Impeaching Fast or Slow

What a difference a couple of weeks make. It seems but yesterday that Donald Trump looked impervious to all the challenges that Democrats and Reality had thrown his way. Today, not so much. Poll after poll shows that a solid majority favors proceeding with the recently started impeachment inquiry. And several polls – including the latest Fox News poll – have begun to show majority support for actual removal. No wonder the President is losing his mind.

Until recently, Democrats and others who oppose Trump were extremely frustrated by the timidity of the Democrats in pursuing impeachment – especially after the Mueller report. I have written elsewhere about the possible strategy there. Now that the impeachment process is underway, another – more important and concrete – question has arisen: How fast or slow should impeachment proceed?

There are two schools of thought on this. One – espoused by those who can barely put up with Trump any more – is to move ASAP. After all, there is more than enough concrete evidence to frame one, two, perhaps even three irrefutable articles of impeachment. The argument is that the Democrats must strike while they still have the public’s attention, and that, with time and Christmas holidays, the air will begin to go out of the balloon. Those who hold this opinion are still frustrated by the pace at which the Democrats are moving, though not at previous levels. Among other things, this group implicitly concedes that there is no possibility of a Senate conviction, and impeachment per se is the most that can be done.

A second school of thought is that the Democrats should proceed methodically, laying bare as much more evidence as possible for the public to see, thus building such inexorable public pressure that Trump’s numbers collapse and even the Republicans in the Senate begin to turn on him. The logical conclusion of this scenario would be a Watergate-style Republican delegation convincing Trump to see the writing on the wall and quit. Events so far have provided more support for this approach in that: a) Every day since impeachment began has brought more Trumpland corruption to light; and b) Public support for impeachment and removal is growing at an unthinkably rapid rate. However, those who think that this process will lead to Trump’s resignation are almost certainly wrong. What is likelier is that Trump will frog march that hypothetical Republican delegation right back to Capitol Hill and refuse to leave the White House, thus precipitating a crisis that does not even have a name in the American political lexicon.

While none of the Democratic presidential candidates have said so publicly, it is possible that most of them also favor the first option so they can get back to the business of their campaigns without being drowned out in the media. But it is also instructive to look at things strategically from the viewpoint of the people who are now – finally – driving the process: Nancy Pelosi and the House Democrats. Consider the situation today. Every day that passes brings new scandals to light. Trump’s poll numbers look worse. He himself says even more unacceptable things that his cowardly sycophants in the Republican Party then have to go and defend in public – or tie themselves into knots trying to avoid doing so. Every day that the Republicans spend stuck in this dignity-losing quicksand is a day they sink deeper into its muck – and a day they are not trying to win the next election. And the longer the Democrats can keep the Republicans in this awful posture, the more public opprobrium they bring upon the Republican Party as a whole. This strategy of slowly sinking the entire Republican Party into oblivion by tying Trump’s misdeeds more and more tightly around their necks every day is a political winner for the Democrats. Their goal should not be to remove Trump or allow the Republican Party to abandon him and nominate a more electable candidate such as Nikki Haley or Mitt Romney for the 2020 ticket. Rather, it should be to stretch out the impeachment process in a way that allows Trump to remain extremely popular with the 30% Republican dead-enders while becoming extremely unpopular with the rest of the country. This will make sure that Republican candidates cannot turn on Trump but are forced to defend behavior that is becoming more and more unpopular with voters in general. Ideally, the process would end with a weakened but defiant Trump still on the Republican ticket in 2020 and going around the country shouting cringe-inducing profanities to anyone who shows up to his rallies. If that can come to pass, the Democrats would have a great chance of winning the presidency and both houses of Congress.

All this will require exquisite timing by the Democrats. If the impeachment inquiry ends too quickly and Trump is acquitted in short order, there may be time for the Republicans to recover. If the inquiry goes on too long, becomes too complex by picking up a lot of extraneous issues, gets swamped by the Holidays and the Superbowl, and bleeds into the primary season, that too may be a problem. It will keep the Democratic presidential candidates from talking about the bread-and-butter issues that win elections, and dissipate the attention of the non-activist voters that the Democrats desperately need in 2020. Somewhere between impeaching next week and waiting for months, there is a sweet spot for the Democrats, and the Pelosi-Schiff team needs to hit it.

