More Sycophancy in the Congress

This is of course a pet peeve of mine, but it’s a peeve that is reinforced and confirmed with frightening regularity while the Indian press plays ostrich (or does it ignore the elephant in the room — pick your favourite metaphor).

Under fire from Baba Ramdev’s supporters, and questioned by his own party, Congress MP from Arunachal Pradesh Ninong Ering began blabbering about his loyalty to the Family.  After getting into trouble for making apparently anti-Indian comments, Ering began talking about how it was all because of his love for Rajiv Gandhi.  See here. That is the best hope for a Congressman nowadays: when in trouble, express love and devotion for the Family.  It might save you.

Watson and the Singularity – IV

On the last day, Watson was not nearly as dominant as on the second day, but won a big packet on the Final Jeopardy question to take its total winnings to much higher levels than either of its contestants.  For a large part of the third day, Ken Jennings was in the lead.

Watson uses the correctness of previous answers to try to understand what a category means.  For example, in a “Name the Decade” category, Watson wasn’t sure what the category name meant — so it didn’t know that it could restrict its answers to the set {1900, 1910, 1920, …, 1990, 2000}.  When it is unsure in such a manner, Watson uses previous correct answers in that category — possibly by its competitors — to narrow down what the category means.  It was clear on the third day that Watson didn’t understand what certain categories meant, even after observing opponents’ correct responses.  This meant it did poorly on those categories as a whole.

Watson’s natural language processing, I think, is tailored to the task at hand — winning Jeopardy.  Like I said before, it doesn’t understand sentences the same way a human would.  While watching the show, I found that many of the answers were found in the intersections of two or more sets, but Watson didn’t identify all of the sets.

I would classify Watson as a mild sub-Singularity event at this point.  If indeed programs like Watson proliferate the way chess programs have — if Watson clones become much more powerful and lightweight enough to run on the computer as personal assistants, perhaps with some help from the cloud — we will be on our way to real artificial intelligence.  Sequential improvements in such programs will eventually lead to super-human intelligence, much as the Singularity gurus predicted.  Eventually, APIs for NLP and this type of reasoning might become commoditized — unless companies like Google prefer to provide access APIs only, and keep all the computation hidden on their servers.

The question that immediately popped into my head when I heard about this Jeopardy challenge for the first time was “why not Google?”.  IBM has a fantastic record of innovation, of course, but the things Watson does are right up Google’s alley.  Search would be greatly improved if you could ask a question and have it answered in addition to being served a bunch of related pages.  Personally I believe Google has already developed a system like Watson.  So why don’t we know about it?  Two possibilities.

First, Watson needed 2800 processors to answer questions one at a time.  The technology that Google has may or may not be equally advanced, but perhaps doesn’t scale up to allow answers for millions of questions yet.

Second, this is a card in Google’s hand that it doesn’t want to show unless necessary.  If a competitor (mostly Bing at this stage) appears to be making significant inroads into its search space, it can add this feature to jump ahead, so it’s insurance.  Revealing everything would just provide Microsoft with a “copy this!” blueprint.

Watson and the Singularity – III

On day 2, Watson comprehensively outscored his human opponents.

To some extent, it seems Watson is at an advantage because his pneumatic button-pressing system can react faster than any human possibly could.  This severely affected Ken Jennings, who obviously had some of the answers and showed frustration at never being able to get to the buzzer first, shaking his head on occasion.

Perhaps a more fair way to assess Watson’s intelligence (as opposed to his button pushing prowess) is to adjust Watson’s button presser to be more commensurate with the pressing rate of human nerve systems.

Although Watson is doing great, it is becoming more apparent that Watson doesn’t understand the nuances of language in the clues as well as a human could.  There’s a document here (PDF) detailing some of Watson’s programming.

Watson and the Singularity – II

So, I watched the first part (of three) of the IBM Watson Jeopardy challenge.  So far, Brad Rutter and Watson are tied at first place, with Ken Jennings somewhat behind.  I was interested in getting a sense for how Watson “thinks”.  One of the things I tried to do was get a sense of the extent to which Watson is “understanding” human language.  At this point the answer seems to be “not very well”.

