{"id":1825,"date":"2013-08-30T14:25:51","date_gmt":"2013-08-30T21:25:51","guid":{"rendered":"http:\/\/www.oliverk.org\/weblog\/?p=1825"},"modified":"2013-09-03T11:34:58","modified_gmt":"2013-09-03T18:34:58","slug":"slime-oracle","status":"publish","type":"post","link":"http:\/\/www.oliverk.com\/blog\/?p=1825","title":{"rendered":"slime oracle"},"content":{"rendered":"
\n
\"Fuligo\"<\/a>

Fuligo septica <\/em>in my back yard<\/p><\/div>\n<\/div>\n

\n
\"slimecu\"<\/a>

Physarum polycephalum<\/em> close up<\/p><\/div>\n<\/div>\n

<\/div>\n

Those of you who\u2019ve followed my blog for a while will know I am fascinated by the intellectual capabilities of slime moulds…
\nThis is the time of year when the dark coniferous forests near my house get brightened up by their white and yellow blobs of sentient goo. Around here, they are mostly the common slime mould Fuligo septica<\/em><\/a> \u2013 otherwise known as\u00a0 \u2018Dog Vomit slime mould,\u2019 a rather harsh\u00a0name,\u00a0 in my opinion, for such an engaging life form.
\nFar from being vomit or mould, these slimy masses are actually swarms of social amoebae that have combined their protoplasm and nuclei into a single super-organism called a plasmodium<\/em>.<\/p>\n

Once formed, the plasmodium creeps along the forest floor until it reaches a tree stump or log, at which point it climbs to a high enough vantage point from where it releases its spores into the slightly more active air. These drift, land, and eventually germinate into another generation tiny amoebas that one day will seek each other out, meld together and then start their complicated social migration process again.<\/p>\n

This is all very fascinating, but the reason slime moulds have been the subject of such intensive research is that their plasmodial masses show an uncanny ability to solve computational problems – which is all the more remarkable given that slime mould\u00a0 has neither brain or a nervous system and its constituent amoebae are about as rudimentary a life form as you can get. But when large numbers of these simple beings are networked into a swarm, the principle of emergence<\/a> takes over and we begin to see a unified intelligence.\u00a0 The rules governing emergence are fairly simple – as long as each individual is aware of and able to exchange basic sensory information with its neighbours, stimuli can be transmitted throughout the network, which can then react as a sentient whole, flowing toward attractants or away from danger.
\nThis simple yet powerful mechanism is behind the flocking of birds, the flow of traffic and the evolution of memes on the internet. It is a non-hierarchical, \u2018ur\u2019-intelligence, an organizing tendency imbued in the very vibrancy of matter itself.<\/p>\n

The lab rat of the slime mould world is the easy-to-raise species Physarum polycephalum<\/em><\/a> and as such it has been the subject of much experimentation.\u00a0 Lured by oatmeal flakes, Physarum<\/em> can outperform robots in solving mazes and it will figure out mathematically complex problems such as how to most efficiently connect an arrangement of arbitrary points, making it astonishingly good at designing highway networks and other mesh-like systems.\u00a0 Though brainless, Physarum<\/em> has a rudimentary spatial memory<\/a>. It recognizes where it has already explored by sensing the slime trails it has left behind, allowing it to focus its foraging energy on covering new ground.<\/p>\n

Inspired by my time hanging around with the bio-geeks at Brooklyn\u2019s GENSPACE<\/a>, I decided to put old Physarum<\/em> to the test by giving it some problems I thought worthy of its impressive powers.<\/p>\n

\"slime<\/a>

slime mould navigating a model of an IKEA<\/p><\/div>\n

1) The IKEA vexation:<\/em>
\nFor years I have been irked by IKEA stores. What is the most efficient way to navigate through one of them? I find IKEA to be especially frustrating and always have had trouble finding my way toward the exit without being overcome by an anxious ‘trapped’ feeling. For me, IKEA is prime habitat for what it called the Gruen transfer<\/a> – a deliberate architectural strategy aimed at disorienting people in shopping malls; an invocation of spatial confusion that makes us lose track of our original intentions and buy stuff we don\u2019t need. I wondered if the slime mould could resist such distractions and efficiently find its way through\u00a0an IKEA without getting sidetracked? To put this to the test, I built a tiny petri-dish sized model of the floor plan of the IKEA, modeled on the store in Richmond, British Columbia, which is more or less typical. Each department was baited with an oatmeal flake and I plunked a little bit of Physarum<\/em> culture down near the store’s entrance to see what it could do.<\/p>\n

