The Streeb-Greebling Diaries

Sunday, September 04, 2011

A new blog

I've started a new experimental blog over here, where I'll be continuing to speculate on all sorts of topics (but probably mostly robot and AI related).

This is part of my "out of the cloud" strategy, to see if I can gain a greater degree of digital autonomy.

Tuesday, June 14, 2011

On wasting money on robots

Does the US government waste money on robotics? Probably it does.  In the past I've complained about "robot bees", and there does appear to be a lot of robotics research which seems to be poorly coordinated and of questionable value.  However, it's difficult to pick winners when it comes to research (otherwise it wouldn't be research), and it may turn out that robot bees are in fact really useful for carrying out inspections of damaged nuclear reactors or collapsed buildings or mines, in places where other types of vehicle wouldn't be suitable.

Criticising the PR2 for folding towels slowly seems pretty lame though, and this sort of childish characterisation or misrepresentation of technology projects is the main reason why on the couple of occasions where I've been asked to appear in mainstream media I've always declined.

Since as far as I'm aware Willow Garage is a private company the criticism here seems to be against the University of Berkley which carried out the aforementioned towel folding. This argument is easy to refute if you think about how many towels and other items of clothing get folded in the US each year, how much time that takes and what the economic value of that time is.  I suspect that it's a really big number.  If you can produce a general purpose textile folding robot at the right price level then this is potentially a huge industry.  Also the speed of folding on this research robot isn't a big issue.
"Computationally,the throughput was primarily limited by the optical flow calculation: the NVIDIA GTX 295 GPU provided a through-put of 1 optical flow computation per 5 seconds, and the processing of each frame requires either one or two optical flow computations, depending on whether a 2-D grasp point detection is made"
This process only needs to be speeded up by an order of magnitude or so for the system to be more practical, and it seems reasonable to believe that this will happen within less than a decade, especially given the amenability of vision algorithms to parallelisation. Also there are many possible alternative ways to calculate optical flow.  Beyond a certain point it makes no difference how fast the vision computations are, because physics sets the pace.

Also the appearance of the Microsoft Kinect sensor, which happened after the towel folding research was done, may make all the above limitations irrelevant.

As far as government prudence goes probably far more substantial savings could be made in the military and security areas, where there seems to be a lot of profligate spending on projects of questionable value even within their particular scope of interest. Is there really any need for ICBMs to be sitting unused in silos for decades, long after the cold war ended? Was it really necessary for the US to build a second smaller version of the space shuttle? Is the US tax payer getting value for money out of the Afghanistan war? All of these things are basically just parasitic costs.

Friday, June 03, 2011

One more personal robot

Another personal robot, similar to the Turtlebot, but with a simple arm and gripper.

The price is $1200, so this is an affordable machine which anyone could buy and start developing software for. Having a somewhat standardised and affordable PC based robotics platform on which to do research and development, in the style of early home computers, has been something which was aspired to for the last couple of decades, but is only now a realistic proposition. The only drawback is the timing of its introduction - something which Ray Kurzweil talks about in regards to the appropriateness of inventions - isn't very optimal since there aren't currently many people with much disposable income. As a consequence it's possible that new products like this may fail simply because of the ambient economic conditions, but the longer term prospects look good.

Monday, May 09, 2011

Reinventing money

Money is a very old invention. As soon as you go beyond subsistence farming, such that you're creating a surplus which you might want to exchange then money becomes an obvious trading technology. Money can be represented by anything, so long as that thing is difficult to reproduce without expending effort commensurate with things of value and it's agreed upon within a community as a token of exchange.

There are problems within the current money system which result in the creation of tokens without corresponding goods or services - a kind of fraud - and monetary reform seems to be a desirable goal to work towards. Cryptocurrencies, although really just another kind of token, offer hope that the current bugs in the monetary system have a technological solution. This takes the creation of money and the administration of its movement out of the hands of a few cronies and makes it algorithmic in nature - independent from any particular individual, organisation or government.

Here is a discussion about Bitcoin, which is a type of cryptocurrency. Whether Bitcoin succeeds or not in becoming a common method of value exchange remains to be seen, but cryptocurrencies in general seem to be the way to go with regard to eliminating the current inefficiencies and inconveniences in the money system. Efficient online micropayments is just one of the possible applications.

There are a couple of problems with Bitcoin that I can think of.

The most obvious one is that the people who are likely to become the wealthiest are those with large computational resources to do the mining in the early years after the currency was initiated, and people with such resources are likely to be already wealthy groups or institutions - such as financial trading companies who own supercomputing clusters.

Another issue is that there is no algorithmic mechanism within Bitcoin to encourage circulation of money, and this is compounded by the fact that after twenty years or so no further Bitcoins will be created. A possible failure case is that many users of this currency just sit on it, like a stash of gold. Instead there should probably be a modest but constant inflation rate (2% per year for example), with no upper limit. This would still allow a high degree of currency predictability which is good for business, but discourage organisations or individuals from sitting on their heap of gold for too long, otherwise it gradually loses value.

Solar Fire

This looks quite neat, and would be especially suited for equatorial regions. The main limitations are that you need a water supply, and that in this case it looks as if the adjustment of the mirrors throughout the day is purely manual. There's nothing very high technology here, and systems like this could have been used to generate power at any time in the previous century or more.

It's also possible to obtain mirrored tape, so the mirrors could be made very cheaply using any common material such as plastic or wood for the backing.