Of course, it’s hard to glean much just watching a TV show, but it seems as if Watson isn’t quite understanding language the way we do.  If a clue indicates the answer is a member of two sets, for example, Watson sometimes seems to ignore the second set.  An example (paraphrased): This word can mean the bend in the elbow and also a thief.  Watson’s best guess was “knee”, which has nothing to do with the second set (words that can mean “thief”) though it does have something to do with the first set (words that can mean the bend in the elbow).  The right answer was “crook”.

Watson seems to do superlatively well when there are unique phrases to be matched, i.e. when the clue contains phrases that are pertinent only to the answer and wouldn’t occur anywhere else.  Perhaps this is not surprising at all.

It’s possible Watson’s thought processes are a bunch of shortcuts completely unlike ours.  It may for example simply be finding a bunch of words and phrases based on associations with keyphrases in the clue and then ranking them.  Rather than searching for words/phrases in the sets that the clue is asking for.

Perhaps the right test is this: is it easy to add a subroutine to Watson that would allow it to rephrase the clue in several simpler English sentences?  I don’t know.  So I’m still unsure whether to call the creation of Watson a Singularity defining moment.

Egypt’s Revolution and India

I saw two blog articles today with ideas that surprised me.

The first is pointed out here:

Pentagram’s Vishal Dadlani exultantly tweeted that the band had a rocking concert in Guwahati and Egypt became “free” on the same night! Then he wrote, “All it took was 18 days, and that the Egyptian people stood up for their rights. Come on India, you can do it!!”

The second one is here:

Even as Indian observers debate the question of why and how Egyptian revolution cannot be replicated in India, the unique characteristics of Egyptian revolution will be of interest to the readers here.

Of course we Indians are interested in what happens in Egypt, some in a general world-news way, some more intellectually.  But these two people (i.e. Vishal Dadlani and V. Venkatesan) are talking about the feasibility of replicating Egypt’s revolution in India.

Wait, what??

How did that idea even enter the mindspace of the Indian public?

Presumably because India has a lot of problems.  A revolution might be an improvement in a country (like Egypt) that has been dictator-controlled for 30 years.  But in a naturally feudal country with an active democracy (like India) that is seeing steady improvements, a revolution is like a roll of a pair of dice — with things getting better if both dice show 6.  Most likely an Indian revolution will lead to a messy neo-feudal society with various warring factions tearing the country apart.

Are these commentators insane?  They have a personal desire for a little excitement, and their way of getting their fix is to encourage revolutions in democracies.

Nokia and Microsoft?!

Nokia, one of Open Source’s biggest advocates and sources of strength, has practically merged with Microsoft, Open Source’s biggest enemy and saboteur.  The agreement goes beyond simple cooperation.  Nokia is killing Meego, an important open source initiative.

This has been on the horizon for quite a while, ever since Nokia hired long-time Microsoft insider Elop as its CEO, and intensifying with a leaked internal memo Elop supposedly sent to Nokia employees.

This probably will help Nokia in the long run, but it fundamentally changes the company’s character.  This is a sellout by a biased CEO.  I was with Nokia so far, but I’m switching to Android as soon as I can.

EDIT:

I had a conversation yesterday that got me thinking, and the more I think about it, the more it seems to me that Elop was a Microsoft implant into Nokia, sent in for this very purpose.  Elop apparently made some noises about why he picked Windows over Android, but I doubt he ever considered Android.  The deal is hugely helpful for Microsoft, which was struggling to get anyone to show interest in Windows. It also seems vaguely helpful for Nokia, but it’s simply not commensurate.  Microsoft gets Navtec maps, various services, a ready-made phone distribution net, and a huge share of phone profits from Nokia at absolutely no cost to itself!  Nokia gives all this potentially revenue-generating stuff to Microsoft for free!  All Microsoft is doing is providing the already-developed Windows OS — and even that for a share of the profits!

Watson and the Singularity

The Singularity is a recurring theme in artificial intelligence and in science fiction.

It refers to an event where a computer achieves certain significant feats of intelligence.  Different authors use it slightly differently, or use different terms.  Sometimes the word means achieving “self-awareness” (whatever that is).  Sometimes it means the creation of a computer that is as smart as a human.  Sometimes it means the addition of a technology that results in a massive increase in ability (the new state is usually one of higher-than-human intelligence).  Some authors speak of multiple singularities — computers caught in an ever-rising spiral of super-intelligence.