And what of the Republicans? The key point is to realize that going into the 2020 election with Trump as their candidate is already a liability for the Republican Party, and will keep becoming more so as impeachment proceeds. The Republicans have much better options than Trump – especially given what they see as a weak Democratic field of candidates. Surely, Mitch McConnell is more aware of all this than anyone else. Some gears must be turning in that scheming mind of his.

One of the more interesting events in the young impeachment saga so far is a letter signed by sixteen prominent conservative lawyers – including George Conway, husband of presidential mouthpiece, Kellyanne Conway – recommending “expeditious” impeachment in the House. The letter is surely motivated in part by righteous indignation, but there may also be a more complex purpose at work. For those committed conservatives who would like to see the recent era of conservative dominance in the courts and legislatures continue, the best of all the bad outcomes would be a very quick end to the impeachment saga – preferably with Trump gone. That is the only rational course for Republicans, and the sooner it is deployed, the better it is for their prospects in 2020. If this strategy is in the works, look for McConnell and the Senate Republicans to gradually start pivoting to a “minus-one” strategy. Of course, if it comes to that, Trump is unlikely to go quietly. Apart from injury to his ego, he and his family members are looking at many potential prosecutions if he ceases to be President before the statute of limitations on their shenanigans has run out. The only way to convince him to leave would be under an immunity deal. If the Democrats take a hard line and refuse to agree to any deal, it would ensure almost certain catastrophe for the current version of the Republican Party. And perhaps, in putting himself over his party, Trump will finally have performed his sole act of patriotism for his country.

Mission Accomplished?

India's decision two days ago to revoke most of Article 370 of its constitution and annex the part of Jammu & Kashmir it holds has sent Subcontinental and transcontinental punditocracy into a frenzy of analysis, interpretation, speculation, and prediction. Several scenarios have risen to the surface.

The most interesting of these, generating a lot of chatter on the Internet - and elsewhere, no doubt - is a conspiracy theory that India's move is part of a brilliant coordinated strategy between India, Pakistan, and the US to eventually make the LoC an international border with minimal political cost to either government. There are many variations of this theory, but the basic idea is this. First, India moves into its part of Jammu & Kashmir and annexes it, allowing the BJP government to look heroic and turning the LoC into an international border, with a buffer territory - Pakistani Jammu & Kashmir - on the other side. Then, after a suitable interval of making noises and writing plaintive but futile missives to the UN, Pakistan declares that the situation is intolerable and annexes its part, thus making the LoC an actual international border. Uncle Sam rewards Pakistan for this daring act by allowing it to negotiate a favorable settlement in Afghanistan, thus fulfilling Pakistan's dream of "strategic depth". Some sort of free cross-border movement is negotiated for Kashmiris on either side of the border. China secures CPEC. Everyone is left happy and dreaming of visits to Oslo.

I think this scenario is extremely unlikely to be true - though it makes for a good movie plot. First, it assumes that the Muslim population of Kashmir will just roll over, which it absolutely won't. Second, the institutional commitment to "all-or-nothing" is too strong in both India and Pakistan to make this an easy process. In particular, Pakistan has a large number of uncontrollable militants who can create complete chaos in the country at the slightest suspicion that Kashmiris had been "sold out". And third, it will leave the Pakistan Army wondering where its next meal will come from. In other words, this fanciful solution is too cold-bloodedly rational to be realistic. And it requires more finesse than politicians in India, Pakistan, or the US can currently muster.

So what else can happen?

