Tag Archives: artificial intelligence

2014 Report Card

It’s taken me a while to get out a “year in review” post for 2014, but anyway, here it is. This won’t be a masterpiece of the essay form. I’m just going to ramble on about some interesting trends and themes from the year, along with a few relevant links.

The critical question this blog tries to answer is, is our civilization failing or not? I’ll talk about our human economy, our planetary system, and make some attempt to tie the two together.

Overall Human Health and Wellbeing. First, there are some very happy statistics to report. For example, worldwide child mortality has dropped almost by half just since 1990. What better measure of progress could there be than more happy, healthy childhoods? And it’s not just about increasing wealth – people in developing countries today have much better health outcomes at the same level of wealth compared to developing countries of the past (for example, Indonesia today vs. the United States when it passed the same income level). It’s hard to argue against the idea that economic growth and technological change have obviously eliminated a lot of human suffering. So, I think the important questions are, will these trends continue? Is the system stable? Can the natural environment continue to support this trend indefinitely? There may also be an important question of whether we had the right to exploit the natural environment to get us to the point where we are now, but that is an academic question at this point.

Financial System Instability. Let’s talk about the stability of our human economic system. The U.S. economy may finally seem to be picking up from the aftermath of the severe 2007-8 financial crisis, but it is certainly far below where it would be if that hadn’t happened and the prior growth trend had just continued since then. The rest of the world isn’t doing so well, however – Europe and Japan are looking particularly slow if not in an outright deflationary spiral, at the same time developing countries appear to be slowing down. Some are calling this a “new normal” for the world economy. More scary than that, the industry-written regulations and perverse incentives allowing the excessive risk taking that caused the crisis have not been fully addressed and the whole episode could recur in the short term.

Thoughts on Ecosystem and Economic “Pulsing”. 2007-8 was a textbook financial crisis – although it was caused by novel forms of money and risk taking beyond the direct reach of government regulators and central banks, it was not that different from crises caused by plain old speculation and over-lending back when there were no central banks around. It’s hard to draw a direct link from the financial crisis to ecosystem services, climate change, or natural resource scarcity. However, if we think about natural ecosystems, they are resilient to outside stressors up to a point – say, moderate fluctuations in temperature, hydrology, or pressure from non-native species. However, say a major fluctuation happens such as a major flood or fire that causes serious damage. In the absence of major outside stressors, the system will eventually recover to its original state, but in the presence of major outside stressors, even if they did not cause the flood or fire, it may never bounce back all the way. In the same way, our human economy may appear resilient to the effects of climate change, ocean acidification, soil erosion, and so forth for a long time, but then when something comes out of left field, like a major financial crisis, war, or epidemic, we may not be able to recover to our previous trend. This probably also applies to the effects of technology on employment, as discussed below. In the absence of major shocks coming from outside the system, we’ll see a long, slow slide in employment and possibly a long, slow rise in energy and food prices, with so much noise in the signal that it will be easy for the naysayers to hold sway for long periods of time. But when those major events happen, we may see sudden, painful changes that we have no obvious way of mitigating quickly.

Technological Change: Artificial Intelligence, Robots, Automation, and Employment. After decades of slow but steady progress, these technologies are really coming into their own. Robots are being used to keep miners in line and to drive cars, for example. Manufacturing has become a high-tech industry. As computers and machines get better at performing more and more skilled jobs (book-keeping is one example), there is gradually less demand for the medium-skilled workers who used to do those jobs. High-skilled workers like computer programmers are doing very well, although I presume the automation will gradually creep higher and higher up the chain, so today’s safer jobs will be less safe tomorrow. At the same time these medium-skilled workers in developed countries are getting squeezed out, developing countries are not benefiting like they used to from their large pools of low-skilled workers as manufacturing becomes more and more automated, and can be done cost-effectively closer to consumers in richer countries.

Will our society recognize and solve this employment problem? American corporate society, and its admirers around the world, are unlikely to. Something very similar to this happened with agricultural automation in the early- to mid-20th century, and with globalization in the mid- to late-20th century. As agriculture became more automated, many displaced workers moved from rural areas in the U.S. southeast to urban areas in the U.S. northeast, looking for factory work. Unfortunately, the factory jobs that existed previously were being moved to developing countries with abundant low-wage labor. The pockets of poverty, unemployment, and social problems created by these forces have not been adequately addressed to this day. To the individual worker, it doesn’t much matter whether your job is being taken by a local robot or an overseas human. Unemployment created by technological forces today could resemble what was created by globalization yesterday, only on a much larger scale. We can only hope that the larger scale will drive real political solutions, such as better education and training, sharing of available work, and more widespread ownership of the labor-saving technology.

