May 30, 2005

Mapping What We Know We don't Know

Mapping Great Debates: Can Computers Think? is an interesting attempt to map the arguments and counter-arguments in the AI debate in the form of 7 poster-sized argumentation maps. Positions and arguments are represented by boxes, linked by arrows shoing how they support and dispute each other.

This is a very good idea, although the space and layout requirements for many interesting debates might be extreme (imagine a capitalism-marxism map) even at this kind of overview level. Some of the other applications mentioned in the To Think Bigger Thoughts lecture by Robert E. Horn such as GMO safety would also fit this kind of map very nicely. The only problem is that in this case the debate and science is moving so quickly the map would need to be software and easily updateable, in turn making it harder to get the high quality physical overview you get from a poster. Information walls are wonderful things, but few can afford having a huge hi-res monitor cover a wall. Sure, a web might do the job, but much of the overview power is lost. This can be seen in the still wonderful consciousness debate map. Here the debate becomes neatly divided into sections one cannot initially see the shape of. It is even worse in the GMO map, which however has very nice literature citations.

I am reminded by the attempts to break away from the strip structure in online comics by Scott McCloud. These diagrams might be conceptually constrained by a legacy of paper and then web, which unnecessarily limits their expressiveness. It is not inconceivable that they could be published as zoomable PDF or (shudder) flash. Actually, McCloud seems to already have solved the problem. Here the debate could be turned into a zoomable information mural.

Overall I have a bit of problem with the design of these argument graphs; the focus appears to have been more on the logic and sending simple iconic messages (a very good start) than turning it into information art. The heavily lined boxes, indistinct arrow icons and colorings grate on my sensitive nerves. I wonder what Edward Tufte would do with them?

But enough quibbling, this is very good stuff. I would love to do something similar for transhumanism, helping disambiguate its various strands and internal debates.

While we are still exploring Horn's site, I would like to point out the map over
What we don't know in the science of genetically modified crops and food
. It is especially nice in that it immediately suggests research programs to bridge the "stepping stones" in the dark area.

The whole dark graph reminds me of that profound (and often ridiculed) insight from Donald Rumsfeld:

"Reports that say that something hasn't happened are always interesting to me, because as we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns -- the ones we don't know we don't know."

Now we just need to invent better ways of mapping what we don't know what we don't know.

Posted by Anders3 at 02:21 PM | Comments (1127)

May 27, 2005

The Art of Line Drawing

This week I attended the International Conference on Sport Medicine Ethics organised by the Stockholm Bioethics Centre and the Uehiro Centre for Practical Ethics at Oxford University. I of course briefly blogged about it at CNE Health, but here is the long version.

I'm utterly uninterested in sports, except for doping. Doping is many ways a testbed of future human enhancement. It also gives enhancement a bad name, and opposition to doping is also the source of many rules that prohibit enhancing drugs. Hence the ethics and politics of doping is a highly relevant area, with many interesting issues that carry over to other enhancements.

Sports Medicine

Sports medicine itself is interesting, since it has goals that are different from ordinary medicine. As Christian Muthe explained, normal health care seeks to secure a certain level of health in a just way. Sports medicine seeks to secure a level of health conducive to athletic performance, which is a far greater level than ordinary health and with no apparent requirement to be just - here there is no rationing of health care resources. Claudio Tamburrini pointed out that sports medicine in some ways is more paternalistic towards athletes than would ever be allowed in public health care; the privacy and autonomy of the patients are infringed to a great extent by doping testing and obligatory treatments. He suggested a patient concept based on being exposed to medicine: a patient is someone who is vulnerable in relation to the medical profession, and hence a bearer of patients rights. This fits well in with a health consumer perspective, regardless of why a person gets into touch with the medical professions.

My own lecture was a brief discussion about how the development of enhancing medicine for non-sports use will force rethinks of doping and enhancement within sports. My basic argument is that use and acceptance of enhancing treatment in society is increasing (but might not become the total mainstream - it is not inconceivable to see a stable split between bioconservatives and dynamists), and that many therapies while not developed for enhancement purposes also have enhancing effects. The result is that non-athletes might get far better therapies and health than athletes, with the doping regulations isolated within an enhancement-accepting society.