If India’s hopes turn out to be true, nothing much. Everyone will give up, recognize the new situation as fait accompli, and live happily ever after. That has about as much chance of coming to pass as Donald Trump converting to Islam. Some have suggested that the situation will escalate inevitably towards war – and then nuclear conflict – between India and Pakistan, which is a prospect too frightening to be contemplated. If that's where this is going, we need to stop worrying about climate change, and focus only on changing the climate – quickly. However, this scenario too seems to be unlikely in the short term. The Pakistani leadership has been caught by surprise – especially after Trump’s vague comments about American mediation during his meeting with the Pakistani Prime Minister, Imran Khan. For things to escalate quickly into war at this point would require either extreme stupidity from India or extreme recklessness from Pakistan. The likelier scenario – more insidiously cruel – is an endless war of attrition.

Since Partition, the fervor of the Muslim population in Indian J&K for self-determination has been tempered by the concessions of Article 370 and the presence of pro-India leadership from Shaikh Abdullah onwards. Now 370 is gone, and so is the credibility of the leaders who preferred staying with India. The Indian action will likely align all Kashmiri Muslim leadership towards the same purpose: Freedom from India. That will inevitably lead to a much more organized insurgency in the region than has been the case so far - and one with much stronger commitment from the local populace. Politically, Indian leadership will have no choice but to fight the insurgency, leading to a brutal guerilla conflict. The insurgency will have clear supply lines from Pakistan, which will see itself as supporting a just war of liberation with complete support from the Pakistani public. Also inevitably, the Pakistani part of J&K will become a staging ground for all this. There will be cross-border attacks, a refugee crisis, and other ramifications for Pakistan, but Indian forces will be bogged down in difficult terrain and among a hostile population for year after year after year. India will also have squandered much of its international legitimacy on the Kashmir issue, having unilaterally flouted UN resolutions and chosen a maximalist course. It will still get by due to its economic heft, but some of the shine will definitely be off. Clearly, the Indian government has decided that this is a price worth paying, but the judgment of history is yet to come.

There have been comparisons of India's annexation of J&K with the hypothetical case of Israel annexing the West Bank. A more apt and real comparison is with the US in Iraq or even the Soviets in Afghanistan: A powerful military fighting a deep-rooted insurgency in a difficult region that has a long border with a hostile power. This sort of thing never ends well, though if it comes to pass, those who suffer the most will be the people of Jammu & Kashmir - as was the case in both Iraq and Afghanistan. Another comparison that is perhaps too apt and sensitive to be contemplated is with the Pakistani army in East Pakistan in 1971. In that case, things escalated to a real war, and one side was decisively defeated. Many in Pakistan have, ever since, thirsted for revenge, but there seemed no prospect of an opportunity. And suddenly, here we are!

There is currently much rejoicing in large parts of the Indian Right about Prime Minister Modi’s masterstroke. The bereft looks of Pakistani leaders and the ineffectual protests of Indian liberals only reinforce this triumphalist euphoria. To a dispassionate observer, however, the current situation is more reminiscent of George W. Bush’s “Mission Accomplished” moment. The war in Iraq had just begun then.

One thing, though, is clear. South Asia is changed forever by this hinge moment in its history. Many things that were not possible are suddenly possible, including some very, very bad ones – and perhaps a few less bad ones. It is a time when visionary leaders could remake the entire future of the Subcontinent, but the path to that future is perilous, and vision is too rare in Kaliyuga.

What AI Fails to Understand - For Now

What AI Fails to Understand - For Now

My latest piece on AI for 3 Quarks Daily.

Pearl on AI

I read this interview of Judea Pearl on AIwhen it first came out a year or so ago. Lots of important points in there. He's absolutely right about the rut AI is stuck in, but I think he is partially wrong about the way out, which he thinks will involve engineers building models of reality inside robots:

"We have to equip machines with a model of the environment. If a machine does not have a model of reality, you cannot expect the machine to behave intelligently in that reality. The first step, one that will take place in maybe 10 years, is that conceptual models of reality will be programmed by humans"

