Fascinating interview on Edge.org with Tom Griffiths of Berkeley. For me, the most interesting passage is this:
One of the mysteries of human intelligence is that we’re able to do so much with so little. We’re able to act in ways that are so intelligent despite the fact that we have limited computational resources—basically just the stuff that we can carry around inside our heads. But we’re good at coming up with strategies for solving problems that make the best use of those limited computational resources. You can formulate that as another kind of computational problem in itself.
If you have certain computational resources and certain costs for using them, can you come up with the best algorithm for solving a problem, using those computational resources, trading off the errors you might make and solving the problem with the cost of using the resources you have or the limitations that are imposed upon those resources? That approach gives us a different way of thinking about what constitutes rational behavior.
The classic standard of rational behavior, which is used in economics and which motivated a lot of the human decision-making literature, focused on the idea of rationality in terms of finding the right answer without any thought as to the computational costs that might be involved.
This gives us a more nuanced and more realistic notion of rationality, a notion that is relevant to any organism or machine that faces physical constraints on the resources that are available to it. It says that you are being rational when you’re using the best algorithm to solve the problem, taking into account both your computational limitations and the kinds of errors that you might end up making.
This approach, which my colleague Stuart Russell calls “bounded optimality,” gives us a new way of understanding human cognition. We take examples of things that have been held up as evidence of irrationality, examples of things where people are solving a problem but not doing it in the best way, and we can try and make sense of those. More importantly, it sets up a way of asking questions about how people get to be so smart. How is it that we find those effective strategies? That’s a problem that we call “rational metareasoning.” How should a rational agent who has limitations on their computational resources find the best strategies for using those resources?
Worth reading (or watching or listening to) in full.
I can see the point of trying to understand why humans are so good at some things. The capacity to make rapid causal inferences was probably hardwired into our DNA by evolution — it’s ‘System 1’ in the categorisation proposed in Daniel Kahneman’s book, Thinking Fast and Slow, i.e. a capacity for fast, instinctive and emotional thinking — the kind of thinking that was crucial for survival in primeval times. But the other — equally important — question is why humans seem to be so bad at Kahneman’s ‘System 2’ thinking — i.e. slower, more deliberative and more logical reasoning. Maybe it’s because our evolutionary inheritance was laid down in a simpler era, and we’re just not adapted to handle the complexity with which (as a result of our technological ingenuity) we are now confronted?
This has interesting contemporary resonances: climate change denial, for example; fake news; populism; and the tensions between populism and technocracy.
Interesting essay by Gary Marcus. I particularly like this bit:
Although the field of A.I. is exploding with microdiscoveries, progress toward the robustness and flexibility of human cognition remains elusive. Not long ago, for example, while sitting with me in a cafe, my 3-year-old daughter spontaneously realized that she could climb out of her chair in a new way: backward, by sliding through the gap between the back and the seat of the chair. My daughter had never seen anyone else disembark in quite this way; she invented it on her own — and without the benefit of trial and error, or the need for terabytes of labeled data.
Presumably, my daughter relied on an implicit theory of how her body moves, along with an implicit theory of physics — how one complex object travels through the aperture of another. I challenge any robot to do the same. A.I. systems tend to be passive vessels, dredging through data in search of statistical correlations; humans are active engines for discovering how things work.
Marcus thinks that a new paradigm is needed for AI that places “top down” knowledge (cognitive models of the world and how it works) and “bottom up” knowledge (the kind of raw information we get directly from our senses) on equal footing. “Deep learning”, he writes,
“is very good at bottom-up knowledge, like discerning which patterns of pixels correspond to golden retrievers as opposed to Labradors. But it is no use when it comes to top-down knowledge. If my daughter sees her reflection in a bowl of water, she knows the image is illusory; she knows she is not actually in the bowl. To a deep-learning system, though, there is no difference between the reflection and the real thing, because the system lacks a theory of the world and how it works. Integrating that sort of knowledge of the world may be the next great hurdle in A.I., a prerequisite to grander projects like using A.I. to advance medicine and scientific understanding.”