The AI promises of the 1980s turned out to be too grand and ill-founded to be realistic.  People thought then that they could program intelligence by programming the minutiae of thought.  This turned out to be a vastly bigger task than anticipated.

For a while, it seemed there were things humans could do that computers would never be able to.  One of the biggest, and most visible, blows was Deep Blue’s defeat of Garry Kasparov.  Today, software (Rybka, Glaurung, Stockfish) running on the ordinary desktop computer will easily defeat the best human chess players.  But Deep Blue and its younger cousins don’t really have intelligence, at least not what we usually mean by it.  They’re “on-rails”, and can do very restricted things on very restricted input sets.

But all this doesn’t mean man-made intelligence is impossible.  Instead of programming intelligence, we can perhaps include techniques like evolving it or learning it.  Is that what Watson has done?

Watson is a massively parallel supercomputer, developed by IBM, that will participate on Jeopardy against Ken Jennings and Brad Rutter.  Watson can understand a variety of language nuances and sift through its massive database to attempt to find answers.  Given the diversity of subject matter as well as question phrasing on Jeopardy, this is quite a feat.  Does this qualify as intelligence?

It’s not clear to me how Watson works, but the bits I’ve gleaned indicate that it is a collection of a number of hand-written subroutines that interact in carefully human-tuned ways.  There’s no automated evolution or search of algorithms to try to make it better.  In that sense, it is still algorithmic, much like Deep Blue.  But Watson’s algorithm is much more complicated and chaotic than Deep Blue’s.  It sounds complex enough that I view it as a limited form of intelligence.

Perhaps we are hitting the first technological singularity, although it’s not the single explosive moment some have imagined.

Watson might have been a good learning experience — the engineers at IBM must have figured out a lot about how to make computers think.  But it still lacks the essential ingredient that sci-fi authors fantasize about.  We still don’t have automated techniques to take a given computer and make it better.  That is, we don’t know how to make computers improve other computers.  That would be the a real Singularity.

Manmohan Singh and “Political Mileage”

The position of the Prime Minister of India is practically a sinecure nowadays.  The PM doesn’t make any decisions, formulate policy, know what his own cabinet is up to, or do anything worthwhile at all.  Instead, like the President, the PM’s job nowadays is to sign paperwork and meet and greet various foreign dignitaries.  All the real decisions are made by Sonia Gandhi, of course.

I forgot to mention one thing the PM does have to do.  Whenever the opposition brings up an instance of corruption or ineptitude on the part of the government, Dr. Singh makes a public comment.  He says, “The opposition is playing politics with this issue.”  Corruption at the Commonwealth Games?  That’s just the Opposition Playing Politics (OPP).  The 3G spectrum scam?  OPP again.  Black money stashed in Switzerland?  OPP.  That’s his explanation for everything.

(Like Paris Hilton tried to trademark “That’s Hot”, and Trump tried to trademark “You’re Fired!” (did he trademark it?), Manmohan should try to patent “OPP”!  He’d earn royalties from politicians in democracies everywhere.)

So I wonder what his reaction would be to this bit of ridiculous political grandstanding by Rahul Gandhi.  To summarize, Rahul Gandhi used a rape victim to make a surreal attempt at scoring a political point.  He dared Mayawati to visit the victim.  Presumably he believes Mayawati is personally responsible because the crime occurred in the state Mayawati is Chief Minister of.  Perhaps he also thought Mayawati was responsible because the victim was a Dalit, as is Mayawati.  (It doesn’t make any sense to me, but then this is Rahul Gandhi.)

Now where’s Manmohan with his comments about political mileage?

Hayek, Predictability, Justice and Affirmative Action

I had a rather exhausting discussion with Sanjeev Sabhlok in the comments of this post on whether reservations (or more generally affirmative action) contradicts basic principles of justice.

Hayek

The only good thing that came out of that discussion (from my viewpoint) is it prompted me to glance through some pages of Friedrich Hayek’s Law, Legislation and Liberty.  Specifically, I looked at his chapter on Social Justice.  The entirety of my exposure to Hayek’s ideas is that one chapter, so I’m quite happy to be corrected.