That representation-based, "information processing" view of intelligence is the problem, not the solution. The models real intelligence uses are, for the most part, implicit in the system, not built explicitly as models.They emerge naturally from the interaction of the adaptive organism with its environment, and become embedded in the physics of the system (its tissue, its joints, its neural networks, etc.) The capacity to build explicit models appears very late in evolution, and even then, it is more the capacity to "feel as though" there is a model rather than there actually being a model that is used in decision-making. Robots will have the same capacity when they become able to make sufficiently complex decisions. At that point, they too will have theories and hypotheses about the world, i.e., models of the kind Pearl talks about. Free will, consciousness, and other such fictional things will also emerge then, as Pearl says too. I don't think we should worry about implementing these things. I am also very skeptical about correct causality as the basis of intelligence. The bee does not "know" the cause of anything but does very intelligent things. The estimation of "true" causes in complex systems is mostly futile; what we care about are relationships, and since some of them are temporally ordered, they can be seen as cause and effect, but only in a post facto descriptive framework such as language. And yes, statistical learning alone is likely not sufficient to discover relationships, as too many machine learning people seem to think today, but that is just an issue of levels, Ultimately, all our knowledge about the world is statistical, except that a lot of it is acquired at evolutionary scales and is encoded in genes that generate specific bodies and brains, and in the developmental process. Learning comes in late to build on this scaffolding of constraints, instincts, and intuitions.

A robotics/AI colleague and I had an interesting discussion yesterday, and agreed that, rather than projects like the Human Brain Project, AI should have a Real Insect Project - building an insect that can live independently in the real world, survive, find food, find shelter, etc., completely autonomously. Once that is possible, it's just a question of scaling up to get human intelligence :-). We can call it Project Kafka! I once said something like this at an AI conference. People were not pleased....

(Machine) Learning Biases

 In a recent tweet, Congresswoman Alexandria Ocasio-Cortez - widely known as AOC - responded to a report from Amazon that facial recognition technology sometimes identified women as men when they have darker skin. She said:

"When you don’t address human bias, that bias gets automated. Machines are reflections of their creators, which means they are flawed, & we should be mindful of that. It’s one good reason why diversity isn’t just “nice,” it’s a safeguard against trends like this"

While I agree with the sentiment underlying her tweet, she is profoundly wrong about what is at play here, which can happen when you apply your worldview (i.e. biases) to things you're not really familiar with. To be fair, we all do it, but here it is AOC, who is an opinion-maker and should be more careful. The error she makes here, though, is an interesting one, and get to some deep issues in AI.

The fact that machine learning algorithms misclassify people with respect to gender, or even confuse them with animals, is not because they are picking up human biases as AOC claims here. In fact, it because they are not picking up human biases - those pesky intuitions gained from instinct and experience that allow us to perceive subtle cues and make correct decisions. The machine, lacking both instinct and experience, focuses only on visual correlations in the data used to train it, making stupid errors such as relating darker skin with male gender. This is also why machine learning algorithms end up identifying humans as apes, dogs, or pigs - with all of whom humans do share many visual similarities. As humans, we have a bias to look past those superficial similarities in deciding whether someone is a human. Indeed, it is when we decide to override our natural biases and sink (deliberately) to the same superficial level as the machine that we start calling people apes and pigs. The errors being made by machines do not reflect human biases; they expose the superficial and flimsy nature of human bigotry.

There is also a deeper lesson in this for humans as well. Our “good” biases are not all just coded in our genes. They are mostly picked up through experience. When human experience becoming limited, we can end up having the same problem as the machine. If a human has never seen a person of a race other than their own, it is completely natural for them to initially identify such a person as radically different or even non-human. That is the result of a bias in the data (experience, in this case), not a fundamental bias in the mind. This is why travelers in ancient times brought back stories of alien beings in distant lands, which were then exaggerated into monstrous figures on maps etc. This situation no longer exists in the modern world, except when humans try to create it artificially through racist policies.