Of course, one of the earliest and probably the most shameful example of a modern capitalist system generating wealth for an elite few at the expense of workers is the American slavery system of the 18th and 19th centuries. We just can’t trust amoral, self-interested private enterprise to maximize welfare in the absence of a strong moral compass coming from the larger society. Let’s stop pretending otherwise.

Another example of extreme corporate immorality: Public apathy over climate change in the U.S. may have been manufactured by a cynical, immoral corporate disinformation campaign over climate change taken right out of the tobacco companies’ playbook.

The Gospel of Shareholder Value. There is an important debate over whether people who run corporations have any ethical responsibility to anything other than profit seeking. Well duh, everyone on Earth has an ethical responsibility. Case closed, as far as I’m concerned. There is even evidence that the ideology of profit maximization is a drag on innovation. Except billions of people out there who have worshiped at business schools would disagree with me. And I don’t want to offend anyone’s religion. Noam Chomsky had a quote that I particularly loved, so I am going to repeat it here:

In market systems, you don’t take account of what economists call externalities. So say you sell me a car. In a market system, we’re supposed to look after our own interests, so I make the best deal I can for me; you make the best deal you can for you. We do not take into account the effect on him. That’s not part of a market transaction. Well, there is an effect on him: there’s another car on the road; there’s a greater possibility of accidents; there’s more pollution; there’s more traffic jams. For him individually, it might be a slight increase, but this is extended over the whole population. Now, when you get to other kinds of transactions, the externalities get much larger. So take the financial crisis. One of the reasons for it is that — there are several, but one is — say if Goldman Sachs makes a risky transaction, they — if they’re paying attention — cover their own potential losses. They do not take into account what’s called systemic risk, that is, the possibility that the whole system will crash if one of their risky transactions goes bad. That just about happened with AIG, the huge insurance company. They were involved in risky transactions which they couldn’t cover. The whole system was really going to collapse, but of course state power intervened to rescue them. The task of the state is to rescue the rich and the powerful and to protect them, and if that violates market principles, okay, we don’t care about market principles. The market principles are essentially for the poor. But systemic risk is an externality that’s not considered, which would take down the system repeatedly, if you didn’t have state power intervening. Well there’s another one, that’s even bigger — that’s destruction of the environment. Destruction of the environment is an externality: in market interactions, you don’t pay attention to it. So take tar sands. If you’re a major energy corporation and you can make profit out of exploiting tar sands, you simply do not take into account the fact that your grandchildren may not have a possibility of survival — that’s an externality. And in the moral calculus of capitalism, greater profits in the next quarter outweigh the fate of your grandchildren — and of course it’s not your grandchildren, but everyone’s.

Our Ecological Footprint. WWF issued an updated Living Planet Report in 2014 suggesting that our annual consumption of natural resources (including the obvious ones like energy and water extraction, straightforward ones like the ability to grow food, but also the less obvious ones like ability of the oceans and atmosphere to absorb our waste products) is continuing to exceed what the Earth can handle each year by at least 50%. We’re like spoiled trust fund babies – we have such incredible resources at our disposable, we never learn to live within our means and one day the resources run out, even if that takes a long time. As we recover from the financial crisis, we have a chance to do things differently, but the connections are not being made to the right kinds of investments in infrastructure, skills, and protection of natural capital that would set the stage for long-term sustainable growth in the future.

Other Big Stories from 2014:

  • World War I. 100 years ago, World War I was in full swing. Remember The Guns of August? Well, that was August 1914 they were talking about. Let’s hope we’re not about to blunder into another conflict. But (and I’m cheating a little here because I read this in 2015), the World Economic Forum named “interstate conflict” as both high probability and high consequence in its global risk report.
  • Ebola. Obviously, Ebola was a very bad thing that happened to a whole lot of people. To those of us lucky enough that we weren’t directly in its path, it is a chance to selfishly reflect whether Ebola or something even worse could be coming down the pike. Let’s hope not.
  • Severe Drought and Water Depletion in the Western U.S.: California has been in the midst of a historic drought, although they got some rain recently. Some are describing this as the new normal. Besides rainfall, glaciers, snowpack, and groundwater all seem to be disappearing in some important food-growing areas.
  • Solar grid parity is here! At least some places, some times…