There is some free software (GPL) available to calculate the mirror angles, and in principle this simulator could also be extended to include an estimate of power output given the current or average cloud cover. One of the problems with solar systems is being able to make an estimate of whether it would be capable of producing a usable amount of power at a particular location, and that's where a simulator could help in making investment decisions.

Of course this is likely to be a higher maintenance system than photovoltaics and batteries, but at current prices it's far cheaper than that. Cheap photovoltaic solar has been promised for many years, and so far hasn't materialised, but new ways of producing solar panels should mean that they eventually become the lowest cost and least maintenance solution.

Thursday, May 05, 2011

Cleaning up

Assuming that this demo can be taken at face value as representative of typical performance then this is pretty impressive, and probably indicative of early uses of service robotics.

Even if the robot is limited to detecting cylindrical objects that's still quite a good achievement. It's also possible to envisage other light duties, such as sweeping, watering, opening and closing curtains or blinds, surface cleaning, feeding pets and dusting. Individually all of these jobs might be automated using special purpose machines, but the main advantage of a general purpose robot is that you don't need to buy, learn to use and maintain lots of separate gadgets.

Service robotics is probably going to be more of an evolution than a revolution. I've learned my lesson with regard to making predictions, and just because technology exists at a certain price level does not mean that it will be used or commercialized. For this kind of application there are additional psychological issues, such as displacing habituated practices, being confident that the robot isn't going to damage anything or be taken over by Anonymous, and so on. There might also be issues around home insurance and possible fire hazards, so all of these things need to be worked out and they will probably work themselves out slowly. Nevertheless, service robotics - which includes telepresence and teleoperation - represents a huge and so far mostly untapped market capable of delivering real value and transforming many aspects of the overall economy.

On not voting in the referendum

Under ordinary circumstances I usually always vote, even if it's in a pseudo-random way by the flip of a coin or some other arbiter.  The 2010 general election was my most systematic voting attempt to date, via data mining manifestos, but since the victors promptly dumped most of their manifesto pledges shortly after arrival in office the whole exercise turned out to be one of utter futility anyway.

Referenda in the UK are a really unusual phenomena, and there havn't been any within my adult lifetime.  In this case it's over a proposed change to voting procedure - essentially the algorithm used to count votes. The choice is between a simple majority, which in computation terms would be known as a winner takes all (WTA) algorithm, or an "alternative vote" (AV) system which is more akin to a weighted sum or probabilistic method based on ranking of candidates by order of preference.

No matter which algorithm is selected they both follow a democratic principle.  I've also voted under both systems in the past, so it really isn't any big issue.  Elections of Euro MPs uses something similar to, or the same as, the proposed AV algorithm.  So in this case no matter which is selected that's ok with me, provided that the algorithm is consistently applied, and I'll therefore abstain from voting as an indicator of no preference.

There seems to be a lot of political-type stuff going on in the UK at the moment, which is mostly connected to the current situation of stagflation and the fallout from decisions over public service and education cuts, so there are many things which people want to change or preserve.  In this context offering a referendum on an issue which very few people care about seems rather perverse.

For more ranting on this topic, and photos of some of the referendum campaign leaflets, see here.

War and information

I think the final point about wars, and the propaganda required to sustain them, might be a very optimistic sign for the future. The Iraq war which started in 2003 is a good example of this kind of scenario, where more transparency might have perhaps fatally undermined the narrative emerging from Blair and his co-workers.

Imagine for example if two minutes after Colin Powell delivered his infamous Iraq speech to the UN people at the claimed locations of the blurry missile silo photos were able to tweet a link to a live video stream from the same site showing that no such silos existed. With advanced editing tools and augmented reality fake videos will become increasingly possible to produce, but if enough corroborating crowdsourced evidence exists this might be sufficient to shift the narrative in a very different direction.

Another strategy which sustains wars, rather than generating a highly elaborate deception narrative, seems to be to avoid talking about them altogether. So for example in the previous five years, apart from the exposure by Wikileaks, there has been extremely minimal media coverage of the war in Afghanistan apart from occasional reports of soldiers being killed. Essentially Afghanistan has become a forgotten war as far as public awareness is concerned, and the propaganda used to prop it up seems to be very minimalistic indeed.

A reason for optimism is that we now live in a world of ubiquitous information, and maintaining largely or wholly fictitious narratives is going to become an increasingly difficult and high maintenance task. The fragility of propaganda narratives in an environment where multiple independent countervailing observations are available, often in real time, is likely to make the buildup to wars more difficult if they are to enjoy any degree of support amongst the populace.

So, one possible response to this in future might be wars which are launched rapidly, with almost no propaganda buildup, and in a cultural environment in which citizens are encouraged to be ultra-loyalists who bear more emotional allegiance to the flag than to any rationalizing narrative involving ideas about democracy or WMD.

Wednesday, April 20, 2011

A strange radio announcement

Listening to the radio whilst driving I heard a very peculiar advert. At first I thought it was a spoof, but it seemed not to be. The announcement went something like:

"Do you suspect a member of your family of being a terrorist? If so, report them to the police now! If the accusation turns out to be false, then no harm is done. This message is on behalf of Durham county council".

I don't spend a lot of time listening to the radio, so perhaps this is a common message. The first thing which came to mind was that this sounds like something you might hear announced in a totalitarian state - inform on your friends and relations. The second thing which occurred to me is that making such accusations about members of your family is far from a harmless activity, and could have very significant consequences for that individual even if they're innocent.