Yesterday I went to my place of work and, from 0930-1430, did not do a single billable hour for The Institute. I was working instead for the Future of Ireland as a Technological Nation FITNa. The Institute was hosting another round of Young Scientist projects from the five surrounding counties. It was an elimination round, with the winners going forward to a national event and the winner of that going to America to compete with the world. I've done this before in 2014 and 2013, and the science is not really getting better, not worse either though. Partly this must be because we are not giving, as far as I can tell, any cross-the-board feedback. It's harsh and brutal but necessary to tell kids that they could do better. It serves nobody if you say that a project is very good when it is clearly underpowered, improperly analysed and the experimental procedure is poorly designed. These traits are so wide-spread that it would probably be helpful to write a generic document for How to Win Prizes at Science Fairs. This year things were quite a bit better in that there were fewer experiments where the kids surveyed the opinion of a bunch of their pals about something rather than getting down in the long grass and measuring the thing itself. It's possible to reveal some useful things by a focused quizzing of your classmates - but you have to have an interesting question and you need a lot of questionnaires if it is to mean anything.
Project #1 was like that, but they were asking their pals about the thing they were measuring: how do we perceive colour. Here the science is in the perception, so the only way you can address it is asking epopel what they perceive. What the lads from Avondale Community College had done was print out 10 3x3cm blocks of red and asked their samplees to put them in order of brightness. Their take off is the work of Prof Jay Neitz in Wash U. in the US, which I'm going to have to follow up on later. At first glance they all looked the same but when pushed I could say this was the darkest and this next and so on. They carried out this procedure for pretty much the whole school and then plotted how far people were off the true order. One guy did the task perfectly, and everyone else did less well. But they hadn't asked Visionboy to do the same test a week later to determine if the whole thing was a fluke. It's like me asking the Masters of Immunology to put the following in order of size. If you ask enough people to do this, someone is bound to get the right answer. Heck a bunch of monkeys pointing with a stick would do as well [i.e. badly]. Clearly that idea has a way to go, but I liked their project very much because I could engage with their spokesman in a serious scientific conversation. It's really difficult to work in isolation, so getting their ideas out there can only improve their science. The dialogue helps you articulate your ideas. The brightest and most engaged young man I've encountered this century went to a regular Irish school and wanted to learn. He had to endure a hail of spit-balls and small stones striking the back of his head pretty much continuously as he sat at the front of class doing schoolwork. But for the fact that he was a dedicated cross-country runner who won prizes at local meets, he would have had his scholarly aspirations and his head flushed down the school toilets by the know-nothings.
But in Avondale, which on the face of it is another regular Irish school, science gets done. I was equally impressed by two lads from the same school who had done an ergonomic study to determine the most efficient way to wrastle a sheep into submission so that you can shear it or doctor its feet. The uncle of one of the pair had sustained a back injury and a broken arm when he was thrown over by a particularly feisty ewe built like a block of flats. There's got to be a better way, they figured: not wanting to get whacked themselves probably, th'wimps. Their deliverable is an explanatory leaflet which will help farmers do this ever-recurring task right rather than the way it's always been done. Everyone wins here: fewer trips to A&E, less stressed ewes, faster processing.
Next to them were a trio of girls who are obsessed by loom-bands which are going the rounds at the moment as a handy way of parting younger teens from their money in exchange for a bunch of colourful rubber bands [Advertisement: http://www.loombands.ie/ ] [critique]. But these girls, from Avondale almost needless to say, heard that a key rubberising ingredient in the product is the carcinogen phthalate. Their study involved sampling loom-bands which were EC certified and those, cheaper, which were not. Remarkably (score 10 for enterprise in their teacher) they had all gone down to WIT and ran their samples through a Gas chromatography–mass spectrometry GCMS instrument to measure the concentration of phthalate in certified and rip-off loom-bands!
Next to them, we were back in the world of agriculture. Young chap from farming stock had recently had half the family herd condemned as positive for tuberculosis. It turned out that, these were 'false-positives' but under the rules of the game a false positive is still condemned. Chap reckons that vets habitually gets the spacing of the two injections wrong and this increases the chance that a false positive will result. You can see how easy this might be as you have to make the injections separately into the neck of a cow tossing its head about in the cattle-crush. Accordingly, he's gone to the shop and made a Steampunk machine from a re-purposed silicon-gun and some sheet metal that does both the necessary injections precisely 7cm apart which is the optimum distance. An unintended benefit of this is that the vet can do each cow with one action rather than two, save a couple of minutes per cow and get back in his jeep quicker for the same amount of money. The prototype weighs a kilo and doesn't fit in a normal pocket but that's what prototypes are about. He has field tested it in a blind experiment in a different county and gotten positive feedback. He's also teamed up with a pal who is a whizz at the graphic design and publicity, so they have the CSO and the VP Sales and Marketing, all they need now is some venture capital. I am in awe. These lads go to school in . . . Avondale.
But it's not all Avondale. The best geek-maths talk I has all day was with a young chap who has proved that the 2048 puzzle that was going viral last year in "NP complete". NP-complete is a class of mathematical problems that can be verified easily - in finite time with a suitable computer - but cannot be easily solved. It is a subset of NP-hard problems. NP stands for non-deterministic polynomial time and I couldn't explain it better than wikipedia, so I won't try. If you haven't seen the 2048 puzzle yet, don't go there, that way madness lies. The niftiest aspect of this project, is that it was solved in exactly the same way as Andrew Wiles solved Fermat's Last Theorem. He showed that his representation of the algorithm which solves the 2048 puzzle could be mapped onto another problem that was known to be NP-complete; that is a sufficient proof in modern mathematics. This fellow doesn't go to school in Avondale, it must be lonely in St.Mary's CBS in Portlaoise, but not too lonely because our jury short-listed at least one other prize from that school.
FITNa? Looking up!
Post a Comment