The machine too is at the mercy of data bias, but its situation is far worse than that of a human. Even if it is given an "unbiased" data set that includes faces of all races, genders, etc., fairly, it is being asked to learn to recognize gender (in this instance) purely from pictures. We recognize gender not only from a person’s looks, but also from how they sound, how they behave, what they say, their name, their expressions, and a thousand other things. We deprive the machine of all this information and then ask it to make the right choice. That is a huge data bias, comparable to learning about the humanity of people from distant lands through travelers’ tales. On top of that, the machine also has much simpler learning mechanisms. It is simply trying to minimize its error based on the data it was given. Human learning involves much more complicated things that we cannot even fully describe yet except in the most simplistic or metaphorical terms.

The immediate danger in handing over important decision-making to intelligent machines is not so much that they will replicate human bigotries, but that, with their limited capacities and limited data, they will fail to replicate the biases that make us fair, considerate, compassionate, and, well, human.

The Perils of Unmotivated Thinking

This article just gets lost in anecdote, but Gary Smith is making an important point here: Good priors matter. When we look at lots of data and try to make sense of it, the expectations we start with make a crucial difference. If those expectations are productive and realistic, data will tell us important things. If the expectations are wrong or poor, we'll just find lots of garbage and think it is gold. Unfortunately, when we're faced with data, we often don't know what expectations to go in with, and the reasonable option seems to be to assume nothing, i.e., to go in assuming that everything is equally possible - a uniform prior. After all, what could be better than looking at data with an open mind? It turns out that this is usually a bad choice. Without the inherent discrimination provided by prior expectations, any large dataset can show all sorts of "patterns" leading to false conclusions.

There has recently been justifiable criticism of the use of motivated thinking, confirmation biases, and such in scientific investigations. What is often not discussed is that unmotivated thinking and unbiased analysis is often far more dangerous. The inevitable lesson is that, rather than rejecting all motivation and bias, we need to identify good motivations and appropriate biases. Unfortunately, there is no good way to do this in a purely mathematical or computational way. In animals (and, in a broad sense, in all living organisms), evolution has successfully configured useful biases. In fact, that can be seen as its most amazing accomplishment. We give these prior biases many names: Instinct, intuition, heuristics. But ultimately, they shape expectations based not only on the animal's own experience, but the experience of all its ancestors going back to the origin of life. As animals, we sense everything in the context of our biases, and make instinctive sense of it. Our physical body is an instrument sculpted by evolution to accomplish this task every instant of our lives. This is the essence of cognition, consciousness, and intelligence. And this is what even our best AI systems lack. They do have biases of course - every computer program, every circuit, every robot is biased by its architecture - but these biases are not the result of an adaptive process such as evolution. Rather, they reflect mathematical convenience, engineering constraints, and sometimes just plain ignorance or laziness. Not surprisingly, then, such systems have a hard time learning the right thing.

This also leads into another subtle point. Our most successful AI systems are those that use supervised learning, i.e., where some type of "ground truth" is used to correct the behavior of the system during learning. But that is just an implicit and very strong way to bring in prior biases based on reality. Where we have the most difficulty is in unsupervised learning, where the AI system goes looking for patterns in data without much prior bias, e.g., finding correlations or clusters. Unfortunately, the use of supervised learning is limited by the fact that, most of the time, the ground truth just isn't available. Real animals do almost all of their learning unsupervised, and mostly succeed because their instinct substitutes for the absence of the ground truth. That is what we will need in any real AI systems, and that is where the AI project should concentrate its greatest effort.

There is one type of AI that does try to approximate this: Reinforcement learning, where a system learns to critique its own options, and ultimately to make better decisions. The spectacular success of AI programs like AlphaGo and AlphaZero is based on a good marriage between the algorithms of supervised learning and the principles of reinforcement learning. However, there is still one big difference. In (most) reinforcement learning, the internal critic itself learns by pattern recognition. It builds instinct from the ground up based on data, albeit in collaboration with the decision-making and feedback from the environment. This is why reinforcement learning works best when the system is operating in the real world with real feedback, and why it is so slow. It's trying to build a new mind every time it is applied! In an animal, evolution has already configured a mind in the physical structure of the body (including the brain). That mind already has instincts, and needs very little experience to learn ti be (mostly) right enough.

There's much more to say about this, but I'm going to start by ordering and reading Gary Smith's book.....