Conclusion. Yes, I think we are on a path to collapse if nothing changes. And I don’t see things changing enough, or fast enough. There are glimmers of hope though. Lest you think I offer only negatives and no solutions, here are two solutions I harp on constantly throughout the blog:

  • Green infrastructure. This is how we fix the hydrologic cycle, close the loop on nutrients, begin to cleanse the atmosphere, protect wild creatures and genetic diversity, and create a society of people with some sense of connection to and stewardship over nature. Don’t act like it’s such a big mystery. It’s known technology. There has been plenty written about trees, design of wildlife corridors and connectivity, for examples. There is simply no excuse for cities to do such a crappy job with these things.
  • Muscle-Powered Transportation. Cars are clearly the root of all evil, the spawn of Mordor, as I pointed out several times (sorry, I just sat through 6+ hours of Hobbit movies). Unless you are perhaps that rare hobbit who can own a car without your morals being completed corrupted by its evil powers. But for the rest of us, I explained several times why getting rid of cars would be good. Here is just one example:

One of the most important things we can do to build a sustainable, resilient society is to design communities where most people can make most of their daily trips under their own power – on foot or by bicycle. It eliminates a huge amount of carbon emissions. It opens up enormous quantities of land to new possibilities other than roads and parking, which right now take up half or more of the land in urban areas. It reduces air pollution and increases physical activity, two things that are taking years off our lives. It eliminates crashes between vehicles, and crashes between vehicles and human bodies, which are serial killers of one million people worldwide every year, especially serial killers of children. It eliminates enormous amounts of dead, wasted time, because commuting is now a physically and mentally beneficial use of time. There is also a subtle effect, I believe, of creating more social interaction and trust and empathy between people just because they come into more contact, and creating a more vibrant, creative and innovative economy that might have a shot at solving our civilization’s more pressing problems.

“robot scientist”

This article is about a robot that can somehow develop its own experiments to test new drugs.

There is an urgent need to make drug discovery cheaper and faster. This will enable the development of treatments for diseases currently neglected for economic reasons, such as tropical and orphan diseases, and generally increase the supply of new drugs. Here, we report the Robot Scientist ‘Eve’ designed to make drug discovery more economical. A Robot Scientist is a laboratory automation system that uses artificial intelligence (AI) techniques to discover scientific knowledge through cycles of experimentation. Eve integrates and automates library-screening, hit-confirmation, and lead generation through cycles of quantitative structure activity relationship learning and testing. Using econometric modelling we demonstrate that the use of AI to select compounds economically outperforms standard drug screening. For further efficiency Eve uses a standardized form of assay to compute Boolean functions of compound properties. These assays can be quickly and cheaply engineered using synthetic biology, enabling more targets to be assayed for a given budget. Eve has repositioned several drugs against specific targets in parasites that cause tropical diseases. One validated discovery is that the anti-cancer compound TNP-470 is a potent inhibitor of dihydrofolate reductase from the malaria-causing parasite Plasmodium vivax.

my vacation reading list

Here’s my vacation reading list, just in case anyone is interested:

The Fear Index (about a hedge fund running an automated trading algorithm – way more fun to read than it sounds)

Count Zero (the sequel to William Gibson’s Neuromancer – great dystopian technology and artificial intelligence fiction although similar to Neuromancer, he uses a lot of made-up slang, weird sentence structure, and multiple points of view that keep me from getting really absorbed in the book)

Monster Hunter International (what it sounds like – pure escape fiction, basically an updated Ghostbusters with a Walking Dead-like gun fetish, violent but lighthearted)

MaddAdam (the third and final book in Margaret Attwood’s Oryx and Crake series)

The original I Am Legend (because why not, I seem to be on a monster kick – if I don’t get to this one, I may save it for around Halloween)

can/should machines run the world?

From “futureoflife.org“, here is a short excerpt on future directions of artificial intelligence research.