It’s not hard to see why Hayek is held in high esteem.  This small chapter covers a wide variety of issues, many of which I think are relevant to the issue of reservations and affirmative action.  I find Hayek’s writing style hard to read, since his passages sound ambiguous to me.  It’s probably true that he had a precise position on the issues, but his book isn’t written that way and feels somewhat open to interpretation.

I don’t claim to understand Hayek’s theories very well, but some things jumped out at me.  Hayek says that we have the right to set rules, but to expect that once the rules are set and the market is set in motion, it is pointless to speak of the justice of the outcomes of the market (as long as everyone follows the rules).

This is an interesting idea, and curiously it parallels Krishna’s “Karmanyeva adhikaraste…“!  We have the right to decide the rules of action, but not to decide the outcomes, which are subject to much randomness!  It is widely accepted that attempts to make sure everyone has what they need is socialism; indeed, that is often treated as the definition of socialism.  Hayek says that attempts to make sure everyone gets exactly what they deserve is still socialist.  He calls this social justice (or rather, says this is what others mean by social justice).  But Hayek goes a step further in thinking about this, and makes two apparently contradictory statements.

Predictability

First, Hayek seems to agree that laws should not be predictably biased towards or against a segment of the society.  This I think is fascinating and in fact a crucial consideration while framing laws.  At the time we frame a law, it should not be predictably unjust.

Elucidating what “predictable” means here is an interesting exercise in itself.  My interpretation is the following.  A predictable set of people is at time t is a set that is determined by events occurring up to time t, and not after.  Thus “Dalits in 2010” is a predictable set in 2010, but “millionaires in 2020” is not very predictable in 2010.  However, “millionaires in 2010” is of course predictable in 2010. As with all social things, a certain level of fuzziness in defining sets is probably convenient.  The set “billionaires in 2015” is probably predictable with 99% accuracy in 2010, although the set “millionaires in 2015” is much less predictable.  Allowing for this slight fuzziness, “Dalits in 2050” is a predictable set in 2010. (People can and do “change their caste”, often through birth certificate fraud, but so few do it that the set is almost determined in 2010.)  Have I defined predictability precisely?  Not mathematically.  But it seems precise enough for society, law and justice.

When we pass a law that increases the relative advantage or disadvantage of a predictable set of people to its complement, we are doing something wrong.  Thus, if we pass a law in 2010 that will widen the advantage gap between, say, the 40-th and 60-th wealth percentile of the population in 2020, the principle of predictability does not prohibit this (since these percentiles are not very predictable sets).  On the other hand, if we pass a law that will widen the advantage gap between the blind and the not-blind in 2015 (a moderately predictable set in 2010), there is something wrong with that law.  Similarly, if we pass a law in 2010 that increases the advantage gap between Dalits and non-Dalits in 2050 (a highly predictable set), there’s something wrong with that law.  This is the gist of my application of Hayek’s predictability criterion to the affirmative action case.

Of course, it’s not quite as simple as that.  It’s not enough to say widening the gap is bad and narrowing it is good — we should also worry about whether things are getting better for everybody.  If we pass a law that predictably reduces everyone to abject poverty, this might reduce the gap — but it’s not what we want.  On the other hand, passing a law that predictably makes Dalits remain poor while increasing most non-Dalits’ wealth is also obviously wrong — even though it is true that some people have gained, and no one has been harmed (relative to where they started off).  Also, most outcomes will invariably be biased if a short enough time-frame is chosen.  For example, it’s certainly true that, no matter what anyone does, those who are poor on Jan 1, 2010 will overwhelmingly remain poor on Jan 2, 2010 — or for that matter on Jan 1, 2011.  Since ALL laws are biased, should we refrain from passing any laws?  A reasonable time frame has to be attached to the term “bias”.

Thus, this is a very loose principle — laws will need to balance fairness, considerations of where people start off, practicality, enforceability, acceptability in society, timeframe and a number of other factors.  Indeed, I think it’s not possible to state a succinct, simple principle that can be the sole guiding principle behind all laws, or even identify all the factors that need to be considered.

Equality of Opportunity

Now, it seems as if equality of opportunity is a natural consequence of this concept of no predictable bias.  After all, if there’s no predictable set of people that is better off than another, isn’t this the same as saying that the law is equally unbiased towards everybody?  It seems any principle for framing laws should lead to laws that give everyone the same opportunities, even though various random events would likely lead to differences in final outcomes.

