Re: So, to sum up. . .
And Bloody Stupid Johnson's move to stop all debate is OK by you still?
8318 publicly visible posts • joined 11 Jun 2009
If you can spare a couple of hours a week you can pop onto futurelearn and do some free courses on Weka and Moa which can give you an reasonable insight into Data Mining and AI which are quite enlightening. It's an art becoming a science and, as illustrated here, coming up with some really useful results which also makes you ask more questions. What more could the inquiring mind want?
You can set up your convolutional networks to give you a 'view' of what the various layers are getting up to in a variety of forms. However you may not get why it makes the decisions it does even if it does tell you in great detail.
Understanding the reasons it makes the decisions will be the next necessary step in AI and should provide us with an interesting tool-box of things we can use to understand how we work,
Break your petrol tank and the fuel runs under the car and warms your cockles. Hydrogen tends to rapidly go upwards once released from its container. It does burn well, but that goes upwards even more. You will get a flash burn from h2 burning but petrol will keep going.
But that was partly due to bad design. Had the cells had sufficient gap between and ventilation them then the hydrogen would have floated away rather than accumulating to the point it could do some damage. Now with modern detectors and cells not made from bull intestines it would be hard to make something dangerous without intending to.
My granddad was an engineer on the R33 and it they discussed all sorts of improvements and modifications that I wish I'd written down to keep. Its criminal we dont have Hydrogen filled air transport 0 you could fly the Atlantic in a day and do some carbon saving in the process.
Having said that horse poo is a very good thing to put on the soil for growing veg. Having had the pleasure of a 20'*20' by 3' deep pile of rotted horse poo I can assure you a family of 5 cant eat the stuff that that will grow fro a couple of years.
No-one can drink that much dandelion coffee even after several tons of rhubarb crumble.!
It doesnt need many. If you can get 1% of your calls to work that's probably around 1 victim a week and that's probably a bitcoin a month or something. I could live on that, the Indian guy running an office of these people is probably going to be having a bit of a party on that.
It may take a while. For some reason we got a family spotify acct and I download music to annoy the kids in the car. We dont get a phone signal here so my phone is often off. If I forget to turn it on and load spotify in the house so it can call home it refuses to play stuff I've downloaded!
All I can say is in the early years of the millennium there was a thing that netscape released and they called it JavaScript2 and it had things like classes and inheritance without prototype bollox and it was easy to write good things. I hope for that day to come again.
I think this is more about 'I know how to do it in Swift, I dont know how to do it in C++' rather than one language being better than the other. Automating reference counting pops up in Boost as does URL handling. I dread to think what the internals of a language look like when things that are better provided by libraries are actually dragged into the language, Or rather I do - I've seen many languages try this over the years and to no benefit.
Having said that I've yet to find a use for Dropbox that doesnt set alarms ringing and would never ever use it in a business unless a gun was pointed at my pay-check.
I did some simulations many years ago and you could get some interesting results in a couple of days of 200MHz Pentiums. I imagine now with GPUs and the like you could get some rather spectacular simulations done, The one thing I thought I learned at the time was you really have to assume a relatively uniform disk at the start or all bets are off.
Anyone know of any current modelling software for this?
and the burps are when accumulated methane geysers past some underground water. That is assuming the 46 tonnes reported is the total release. There are many processes that could generate methane at depth and , while I'd squeal with delight to find life on Mars, Occam says rocks.
We had a machine in the pub toilet for that!
But seriously the number of machines that used to 'go yank' and refuse to function after a power outage until threatened with percussive maintenance because we couldn't find the manual to set it to A4 or, in later years with LCD menus the A4 selection method was modelled on the claw machine at the seaside.
I did help install an office in a 15th century half timbered jobie. Grade 1 listed so one had to dismantle the equipment to get it up the stairs and through doors I'd have trouble getting through these days.
The printer was about to be returned to sender when someone spotted a hoist sticking out of the eaves on the 2nd floor, The printer was hoisted up to the window there and then returned to the ground where another couple of ropes were attached so it could be tipped on its side to get it through the window. Once we'd installed everything and got it working we wandered off for a pint and I realised that half the buildings on the old streets in the town I'd lived in for 8 years had hoists sticking out.
So they knew their shit 600 years ago - I doubt a modern architect would like something useful upsetting their clean corrugated lines.
And as a note on 2+2=5's comment. You dont use wheely chairs on 600 yr old oak floors. You either find yourself slowly sliding under or away from the desk or flipped on the back of your head when you hit the myriad of kerbs bumps and gaps or rucked carpets working in a place of great beauty offers,
TBF Computer Neural Networks were inspired by the biological ones. They are similar at a very simple level - which may be all that is necessary. They have copied models of Nematode NNs with 'interesting' results. At the moment we are playing with basically layers of nodes feeding layers of nodes whereas nature has more convoluted nets, the functions of which we currently dont know. We have wired up NNs in the convoluted form found in nematodes but we have a way to go to work out which keyboards to plug in and where and how to initialise the net before we can simulate a nematode, I think we are circa 1950 digital computer stage with NNs but we have the equivalent interactions of the IOT to evaluate and experiment with. Oh and fucking contrary organic NNs wandering around kicking over the chess sets to deal with too,
I returned to Uni a year or so after I graduated. I bumped into my old tutor in the bar and the look on his face when he discovered I was already earning more than he was was a sight to behold. Which was a fucking tragedy as he was damn good at teaching. He could lecture to 60 kids 3 times a week and three years later generate more tax in a month than he was paid in a year.
I did quite a bit of semiconductor work which required some silly maths, Mostly above and beyond the three years I did in my Electrical and Electronics degree. I find now researching AI and big data I get to use some of that and more, And, to be fair, its all fucking matrices now. Probably always has been.