What policies could help increasingly automated societies flourish? For example, Brynjolfsson and McAfee [12] explore various policies for incentivizing development of labor-intensive sectors and for using AI-generated wealth to support underemployed
populations. What are the pros and cons of interventions such as educational reform, apprenticeship programs, labor-demanding infrastructure projects, and changes to minimum wage law, tax structure, and the social safety net [26]? History provides many examples of subpopulations not needing to work for economic security, ranging from aristocrats in antiquity to many present-day citizens of Qatar. What societal structures and other factors determine whether such populations flourish? Unemployment is not the same as leisure, and there are deep links between unemployment and unhappiness, self-doubt, and isolation [34, 19]; understanding what policies and norms can break these links could signifi cantly improve the median quality of life. Empirical and theoretical research on topics such as the basic income proposal could clarify our options [83, 89].

Please follow the link if you would like to see the references.

Also see The Evitable Conflict, the last story in Asimov’s I, Robot. No, not the Will Smith movie! Just put that out of your head and read the book, it’s short. Anyway, in that story humans have handed control of the global economy over to “Machines”, artificial intelligences which are supposed to smooth everything out and keep everything perfectly balanced. Only it doesn’t work out exactly that way, and the humans are trying to figure out why not, and whether or not they should try to do anything about it. This story was written in 1950, so it should be in the public domain soon. Another great old story that is in the public domain is Forster’s The Machine Stops. In that story (from 1909!), a machine runs the entire world, and is supposed to smooth everything out and keep everything perfectly balanced. Only it doesn’t work out exactly that way. Or, it does for awhile, but then the machine… well, I don’t want to spoil it for you. It’s free and it’s short, so give it a read!

poker bots

It never occurred to me before that there might be computers playing in online poker games, but it makes sense.

Bowling says the program isn’t much of a threat to online gamblers. Heads-up, Limit Hold’em is not the variety of poker most people play. But he does believe that “poker bots” are trying to win in online game rooms. “My guess is there are probably quite strong poker bots out there,” he says. “But you’re not going to hear a lot of talk about them.”

robots learning to use tools

This article from KurzweilAI is about robots learning to use tools by watching videos on the internet.

This paper presents a system that learns manipulation
action plans by processing unconstrained videos from
the World Wide Web. Its goal is to robustly generate the sequence
of atomic actions of seen longer actions in video in
order to acquire knowledge for robots. The lower level of the
system consists of two convolutional neural network (CNN)
based recognition modules, one for classifying the hand grasp
type and the other for object recognition. The higher level
is a probabilistic manipulation action grammar based parsing
module that aims at generating visual sentences for robot
manipulation. Experiments conducted on a publicly available
unconstrained video dataset show that the system is able
to learn manipulation actions by “watching” unconstrained
videos with high accuracy

Robots, pay attention – I just Googled videos on “how to clean a bathroom” and got over 33 million results, so get to work!

December 2014 in Review

At the end of November, my Hope for the Future Index stood at -2.  I’ll give December posts a score from -3 to +3 based on how negative or positive they are.

Negative trends and predictions (-12):

  • When you consider roads, streets, and parking, cars take up more space in cities than housing. (-2)
  • The latest on productivity and economic growth: Paul Krugman says there is risk of deflationary spirals in many countries, and the U.S. economy is nothing to right home about. (-1)
  • There are a few legitimate scientists out there warning of sudden, catastrophic climate change in the near future. (-1)
  • Automation (meaning robots and AI) is estimated to threaten 47% of all U.S. jobs. One area of active research into automation: weaponry. Only one negative point because there are also some positive implications. (-1)
  • Margaret Atwood’s Year of the Flood is a depressing but entertaining reminder that bio-apocalypse is possible. (-2)
  • Before the recent rains, the drought in California was estimated to be a once-in-1200-years event. Major droughts in major food growing regions are not good news, especially with depletion of groundwater, and loss of snowpack and glaciers also in the news. (-2)
  • William Lazonick argues provides evidence that the rise in the gospel of shareholder value correlated with the growth slowdown that started in the 1970s – his explanation is that before that, retained earnings were a cornerstone of R&D and innovation in the economy. Loss of a point because it’s good to hear a dissenting voice, but the economy is still run by disciples of the profits for now. (-1)
  • Elizabeth Warren and Bernie Sanders are warning that the U.S. financial system may still be dangerously unstable. (-2)

Positive trends and predictions (+6):