Yep: ‘superintelligence’ is farther away than we think.
From MIT Technology Review:
You may control your home with your voice, but having it speak back is often impractical. Asking Amazon’s Alexa to play a specific song, for instance, is a joy. But if you’re not sure what to listen to, the voice-only system can feel limiting. At the same time, voice assistant apps grow in number but go unused because people simply forget about them. Speaking to the [Tech Review] Download, Andrew Ng, chief scientist at Baidu, explained that, while a 2016 study by Stanford researchers and his own team showed that speech input is three times quicker than typing on mobile devices, “the fastest way for a machine to get information to you is via a screen.” He continued: “Say you want to order takeout. Imagine a voice that reads out: ‘Here are the top twenty restaurants in your area. Number one …’ This would be insanely slow!” No surprise, then, that Baidu has been working on a smart assistant device called Little Fish that includes a screen, and Amazon is also rumored to be developing a similar piece of hardware. The AI assistant revolution, it seems, may be televised.
Yep. My experience with Amazon Echo chimes with this.
This morning’s Observer column:
On 25 October, the German chancellor, Angela Merkel, wandered into unfamiliar territory – at least for a major politician. Addressing a media conference in Munich, she called on major internet companies to divulge the secrets of their algorithms on the grounds that their lack of transparency endangered public discourse. Her prime target appeared to be search engines such as Google and Bing, whose algorithms determine what you see when you type a search query into them. Given that, an internet user should have a right to know the logic behind the results presented to him or her.
“I’m of the opinion,” declared the chancellor, “that algorithms must be made more transparent, so that one can inform oneself as an interested citizen about questions like, ‘What influences my behaviour on the internet and that of others?’ Algorithms, when they are not transparent, can lead to a distortion of our perception; they can shrink our expanse of information.”
All of which is unarguably true…
From a Guardian column by Paul Mason:
these battles between regulators and the rent-seeking monopolists who have hijacked the sharing economy are, in the long term, irrelevant. The attempt to drive down cab drivers’ wages and reduce their employment rights to zero are, in their own way, a last gasp of the 20th-century economic thinking.
Because soon there won’t need to be drivers at all. Given that there are 400,000 HGV drivers in the UK, that at least a quarter of Britain’s 2.5 million van drivers are couriers, and that there are 297,000 licensed taxi drivers – that is a big dent in male employment.
The most important question facing us is not whether Uber drivers should have employment rights (they should), but what to do in a world where automation begins to eradicate work. If we accept – as Oxford researchers Carl Frey and Michael Osborne stated in 2013 – that 47% of jobs are susceptible to automation, the most obvious problem is: how are people going to live?
This morning’s Observer column:
The question on everyone’s mind as Google hoovered up robotics companies was: what the hell was a search company doing getting involved in this business? Now we know: it didn’t have a clue.
Last week, Bloomberg revealed that Google was putting Boston Dynamics up for sale. The official reason for unloading it is that senior executives in Alphabet, Google’s holding company, had concluded (correctly) that Boston Dynamics was years away from producing a marketable product and so was deemed disposable. Two possible buyers have been named so far – Toyota and Amazon. Both make sense for the obvious reason that they are already heavy users of robots and it’s clear that Amazon in particular would dearly love to get rid of humans in its warehouses at the earliest possible opportunity…
The Economist has an interesting article on how major universities are now having trouble holding on to their machine-learning and AI academics. As the industrial frenzy about these technologies mounts, this is perfectly understandable, though it’s now getting to absurd proportions. The Economist claims, for example, that some postgraduate students are being lured away – by salaries “similar to those fetched by professional athletes” – even before they complete their doctorates. And Uber lured “40 of the 140 staff of the National Robotics Engineering Centre at Carnegie Mellon University, and set up a unit to work on self-driving cars”.
All of which is predictable: we’ve seen it happen before, for example, with researchers who have data-analytics skillsets. But it raises several questions.