Drawing Lines and the Spirit of Sports

Susan Shewin argued against allowing genetic modifications and suggested that society should actively resist their use. Her main argument was that competition would make it irrational not to use genetic doping, everybody would be forced to take it and hence allowing its use would narrow autonomy. She also pointed out that the age where such modifications were most likely to be taken would be adolescence, a period not known for its carefully weighed judgements. While her argument about the bandwagon effect likely is true (especially given Sören Holm's lecture), it is not clear to me that one can enhance autonomy by banning anything. Here our difference might be due to my more individualistic perspective.

A more serious problem, based in the relational theory she used herself, is that active resistance could also reduce autonomy. It would require control mechanisms (formal or informal) that demand compliance and punish disobedience. This would produce obedience, including a reduction in free speech and thought - daring to speak out against the official view is often seen as disloyalty to the group and draw a suspicion that the dissident might be a doper. Hence there would be a tendency to make people officially support the control, strengthening it and leading to a groupthink situation. Add to this the public choice aspects of control organisations that would have an incentive to continue and expand this regime, and you get a fairly serious narrowing of autonomy. Not unlike the current anti-doping regime, of course.

Søren Holm critiqued the idea that doping could be allowed under medical control from an economic/game theoretic perspective. The core of his argument was that if doping was allowed, there would still be an incentive to hide the fact that one had a new kind of doping since it would give oneself an advantage, and this would slow the promised benefits of legalisations such as development of safer enhancing drugs. In the end, the legalised doping system would have to get back to checking athletes anyway, not winning much. While I think his example was a bit contrived, he has an important point in that legalisation doesn't automatically solve the problems.

Åke Andrén Sandberg stated his four main arguments against doping. The key argument was that it was "unhuman" due to its risks. The second argument was that given the tax funding of sports doping would likely lead to a withdrawal of such funding and a decrease in the amount of children participating in sports. The third argument was that the rules of sport are intended to make it fun to participate in and watch, and that doping is incompatible with meaningful competition. The fourth argument was the observation that vast majority of athletes do not want legalised doping.

The fourth argument might be weaker than stated due to the conformity bias discussed above. I do not see the other arguments as insurmountable either: safe enhancements could be legalised, sports might not need tax funding, children might not need traditional sports organisations for their exercise, and it is entirely possible that one could have fun playing enhanced sports too.

Christian Lenk discussed equality of opportunity and sports. He based his talk on Boorse's concept of health ("the state of an organism is theoretically healthy, i.e., free from disease, in so far as its mode of functioning conforms to the natural design of that kind of organism") and a discussion of equality. Sport without equality of opportunity would be unfair, which might make it seem reasonable to allow enhancement to produce this equality. But he argued against this, claiming that these enhancements would just introduce new inequalities of opportunities and reduce the individual contribution to sports prestations. It was never clear to me why high technology was so inimical to equality of opportunity, nor why it would reduce the need for the athlete to exert himself in training.

Torbjörn Tännsjö examined what, beyond legality and safety, was wrong with doping. He reached the conclusion that sports has an ethos of searching for the perfect human being and allowing us vicariously share his experiences and celebrate him - quite often a position gained through genetic differences. But one could create a new ethos aiming to find where the limits of the "new human nature" go, combining the enjoyment of competition, admiration of the science involved and see the athletes as testers of this new condition.

Transhumanism and Sports

Michael McNamee from University of Wales Swansea attacked transhumanism. Or rather, tried to, because it is such a vague term that it is hard to do a real philosophical critique of it - the transhumanist can always claim the critique doesn't apply to this particular brand of transhumanism or the current version. I think transhumanism needs a lot of good criticism to shape up, so I enjoyed his talk immensely.

The key problem in his opinion is that transhumanism follows the logic of expansion but has no clear aims. The object of enhancement is "better" posthumans, but the value scale is often undefined. This produces a "slippery slope to an arbitrary destination". I think he has a good point, even if I look at it more positively. The lack of clear value scale is largely due to the broadness of transhumanist views and the scarcity of solid philosophical systems. I would claim that the system embodied in the extropian principles (together with the encessary liberal-humanistic context) does define a direction to move in and ways of discerning certain "enhancements" as undesirable. But current transhumanism is indeed rather mealy-mouthed about what we are aiming for, perhaps to avoid upsetting anybody. That is probably not a good thing.