This is where Hayek seems to make a contradictory statement.  Hayek says that equality of opportunity is also a socialist ideal — not a-priori, but in its implications:

To achieve this government would have to control the whole physical and human environment of all persons, and have to endeavour to provide at least equivalent chances for each; and the more government succeeded in these endeavours, the stronger would become the legitimate demand that, on the same principle, any still remaining handicaps must be removed-or compensated for by putting extra burden on the still relatively favoured. This would have to go on until government literally controlled every circumstance which could affect any person’s well-being.

This sounds correct — it is obviously impractical to demand that the government provide perfect equality of opportunity to every single individual.  But Hayek himself says that

So far as [equality of opportunity] refers to such facilities and opportunities as are of necessity affected by governmental decisions (such as appointments to public office and the like), the demand was indeed one of the central points of classical liberalism, usually expressed by the French phrase ‘la carriere ouverte aux talents’. There is also much to be said in favour of the government providing on an equal basis the means for the schooling of minors who are not yet fully responsible citizens…

It seems that equality of opportunity is perhaps not inevitably socialist or classical liberal, but rather a mixture of the two tempered by the extent to which it is practical.  That is, government should endeavour to provide equality of opportunity up to the point where it has to start taking socialist actions like controlling people’s lives.  The line between providing equality of opportunity and socialism is blurred — so blurred that it’s silly to pretend there’s a line (my thoughts, not Hayek’s).

Affirmative Action

How does all this tie in with affirmative action?  The principle that laws should not be predictably biased would seem to indicate that affirmative action is necessary.  The current system is extremely harmful for Dalits and certain other backward classes.  Indeed, the state completely failed them for several decades, a situation that is only now starting to be rectified. Under current laws, and under any law that completely denies all forms of affirmative action, Dalits will predictably be disadvantaged and continue to be punished by the system for several decades.

It is important to note that this reasoning does not apply to every group that is disadvantaged.  If a Muslim and a Brahmin are equally smart and study in the same class in the same school (I’m establishing ceteris paribus here), I think the Brahmin has no advantage compared to the Muslim.  They are equally likely, or almost equally likely, to find good jobs.  In addition, opportunities available to Muslim and Brahmin kids are the same modulo their own beliefs.  That is, if a community of Muslims chose to reach out and accept the available opportunities, they would be no worse off than a community of Brahmins.  The same is not true for Dalits.  There are active as well as passive forces arrayed against the Dalits.

Thus, Hayek’s own notion of not predictably harming someone via legislation seems to support the idea of affirmative action for Dalits.

The important question whether this can be classified as actively harming non-Dalits.  I don’t believe so.  Increasing opportunity for Dalits in this way certainly decreases opportunity for non-Dalits, but opportunity was lop-sided to begin with, and the lop-sidedness continues to be maintained using marginally legal methods.  With affirmative action, entrance into various lucrative positions becomes tougher for non-Dalits, but still not as tough as it is for Dalits.

My Position on Affirmative Action

For the record, my own position is a guarded support for certain forms of affirmative action in the short term.

I think it’s important to base affirmative action not only on caste, but on as many major sources of predictable variability as practical.  This is the topic of the MIRAA score discussed in my other post.

I also believe affirmative action is nothing but a temporary pressure valve measure to quickly correct certain imbalances.  It is no substitute for free, high quality universal education.  Education, not affirmative action, should be the method of choice for ensuring equality of opportunity.  Education is the only useful very-long-term sustainable means for equality of opportunity.  The only reason for affirmative action is that it seems impossible to equalize “predictable opportunity” using education alone in the next 30 years.

Sonia Gandhi’s Feet

This makes for very interesting reading: http://www.hindustantimes.com/Angry-Congressman-sings-Telangana-tune/Article1-657742.aspx.

There’s nothing interesting about the politician in question or the backtracking he’s doing; what’s interesting is just how scared this 82-year old veteran is.  Scared of repercussions after he criticized Sonia, he said he’d touch Sonia’s feet if she agreed to splitting Telengana.  The abjectness of his situation, and the culture within the Congress, come through pretty clearly.  Numerous Congress leaders, including his own son, jumped to denounce his statements.