The first is whether this brain brain will, in the end, turn out to be self-defeating? After all, the graduate students of today are the professors of tomorrow. And since, in the end, most of the research and development done in companies tends to be applied, who will do the ‘pure’ research on which major advances in many fields depend?
Secondly, and related to that, since most industrial R&D is done behind patent and other intellectual-property firewalls, what happens to the free exchange of ideas on which intellectual progress ultimately depends? In that context, for example, it’s interesting to see the way in which Google’s ownership of Deepmind seems to be beginning to constrain the freedom of expression of its admirable co-founder, Demis Hassabis.
Thirdly, since these technologies appear to have staggering potential for increasing algorithmic power and perhaps even changing the relationship between humanity and its machines, the brain drain from academia – with its commitment to open enquiry, sensitivity to ethical issues, and so on – to the commercial sector (which traditionally has very little interest in any of these things) is worrying.
For my money, danah boyd is one of the smartest and most perceptive people around. This year she went to Davos, and wrote a stunning essay about what she saw there, and the implications thereof. Well worth reading in full, but here’s a sample:
Walking down the promenade through the center of Davos, it was hard not to notice the role of Silicon Valley in shaping the conversation of the powerful and elite. Not only was everyone attached to their iPhones and Androids, but companies like Salesforce and Palantir and Facebook took over storefronts and invited attendees in for coffee and discussions about Syrian migrants, while camouflaged snipers protected the scene from the roofs of nearby hotels. As new tech held fabulous parties in the newest venues, financial institutions, long the stalwarts of Davos, took over the same staid venues that they always have.
Yet, what I struggled with the most wasn’t the sheer excess of Silicon Valley in showcasing its value but the narrative that underpinned it all. I’m quite used to entrepreneurs talking hype in tech venues, but what happened at Davos was beyond the typical hype, in part because most of the non-tech people couldn’t do a reality check. They could only respond with fear. As a result, unrealistic conversations about artificial intelligence led many non-technical attendees to believe that the biggest threat to national security is humanoid killer robots, or that AI that can do everything humans can is just around the corner, threatening all but the most elite technical jobs. In other words, as I talked to attendees, I kept bumping into a 1970s science fiction narrative.
Yep. The problem is not just that we’re in a tech bubble. It’s that we’re in a Reality Distortion Field which leads those who dominate the tech industry to think that they are the centre of the universe, that Silicon Valley is the Florence of Renaissance 2.0. And — worse still — it’s a RDF that leads powerful and influential non-tech people to believe that maybe they’re right.
Like I said, danah’s piece is unmissable — and wise. Make space for it in your day.
Nice informative obituary by Martin Campbell-Kelly which includes stuff I hadn’t known. This,for example:
Minsky was an exceptional pianist, and in 1981 wrote a remarkable paper, Music, Mind and Meaning, that explored the cognitive processes in musical appreciation. In 1985 he became a founding member of the MIT Media Lab, an interdisciplinary research laboratory devoted to projects at the convergence of technology, multimedia, sciences, art and design.
His last book, The Emotion Machine (2006), which was written for the lay reader as much as the specialist, sought to understand and explain how “thinking” works, and to explain such phenomena as consciousness and common sense. He was the recipient of many academic awards and scientific honours, including, in 1969, the AM Turing award of the Association for Computing Machinery.
Last Sunday’s Observer column:
Last week, researchers at the artificial intelligence company DeepMind, which is now owned by Google, announced an extraordinary breakthrough: in October last, a DeepMind computing system called AlphaGo had defeated the reigning European champion player of the ancient Chinese game go by five games to nil. The victory was announced last week in a paper published in the scientific journal Nature.
So what? Computers have been getting better and better at board games for yonks. Way back in the dark ages of 1997, for example, IBM’s Deep Blue machine beat the then world chess champion, Garry Kasparov, at chess. So surely go, which is played not with six different pieces but black and white tokens – would be a pushover? Not so: the number of possible positions in go outnumber the number of atoms in the universe and far exceed the number of possibilities in chess…