The worry about a "slippery slope to anywhere" as worse than a slippery slope to a bad outcome is interesting in itself. There is a pendulum movement between fearing too great order (1984, Brave New World, Colossus: The Forbin Project) and fearing anything-goes chaos (Our Posthuman Future, Enough). The first was a very modernist fear, this might be the postmodern variant. Maybe it is the result of the realisation that one can never achieve total order, and people instead fear that chaos will overwhelm us. In any case, the freedom pathos and open-ended dynamism of transhumanism does nothing to reassure people that we are not headed for some kind of posthuman singularity mess.

One of his criticisms of transhumanism was its attempts to normalise its aspirations through sports. If people enhance themselves (or are allowed to do it) in sports, we can claim "look, it is acceptable, now let's do it in the rest of society!". I'm not entirely sure this is true, since it seems easier to get enhancements accepted outside sports than inside. He made the point that posthuman sports might be rather pointless to watch, especially if the audience had no chance to themselves become posthuman or even identify with the posthumans. Of course, one might just shrug and ask why that would matter? In that case humans would want to watch human sports instead. Given that people also like watching monster truck racing, it seems at least some find looking at non-human sports interesting.

The worry that humans would have a less moral standing than posthumans came up. While I can see a sociological possibility, it is not clear to me that there is any philosophical problem here. As I see it, moral standing comes from being an entity that is able to act morally, something which requires abilities such as memory, rational thinking, control over one's actions etc. (whether one should include consciousness, the ability to suffer etc is more unclear to me) But once a certain level of such abilities are reached, further enhancement of these abilities does not confer a higher moral status to the entity. A smart person is not judged as inherently morally superior to a stupid one because he can predict consequences of his actions better. I would call this a kind of "moral Church-Turing thesis": there exist a threshold level of ability that allows full moral subjecthood, and all entities above this threshold could in principle reach the same moral conclusions and moral behavior (given similar data, values etc). Some might have a hard time doing it, of course.

Julian Savulescu defended enhancement in sports, suggesting that they were not counter but actually in line with the spirit of sports. Sport is after all a test of ability and talent in a rule guided activity, where we want to see the human spirit developed to its fullest. But isn't the human spirit striving to be better, to live better lives? And for that we always use technology. He recognised safety and the nature of the activity as reasonable limits - no need to allow damaging substances or enhancements that make the game ridiculous.

I think he has a good point here. What sports really is about is autotelic behavior, seeing somebody make the most of one's potential or push beyond it. There is nothing inherently wrong in doing this with technology (especially given the external technological enhancements like fiber glass vaulting poles and shark-skin swimming suits that are already accepted). Achieving something with the help of an enhancement is just as authentic as achieving something lesser without an enhancement: the enhancement allows us to have authentic experiences of greater events that we would otherwise not have access to.

One interesting point was his suggestion to have an independent body test the health of athletes, rather than just look for doping. This fits in with the conflict of interest problems of sports doctors (who are they working for?) and would aim for better outcomes in sport.

Jim Parry talked about supplementation, another grey area. Lots of interesting facts, but few conclusions. Overall, the placebo effect might be a very strong factor in much performance enhancement.

Nick Bostrom discussed status quo bias in bioethics: when a proposed enhancement seems to have negative long-range effects, are these due to real problems or just that we tend to prefer the status quo? He suggested an ingenious thought experiment to subtract the bias. I think his approach is somewhat tricky to apply in many cases, but still a good argument.

Andy Miah discussed the different kinds of possible enhancements. What are the differences between designer steroids, plastic surgery (so far allowed in sports - but what happens when some swimmer gets webbed fingers?) and morphological enhancements like wings? His suggestion is to look for "the wrong kind of risk" - some risks are acceptable in sports, others aren't. This is largely determined by the spirit of sports (and other factors, like liability, I would guess), and could be a basis for drawing lines. For more, see his Bioethics and sport blog.

Kutte Jönsson approached the issue from a radical genus perspective. Given the gender inequalities in sport and how they are institutionalised, he suggested that cyborgization (in Donna Haraway's sense) was needed to liberate women in sports. Very fun - equality through singularity. But while I think the gender dichotomies in the long run are likely to dissolve when we get good morphological freedom, it will be quite a while.

