* Posts by Fenwick

22 publicly visible posts • joined 9 Dec 2008

Sorry, Dave, I can't code that: AI's prejudice problem

Fenwick

Re: Transparency...

How a neural network arrived at an outcome (the algorithm, its parameters and its input) can easily be recorded. Understanding why the algorithm gave a particular output is much harder.

1. Data is used, with another algorithm, to set the parameters of the neural network. There are lots of parameters and they all depend upon each other and the data in complicated ways. Changing one parameter a little bit, may, or may not, have a big impact upon all of the other parameters. This algorithm takes a computer ages to run. It is a huge, perhaps impossibly huge for a human, job to find out and understand why the parameters are set the way they are.

2. Your input is fed to the neural network. It proceeds through the network with numbers growing and shrinking in complicated ways according to the parameters. At the end an output is generated. If you are really careful, you might be able to say that certain parameters fixed the decision and others weren't so important. With more effort you might be able to find out if the decision was robust to small changes in the parameters or your input. But what I think you really want is to know why the parameters are set the way that they are. In a complicated problem, nobody knows.

Stop using USB sticks to move kids' data, auditor tells Education Dept

Fenwick

Real world

"Analyse this data or you are sacked, it is your job."

"No you can't have network access, you can have access to the computer for 10 mins."

"Don't tell me how it is done, just make it happen."

So who exactly will you fine with your easy fix?

Why solid-state disks are winning the argument

Fenwick

"it is a really dumb idea to take the cheapest desktop hard drives you can find"

Loads of people say this, and often have anecdotes about a life being destroyed, business failing, etc.

But the only actual evidence I have managed to find says the opposite (Backblaze, Google, https://www.cs.cmu.edu/~bianca/fast07.pdf).

Of course no study is perfect. I'm sure that many people think that all of this "research" is wrong. So can anyone please come up with a logical or evidence based counter argument for why "it is a really dumb idea to take the cheapest desktop hard drives you can find".

Saucy selfie app Snapchat hits $800m valuation as VCs chuck cash around

Fenwick
Unhappy

Re: What an utter load of bollocks.

Those idiot investors are probably our highly paid pension fund managers, (when it floats).

Whiptail whips out SME-friendly flash array

Fenwick

Re: what kind of IOPS are those?

I guess you answered my questions. You pay for software and support. I should have asked: What is the difference in the hardware?

Yes, I know that on the SI quality scale the box I mentioned was "commodity crap", I know that it is set up in a crazy way, with a reliability rating of "unholy". Looking at their website, I don't know that the whiptail box isn't just the same hardware with an "enterprise" badge stuck on the front.

Is it? If not what is it? Why is the whiptail hardware better than the commodity server? Why is the ACCELA hardware better than 10 (for redundancy) commodity servers? (or if not whiptail some other similarly expensive SSD based device) Is the entire price difference due to the software, support and high margins?

And finally yes you're probably right. Being new at this stuff, I shouldn't ask questions because I might expose how ignorant I am.

Fenwick
WTF?

Why not get one of these instead?

http://www.tomshardware.co.uk/ssd-dc-s3700-raid-0-benchmarks,review-32669-6.html

What do the high end boxes in the article have on this cheap comodity box?

Are people being tricked out of their money or not?

Biz prof disses Big Data as a fetish for info hoarders

Fenwick

Re: @Fenwick

I reckon that I could make more accurate predictions of the average August temperature using a larger data set. Or at least quantify my accuracy with a little more certainty.

Talking of trends. If you want to convincingly separate (and quantify) any long term trend in say the temperature, from things like the annual cycle, turbulence and sea surface temperatures, then 200 years of 2000 samples per hour (at 1x1 mm resolution around the whole world) would do nicely.

For some applications the volume of the sample is exactly the issue (for example in accurately estimating the failure rate of bearings).

Fenwick

Re: @Fenwick

In the UK I have seen that August is warm and February is cold (substitute for correlation). The weather is chaotic and unpredictable to some extent.

I predict that this August will be warmer than next February.

What odds would you require to bet against me?

Fenwick
Thumb Down

Newton's law of universal gravitation is quite a good model based upon "fitting a pattern to some data".

You can make accurate predictions. You don't have to understand why it works. It has stood the test of time.

How do you expect to come up with a theory (model) if you are not allowed to use your observations (data)? Close your eyes and guess?

"[Scientists] simply look at loads and loads of data, fit some patterns to it, and then start making decisions based on that model."

Yup. That is what scientists do. If you come up with a crappy model, perhaps your predictions will also be crappy. There is nothing wrong with the overall approach though.

Cycle fires up 50,000-core HPC cluster on Amazon EC2

Fenwick
WTF?

Re: obesity?

How do you prescribe that?

Met Office wants better supercomputer to predict extreme weather

Fenwick
Thumb Down

Re: You want ANOTHER supercomputer?

