Quantum computing

Visage · 23 Feb 2007 at 10:44

titchard said:
well you say that - but we are able to create and program little robots with the same reactions as small insects - surely that's the first step!

you forget that we have had billions of years of evolution - an incomprehendable ammount of time, yet a computer has less than 100 years. But if you compare the two in graphs the speed of evolution is huge on the computer than us.

They don't have conscious yet, but we do have software that can learn (even in slow and small simple ammounts) and it is working that way.

Rich

I suspect that before we can develop computers that are conscious that we would first need to accurately define what we mean by 'conscious'....

Psyk · 23 Feb 2007 at 10:47

titchard said:
well you say that - but we are able to create and program little robots with the same reactions as small insects - surely that's the first step!

you forget that we have had billions of years of evolution - an incomprehendable ammount of time, yet a computer has less than 100 years. But if you compare the two in graphs the speed of evolution is huge on the computer than us.

They don't have conscious yet, but we do have software that can learn (even in slow and small simple ammounts) and it is working that way.

Rich

Read my post a bit further up. Computer software does evolve but it needs humans to design it. Our limited knowledge of what it means to be conscious and the algorithms needed to simulate it is a bigger limiting factor than raw processing power.

titchard · 23 Feb 2007 at 10:52

Psyk said:
Read my post a bit further up. Computer software does evolve but it needs humans to design it. Our limited knowledge of what it means to be conscious and the algorithms needed to simulate it is a bigger limiting factor than raw processing power.

I understand that - but I am saying that the rate we are learning at how to improve upon our limited knowledge of what consciousness is and how we work on a base level is allowing us to recreate this in a computer-sense much faster than we have evolved. I know that humans came by cumulative selection and that computers are actually generated by us as a product of our learning, but it is feasible to create a computerised product that could even have one tenth of the concsious and rational thinking of us, just takes time.

Rich

dustiestrat · 23 Feb 2007 at 11:02

Psyk said:
Yes I believe there has been research into this area. They can "teleport"* single particles. However the problem with teleporting an large object is really that it would have so many billions of particles there's currently no way to map them all so they can be reconstructed on the other side.

*by teleport I actually mean transfer the exact state of one particle to another, the original would still remain. So in effect the two would be identical.

okay then cheers for clearing that up so basically it would be a cloning device rather than teleportation to an extent?

Johninho · 23 Feb 2007 at 11:13

Inquisitor said:
http://en.wikipedia.org/wiki/Quantum_computing

Interesting Read

Saberu · 23 Feb 2007 at 13:43

Psyk said:
Without a program written by a human it won't "think" at all. So no.

True but a huge processing capacity is required for AI to work effectively in the first place so it's a step albeit not the most important one which as you said would be the programming itself. I envision a program which evolves it's own intelligence through various learning methods.

I remember reading about a guy who made a program which was able to evolve using random mutations of the program then keeping the best one then getting another load from that as another generation and so on. I think he used it for an optimal design of lense for short or long sighted vision.

titchard · 23 Feb 2007 at 13:49

Saberu said:
True but a huge processing capacity is required for AI to work effectively in the first place so it's a step albeit not the most important one which as you said would be the programming itself. I envision a program which evolves it's own intelligence through various learning methods.

I remember reading about a guy who made a program which was able to evolve using random mutations of the program then keeping the best one then getting another load from that as another generation and so on. I think he used it for an optimal design of lense for short or long sighted vision.

A program of that nature (on a simplifed level) is mentioned in the book "The Blind Watchmaker" by Richard Dawkins. Interesting read - revolves around some of the subjects we are talking about at the moment.

Rich

D.P. · 23 Feb 2007 at 13:56

titchard said:
A program of that nature (on a simplifed level) is mentioned in the book "The Blind Watchmaker" by Richard Dawkins. Interesting read - revolves around some of the subjects we are talking about at the moment.

Rich

I work with these algorithms on a daily basis.
A friend in the lab I work in recently had this paper published:
https://imapwww.epfl.ch/web/util/go...009281&Horde=a516f4415426de0cc7c7415c5ff4e6aa

Evolution of robotic communication that was accepted in a high impact factor biology journal.

Kreeeee · 23 Feb 2007 at 13:57

The grammar in this thread is shockingly poor!

n3crius · 23 Feb 2007 at 14:08

titchard said:
but isn't that all essentially we do - on a base level? we just evolved to a point of being able to learn - just like computers will be able to do one day.

Rich

Correct. We understand very little about our own consciousness. But the reality is our 'minds' are based on logic. So what if your gray matter doesn't look like the silicon inside your pc? We are machines, and we're disillusioned quite a bit on just how aware we are, and how much control we possess over our own decisions. Our awareness sits on levels of complex logical processes.