To sum up, this was a very enhancement-friendly discussion. While there are clearly practical problems with allowing enhancements, the philosophical case about whether they are inherently bad is far from settled.

The image on top of the page is a Dihydrotestosterone Receptor, based on the PDB file 1T5Z.

Posted by Anders at 09:00 AM | Comments (1199)

May 26, 2005

Saving Sweden from the Doom that Befell Stockholm

[q-bio/0505044] The effect of travel restrictions on the spread of a highly contagious disease in Sweden

A preprint dealing with how a SARS-like disease might spread in Sweden, and how banning long distance travel could reduce the outbreak.

The outbreak scenario starts with a single infected individual in Stockholm. People move between the categories S (susceptible), Latent (infected but not infectious), I (infectious) and R (recovered and/or immune) with different probabilities, and are assumed to move around based on travel intensity data (from a 2002 study looking at how many people travelled between different municipalities). Three different levels of infectiousness and three levels of travel restrictions were tested - free travel, a ban on trips longer than 50 km and a ban on 20 km trips.

The results were fairly robust: even banning 50 km trips kept the outbreak largely within Svealand (the region around Stockholm) regardless of infectiousness (worst case scenario had tens of thousands infected in Stockholm, the best still a few thousand). Cities far away from the center of infection like Gothenburg and Malmö were well protected.

Is such a ban doable in practice? Legally it seems the government does have the authority to do it in the case of a potentially serious epidemic. Shutting down flights, inter muncipality bus-lines and passenger trains is fairly easy to arrange (after all, they tend to shut down spontanously from time to time :-) Northern Sweden would be hardest hit (long distances between cities), but would also likely avoid infection altogether.

The interesting thing about the study is that it shows that something milder than a full quarantine can limit the spread of an epidemic. This is an approach that does not rely on absolute protection, but just reducing the number of high-throughput links in the disease transmission network. While preventing epidemics might be one of the most legitimate uses of government coercion around, one would prefer to reduce the amount of coercion needed.

It remains to be seen if this approach works in other countries. Sweden has a fairly small population and few metropolitan areas, and intra-area transport is about 2.6 times bigger than inter-area transport (based on the data below). A more evenly populated country with stronger people flows might be harder to handle. But I would bet this is a scale-free result: another country might need another cut-off distance, but by cutting the long-range links first the epidemics is nipped in the bud.

There is still the problem of handling the core epidemic around the initial person; maybe a similar approach but on a smaller scale would work here?

There are some cool graphs in the paper of intra-municipal travel intensities based on data from SIKA. I couldn't find the exact data, but as a consolation I plotted the tonnage of goods transported by road between the counties in the graph at the top of this entry. Color and line width proportional to tonnage, volume of the spheres represent intra-county transports. Note that the translucent arrows allow one to (waguely) see in what direction the goods flow (e.g. Norrbotten sends much less goods to Västerbotten than vice versa).

Originally I did a 2D version using Matlab, as seen to the right. However, I couldn't sleep tonight (honest!) thinking about how bad it looked.

Posted by Anders at 09:54 PM | Comments (513)

May 23, 2005

Long Names for Very Small Objects

Before me lies an Onlifrag Housacculicorn whitcurrcruncotherlik and Nelifrag Tablestabscrach graycylishinotherlik. Thanks to The Collier Classification System for Very Small Objects I finally have names for the little things that accumulate everywhere.

I am reminded by the "The Celestial Emporium of Benevolent Knowledge" classification of animals written about by Borges in "The Analytical Language of John Wilkins":

  1. those that belong to the Emperor,
  2. embalmed ones,
  3. those that are trained,
  4. suckling pigs,
  5. mermaids,
  6. fabulous ones,
  7. stray dogs,
  8. those included in the present classification,
  9. those that tremble as if they were mad,
  10. innumerable ones,
  11. those drawn with a very fine camelhair brush,
  12. others,
  13. those that have just broken a flower vase,
  14. those that from a long way off look like flies.

That essay of deals with a conlang that tries to be its own categorisation system, gently poking fun at the absurdities and arbitrariness of categorisation. It is a nice complement to the criticism of ontologies I blogged about earlier. May doesn't seem to be a very ontological month.

