and tech and myths and plots all for longer time after an episode… I mean, for god sakes, we’re talking about an tempted galaxy-wide biocide! Like, it’s not homicide, it’s not genocide, it is BIOCIDE! My mind just cannot take that in within the span of one sentence! It needs time to experience this.

One cannot praise the acting enough though. Stewart, can blush like Briones and barf like Pill… heck, he can probably play a few half’s of soccer with Cabrera and… and when he smiles, you know it is right. You know the universe, with starfleet, is alright.

The Trek is a little bit less subtle than TNG, IMHO, when Picards explains to Rios the powerful tools of federation: Openness, Optimism, the spirit of curiosity, (all they have is fear, fear is the great destroyer, not…)… failing to use these powerful tools is a failure of imagination (creativity/intelligence/technology) These words while verbalize some ideals behind Star Trek, it is difficult to contemplate that they can be taught in words, until I saw Picard do it. You can actually say these things out loud in a human language. I do not know how I feel about this direct approach, but it hits the nail right on the head, directly.

I am also a little scared of season two, probably for lack of my own creativity. How do they follow a Picard-data story like this… and many bigwig producers are leaving the show.

But, I suppose I should take a cue, and boldly look forward to the unknown, with hope and optimism full blast!