  • There are some new ideas out there for teaching computer programming, even to young children: Loco Robo, Scratch, and for-profit “programming boot camps”. (+1)
  • You can now get genetically customized probiotics for your vagina. (+1)
  • There are plenty of ideas and models out there for safe, walkable streets, some as simple as narrower lanes. But as I point out, the Dutch and Danish designs are pretty much perfect and should just be adopted everywhere. (+1)
  • I linked to a new video depicting Michael Graves’s idea for “linear cities“. These could be very sustainable ecological if they meant the rest of the landscape is left in a mostly natural condition. I am not as sure about social sustainability – done wrong, they could be like living in a mall or subway station. This was one of my all-time more popular posts. (+1)
  • There are new algorithms out there for aggregating and synthesizing large amounts of scientific literature. Maybe this can increase the returns to R&D and help boost innovation. (+1)
  • There will be several international conferences in 2015 with potential to make real progress on financial stability and sustainability. The phrase “deep decarbonization” has been uttered. (+1)
  • Some evidence suggests that the oceans have absorbed a lot of global warming over the past decade or so, preventing the more extreme range of land surface warming that had been predicted. This is a good short- to medium-term trend, but it may not continue in the long term. (+0)

change during December 2014: -12 + 6 = -6

Hope for the Future Index (end of December 2014): -2 -6 = -8

robots robots robots!

Yes, there’s a robot bartender now.

No word on whether this is a bar where everybody knows your name. I suspect not. Here’s a much longer academic study on which occupations are likely to be most affected by computerization/automation in coming decades.

According to our estimates around 47 percent of total US employment is in the high risk category. We refer to these as jobs at risk – i.e. jobs we expect could be automated relatively soon, perhaps over the next decade or two. Our model predicts that most workers in transportation and logistics occupations, together with the bulk of office and administrative support workers, and labour in production occupations, are at risk. These findings are consistent with recent technological developments documented in the literature. More surprisingly, we find that a substantial share of employment in service occupations,where most US job growth has occurred over the past decades (Autor and Dorn, 2013), are highly susceptible to computerisation. Additional support for this finding is provided by the recent growth in the market for service robots (MGI, 2013) and the gradually diminishment of the comparative advantage of human labour in tasks involving mobility and dexterity (Robotics-VO, 2013).

The paper has a detailed appendix where you can look up your specific occupation if you are so inclined. In also has a detailed lesson on the history of technology and labor markets, if you are inclined to read that.

Finally, the Pentagon is also worried about falling behind the curve on automation:

Hagel and DOD officials have been discussing the so-called third offset strategy for months without giving up any specifics as to how they intend to achieve offset innovation. In his speech, Hagel provided a small glimpse into the fields that will attract special Defense Department attention as part of the strategy: “robotics, autonomous systems, miniaturization, big data, and advanced manufacturing, including 3-D printing.”

automated aggregation of scientific literature

I am intrigued by this example from Stanford of computerized review and synthesis of scientific literature:

Over the last few years, we have built applications for both broad domains that read the Web and for specific domains like paleobiology. In collaboration with Shanan Peters (PaleobioDB), we built a system that reads documents with higher accuracy and from larger corpora than expert human volunteers. We find this very exciting as it demonstrates that trained systems may have the ability to change the way science is conducted.

In a number of research papers we demonstrated the power of DeepDive on NMR data and financial, oil, and gas documents. For example, we showed that DeepDive can understand tabular data. We are using DeepDive to support our own research, exploring how knowledge can be used to build the next generation of data processing systems.

Examples of DeepDive applications include:

  • PaleoDeepDive – A knowledge base for Paleobiologists
  • GeoDeepDive – Extracting dark data from geology journal articles
  • Wisci – Enriching Wikipedia with structured data

The complete code for these examples is available with DeepDive.

Let’s just say an organization is trying to be more innovative. First it needs to understand where its standard operating procedures are in relation to the leading edge. To do that, it needs to understand where the leading edge is. That means research, which can be very tedious, and time consuming. It means the organization is paying people to spend time reviewing large amounts of information, some or even most of which will not turn out to be useful. So a change in mindset is often necessary. But tools that could jump start the process and provide short cuts would be great.

This is my own developing theory of how an organization can become more innovative: First, figure out where the leading edge is. Second, figure out how far the various parts of your organization are from the leading edge. Third, figure out how you are going to bring a critical mass of your organization up to the leading edge – this is as much a human resource problem as an innovation problem. Fourth, then and only then, you are ready to try to advance the leading edge. I think a lot of organizations have a few people that do #1, but then they skip right to #4. Then that small group is way outside the leading edge while the bulk of the organization is nowhere near it. That’s not a recipe for success.