If you know how exactly to write a better program than the one we already have. I.e. a program that makes more accurate predictions for less computing power. Perhaps you should let the world know about your world beating understanding of atmospheric physics and numerical algorithms.

Or perhaps you meant that you don't understand the point of knowing the weather in advance.

Beware Freedom of Info law 'privacy folktale' - ICO chief

Fenwick

So are you saying that the problem is not FOI it's people with an agenda twisting someone's words? If so I suggest we solve that problem instead.

Since I start with the assumption that anything I do at work (for the public sector) will at some point in the not too distant future be made public, I find myself trying just a bit harder to do the right thing. Also I spend less time doing things that I would not like to be made public.

‘Want to be more secure? Don’t be stupid’ redux

Fenwick
Unhappy

I think I fall under your definition of stupid, sack me?

Our departments computer admin was stopped because there was no budget (public sector). My boss told me that I had to do it on top of the job that I am paid to do.

I know nothing about computer security or administration.

Many of our applications that we all use are out of date, i.e. a later version costs money and therefore we can't have it. I have a budget of zero for the next year at least.

I don't have admin rights or access to about half of the computers.

Users typically purchase their own software with their own budget and must be allowed to install it for network use. I.e. on a server. I don't know who half the users are and am not allowed to find out because of privacy bureaucracy bullshit. (I'm sure I could find out if I broke the conditions of my employment)

I do not have sufficient status to be allowed to attend the meeting where these policies are decided. Making a fuss is the sure way to prove that you are not capable of making do and to be promoted from the voluntary to the compulsory redundancy list.

Since I am "responsible" for computer security, perhaps when our site gets hacked and we loose a load of data you should sack me because it is clearly my stupidity that caused it.

Solutions?

HP parks Airbus supers in containers

Fenwick
Stop

1600 km apart

Did you mean that each pod has 24192 cores. If so then the top 500 benchmark was done with only one pod and Airbus has two computers that would come in at 29th in the list.

El Reg to unleash rocket-powered spaceplane

Fenwick
Mushroom

Picky Vogon says...

Is a model plane with an auto pilot called a "cruise missile" in some circles?

I thought (probably wrong) that you need a license to develop cruise missiles.

Give the Met Office £10m, says Transport Committee

Fenwick
FAIL

Weather action! Are you mad?

How do you know that Weather action have any more or less skill than the met office.

Oh, you don't.

Fukushima one week on: Situation 'stable', says IAEA

Fenwick
Megaphone

Good information

http://www.bousai.ne.jp/vis/

With the application of google translate.

Bishop Hill: Gonzo science and the Hockey Stick

Fenwick
Unhappy

Appologies

For my over the top comment. I guess anger is what leads to things like Godwin's law.

Fenwick
FAIL

Teleconnections

I got as far as the comment on teleconnections and realized that Andrew is writing from a position of total ignorance. How you have the balls to suggest that someone else is talking shite when you don't have any grasp of the subject yourself is beyond me.

In this context...

For mathematicians: The first teleconnection pattern is simply the first eigen vector of the covariance matrix.

For statisticians: The first teleconnection pattern is the first Emperical Orthogonal Function.

For everyone else: The first teleconnection pattern is a robust pattern that you can observe in the climate data.

Actual physical interpretation is difficult and sometimes controversial. They have nothing what so ever to do with pseudo scientific bullshit like homeopathy, telepathy or anything else, it's only the unfortunate name that suggests so.

After reading this propoganda, I couldn't be arsed to read the rest of the article so did not get what ever point you were trying to make. Irrespective of how correct you may or may not be.

Leaving PCs on costing UK business millions

Fenwick

Calculations

We do calculations on all of our linux boxes overnight, switching them off would cost us dearly.

I'm a sceptic now, says ex-NASA climate boss

Fenwick
Stop

Just to clarify

It is well established that if you add enough CO2 to the atmosphere, it will eventually heat the surface via the greenhouse effect. Or rather it is difficult to imagine anything else.

The question is how much C02 is required to heat the atmosphere by any particular amount? Precisely how much warming do our emissions cause? Is a runaway greenhouse effect possible on Earth, like it is on Venus?

Models attempt to guess at the answers to these questions. As far as I know there isn't a single climate model that is able to definitively answer them, in fact building one that does is a holy grail of climate science. The best IPCC model guess seems to be a warming of 2 to 4 degrees, with an uncertainty in predicting the true atmosphere ranging from 0 to about 100 degrees.

Firm touts anti-radiation chip for phones

Fenwick
Stop

Hoax

This is a con and does not work. I have seen this several of these sort of hoaxes before. One was even advertised by the BBC news website, which they took down after finding out it was a pack of lies.

Mobile phones emit microwaves that can be stopped by a thin sheet of metal. Placing a metal shield between you and the phone was thought to work until researchers found that it didn't unless the shield was so big as to become unfashionable/impractical. Wrapping the phone in foil should do it, but then you won't get a signal.