\n
\"slime<\/a>

slime mould says ‘yes!’<\/p><\/div>\n<\/div>\n

\n
\"second<\/a>

second choice ‘hazy!’<\/p><\/div>\n<\/div>\n

\n
\"at<\/a>

at the moment of ‘yes or no?’<\/p><\/div>\n<\/div>\n

<\/div>\n

2) The Future is out there:<\/em>
\nDecisions involving the future are hard for me to make, especially if there are a lot of equal seeming options. Throughout history, we\u2019ve consulted soothsayers, fortune tellers and ouija boards in hopes we might get some guidance on how to choose between those many forks we encounter in life\u2019s road.\u00a0 I\u2019ve often despair and then make a seemingly random choice. If slime moulds are so good at solving problems in the here and now, how might they do in sussing out what hasn\u2019t yet happened? Could they help me \u2018Magic 8-Ball\u2019 <\/a>style make decisions on what to do next?\u00a0 Well, I was a little too clumsy to build a proper \u20188-Ball\u2019, but I managed to make a reasonably convincing \u20186-Ball\u2019 simulacrum, by cutting six (approximately identical) radial paths into the agar of a petri dish and having the slime mould choose which one it wanted to go down. I baited\u00a0 each \u2018choice\u2019 with a flake of oatmeal and labelled them: \u2018yes,\u2019 \u2018no,\u2019 \u2018outlook good,\u2019 \u2018reply hazy,\u2019 \u2018somewhat likely,\u2019<\/em> and \u2018don\u2019t count on it!\u2019<\/em> I\u2019d ask the slime mould a (secret) question then wait a few hours to see what it decided.
\nWell those of you who have had scientific training are probably about to bust a blood vessel about now, but bear with me!<\/p>\n

\"ballistic<\/a>

ballistic behaviour<\/p><\/div>\n

Perhaps predictably, the IKEA \u2018maze\u2019 problem was quickly solved by Physarum<\/em>. It relentlessly streamed from one \u2018department\u2019 to the next in search of its oatmeal rewards and within a couple of days, it arrived at the exit hungry for more. This is a very basic slime mould-solvable problem. Physarum<\/em> tends to behave in what is called a \u2018ballistic\u2019 manner; once they have sensed a stimulus they will try to flow toward it taking the most direct route possible.<\/p>\n

Except when they don’t
\nBecause unlike a non-sentient being, slime mould will often abort its migration toward an attractant, if it senses a rival has gotten there first. Sometimes though, competitors will merge into one happy super organism and share the the love and the bounty.<\/p>\n

\n
\"logic<\/a>

inputs A and B seeking oatmeal<\/p><\/div>\n<\/div>\n

\n
\"A<\/a>

A and B combining to one output<\/p><\/div>\n<\/div>\n

<\/div>\n

This makes for some interesting possibilities for \u2018non-digital\u2019 logic. Let’s say we place slime moulds, (let\u2019s call them \u2018A\u2019 and \u2018B\u2019) in each of the two forks of a \u2018Y\u2019 shaped path and put an oatmeal flake in the remaining leg. If the slime moulds were to behave consistently that is to say \u2018digitally\u2019, you would have the basis for what is called an Exclusive \u2018OR\u2019 gate, a common component of computer logic. There has been some great research done on this by Andrew Adamatzky<\/a> at the University of the West of England.
\nGiven the slime moulds\u2019 usual aversion to the presence of a rival, the one that gets to the bait first will dissuade the other. So in logic gate terms the presence of ‘A’ or ‘B’ (the inputs) will generate an output, i.e. either<\/em> ‘A’ or ‘B’ gets to the oatmeal.<\/p>\n

\u2018A\u2019 and \u2018B\u2019 can\u2019t both reach the oatmeal because the presence of one will (most of the time) repel the other, so the gate is considered \u2018exclusive.\u2019<\/em><\/p>\n

But because slime moulds, though simple, are living things, with their own agency, they don\u2019t always do what we expect. When the slime moulds \u2018A\u2019 and \u2018B\u2019\u00a0 break convention and flow together to share the oatmeal, the logic gate becomes \u2018non-exclusive,\u2019<\/em> i.e input \u2018A\u2019 or \u2018B\u2019 or both<\/em> can appear as outputs. Sometimes even that doesn’t happen and a slime mould gets distracted and decides to loiter around the perimeter of the petri dish, boycotting the experiment despite the presence of oatmeal. Who knows why? Are they following their bliss? Or are they just confused? To my mind, these \u2018exceptions\u2019 to the behaviour we deem ‘logical’ are the most interesting results. If we develop a new class of biological computers based on slime mould logic, the outputs generated might have a charming and useful variability. This could be used to more closely emulate the real world \u2018twitchiness\u2019 of such social\u00a0 phenomena\u00a0 as how people behave in the marketplace (they don’t always act rationally or in their self interest), or how they participate in riots or elections, which are also less than rational processes.<\/p>\n

\"this<\/a>

this time- both outputs are generated!<\/p><\/div>\n

And as for predicting my future, the slime mould oracle was initially quite decisive.\u00a0 It answered my first (secret) question with a resounding ‘Yes,’\u00a0 but then started creeping around to investigate other nearby possibilities. To me that seems like a perfectly reasonable approach.<\/p>\n

 <\/p>\n

 <\/p>\n","protected":false},"excerpt":{"rendered":"

Those of you who\u2019ve followed my blog for a while will know I am fascinated by the intellectual capabilities of slime moulds… This is the time of year when the dark coniferous forests near my house get brightened up by their white and yellow blobs of sentient goo. Around here, they are mostly the common […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8,9,91],"tags":[94,93,92,129],"class_list":["post-1825","post","type-post","status-publish","format-standard","hentry","category-environment","category-forests","category-slime-moulds","tag-emergence","tag-physarum","tag-slime-molds","tag-slime-moulds"],"_links":{"self":[{"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/1825","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1825"}],"version-history":[{"count":12,"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/1825\/revisions"}],"predecessor-version":[{"id":1848,"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=\/wp\/v2\/posts\/1825\/revisions\/1848"}],"wp:attachment":[{"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1825"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1825"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.oliverk.com\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1825"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}