Posted by Anders at 02:37 AM | Comments (475)

May 22, 2005

Visualising the Eurovision

[physics/0505071] How does Europe Make Its Mind Up? Connections, cliques, and compatibility between countries in the Eurovision Song Contest (via Nature)

A nicely lighthearted paper about network analysis of the voting patterns in the Eurovision Song Contest. To nobody's surprise there are nonrandom patterns, like the Scandinavian countries voting for each other and playing tit-for-tat games.

Although I didn't watch the contest, I couldn't resist making a graph out of the 2005 data, shown above. Color of the nodes represents total score (mapped, somewhat gaudily and uselessly onto hue: blue is highest, red lowest - see The End of the Rainbow? Color Schemes for Improved Data Graphics for some color schemes I ought to have used) while points given are shown as the saturation of the blue arrows.

Analysing the contest data seems to be a popular pastime among researchers. Cultural Voting
The Eurovision Song Contest
by Victor Ginsburgh and Abdul Noury looks for vote trading and the effect of linguistic and cultural similarity. The data can be used for a cluster analysis tutorial and visualised in various ways. Overall, the voting cliques seem fairly robust, while the dynamics over time is much more uncertain.

Posted by Anders at 02:49 AM | Comments (1975)

May 21, 2005

Cockroaches Having a Ball

From Slashdot: Garnet Hertz: Control and Communication in the Animal and the Machine. Controlling a robot by using a cockroach moving on top of a trackball, based on light signals.

This kind of research isn't that radical (compare that with running robots with moth antennae, a lamprey controlled robot, remote controlled cockroaches and rats). But it is still charming in a way, especially as a private research project.

What I especially liked was the answer to the question "Is the cockroach in pain?":

Madagascan Hissing Cockroaches make a loud "alert hiss" when they are angry. They also enjoy feel safe when crammed into a tight space. Their cuticle has no nerve endings in it. Because of these reasons, and because they do not illustrate a fearful hiss when controlling the robot, it is my opinion that they are in no pain, and do not mind being in the robot system.

This is the first time I have heard of an invertebrate experiment where one can actually tell what the animal "thinks" about the experiment. Maybe one should actually aim to use more species that are expressive enough to signal like or dislike in order to construct more ethically solid experiments?

Another good observation on the page is that "The quest for artificial intelligence or artificial life might be more interesting if it was less artificial". I fully agree (and this is IMHO becoming more and more accepted wisdom in the field): most of our experiments have dealt with environments (be they knowledge spaces or simulated worlds) that were far too clean and simple. Much of the wonderous complexity we see in nature comes from interacting with other complexity.

There is of course a reason for keeping experiment worlds clean: without it data from the experiment becomes more unreliable, and some complex responses might not be due to the system being studied itself but rather from the complex environment (however, this is interesting in its own right - quite a bit of our own complex behavior is likely just simple responses driven by complex inputs). One reason sociology, psychology and economics are such a mess is that it is nearly impossible to do unequivocal experiments. However, a simulated world can still be far more regular experiment-wise than the real world where nothing can ever be replayed perfectly.

In the end, it might be interesting to figure out ways of using this project so that the cockroaches can actually achieve their own aims rather than just move the robot around. Imagine cockroaches moving their robots to find shelter, food and partners for themselves, perhaps in ways that go beyond what could be achieved by a normal cockroach, e.g. in climates inhospitable to them. That would be truly "ecological robotics". The cyborg cockroaches would have arrived, and as long as they can't build any robots of their own we would be safe.

Posted by Anders at 10:13 PM | Comments (586)

May 19, 2005

Information Art and the Third Internet Stage

I recently ran into the information aesthetics weblog. What a delight! It links to just the kind of information-as-art projects I enjoy, with many intriguing ideas.

After reading Ontology is Overrated: Categories, Links, and Tags and being presented to I was really primed for this. Because many of the most impressive projects are based on this tagging idea: there is lots of unstructured data around, and it can be visualised and made useful/beautiful in new ways. Information visualisation of the traditional ontologies and databases remain useful but clearly limited.

We have reached a new situation on the Internet, the "third stage" (I better invent a cooler buzzword here, so that I can spend my time lecturing about it :-)

The first stage was the development of the physical infrastructure, basic protocols and software services such as the WWW. It was about building the underpinnings of the net and exploiting them directly. Some meta-services emerged like mailing lists, paving the way for the second stage.

The second stage occured when more people got online and began to set up businesses