It's also hard to understand how consciousness arises, if it exists on different levels. I've read stuff before about how a stone might be conscious on a really low level, and that every time you turn a computer off, you effectively kill a conscious system (albeit one on a level much lower than our own.) Of course, that's all speculative but can't be discounted. Food for thought maybe.

There's no doubt that it should be possible to construct artificial minds, with just as much awareness and intelligence as ourselves if not greater greater than our own.

The difference is the traditional computer with current hardware, and our programming will never 'learn' by itself. An information system that can learn from us is more akin to the 'web'. In a way, that is an entity that we interact with on a massive scale, it is growing, and really could be losing control over it every day. Who knows what it might evolve into at some stage.

More traditionally, you might consider mapping the logic of the human mind in software, or indeed in hardware in the future. Quantum computing will probably help with these processes a lot.

M0KUJ1N · 23 Feb 2007 at 14:11

Visage said:
But while quantum computing may render traditional crypto useless, quantum technology will also enable truly uncrackable encryption.

Not encryption thats really hard to break, but encryption that is impossible to break.

Yes, as to attempt to intercept the message you will have to observe it. However merely by observing it you force it from being in a superposition of states to being in on or other states, thus causing the recipient and the sender to both be informed instantly and a new key broadcast.

Also the "paired" quanta are generally photons, not electrons and the effect is called quantum entanglement. Entangled photons are basically 2 photons which retain the same quantum state irrespective of distance. It also means that if the state of one of them changes the state of the other will change instantly too!

Thats my (rusty and probably out-of-date) take on these points, hopefully someone with more up to date knowledge can confirm this or tell me off for talking out of my bottom

Mat · 23 Feb 2007 at 14:20

Whos to say that if a suitably "large" quantum computer is built and a suitably complex algorith is written for it, the computer wont go away and start to learn at a geometric rate, become self-aware at 2:14 a.m. Eastern time, August 29th and in a panic, they try to pull the plug?

But seriously, if the computer is capable of such massive numbers of calculations, has the necessary storage and algorithms, we have no idea if it will be able learn at a rate capable of producing AI. someone might hit the 'enter' key and 2 days later, *poof*... "Hello Dave..."

(3 movies in one post

)

Inquisitor · 23 Feb 2007 at 14:37

Mat said:
Whos to say that if a suitably "large" quantum computer is built and a suitably complex algorith is written for it, the computer wont go away and start to learn at a geometric rate, become self-aware at 2:14 a.m. Eastern time, August 29th and in a panic, they try to pull the plug?

But seriously, if the computer is capable of such massive numbers of calculations, has the necessary storage and algorithms, we have no idea if it will be able learn at a rate capable of producing AI. someone might hit the 'enter' key and 2 days later, *poof*... "Hello Dave..."

(3 movies in one post )

Which was the 3rd? I got 2001 and Terminator 2

Energize · 23 Feb 2007 at 15:31

Visage said:
Not encryption thats really hard to break, but encryption that is impossible to break.

There are already types of encryption that are impossible to break using brute force.

Inquisitor · 23 Feb 2007 at 15:33

Energize said:
There are already types of encryption that are impossible to break using brute force.

Theoretically, but as always, the weakness lies in the implementation.

Visage · 23 Feb 2007 at 15:36

Energize said:
There are already types of encryption that are impossible to break using brute force.

Such as?

Inquisitor · 23 Feb 2007 at 15:51

Visage said:
Such as?

One-time pad.

Of course, it's not the most practical. It also relies upon the key being truly random and kept secret.

Visage · 23 Feb 2007 at 15:57

Inquisitor said:
One-time pad.

Of course, it's not the most practical. It also relies upon the key being truly random and kept secret.

To be truly unbreakable it also requires the pad length to be at leat as long as the cumulative length of all messages you wish to send.

So you've replaced the problem of sending an encrypted message with one of sending a pad, which is at least as long.

So yes, technically its an unbreakable cypher, but in practical terms its unusable.

D.P. · 23 Feb 2007 at 17:29

Visage said:
To be truly unbreakable it also requires the pad length to be at leat as long as the cumulative length of all messages you wish to send.

So you've replaced the problem of sending an encrypted message with one of sending a pad, which is at least as long.

So yes, technically its an unbreakable cypher, but in practical terms its unusable.

Indeed, it is a recursive problem, instead of sending an encrypted message you are sending an encrypted keypad,except the keypad is very big to cover all possible messages you will send.

+ the keypad has to be truely random which it may not be.