Neural Network

[moodz]

Hi Qiaozhi,

nice weather detection example. Brilliant. The NN did work. I like it.
This is a good example of one of the pitfalls in NNs.

Is it worth at all?
Metal detection: No, metal detection is trivial and a well-known science. You know, how to process the data.

But there are many many non-trivial problems that might be challenged. If you don't know, how to solve a problem. If you don't know, what parameters do affect the problem. If you don't know, whether a problem can be solved or not. In this case, the NN is a very very useful data mining tool.

Nevertheless, it causes a lot of work. And it isn't justifying the lots of effort in many cases.

Cheers,
Aziz

don't discount neural nets so easily ....after all what is that came up with all metal detector technologies ... A neural net (your brain LOL ). See link for clues on neural net DSP .....

http://www.asel.udel.edu/icslp/cdrom/vol1/289/a289.pdf

[Aziz]

I suppose that's your prerogative.

Neural networks (NN) could theoretically be used in metal detectors, but whether the added complication would be a step-function increase in performance is open for debate. Dave Johnson obviously considers the idea to be of little practical value, and I tend to agree. I have some small experience using back-propagation neural nets, and I can see that the concept is do-able, but only of academic interest. Neural networks are very useful for finding patterns in imprecise data, which you could argue is the case for buried metal targets, as there is some correlation between the signals received by the metal detector and the actual target, and a trained network can also make decisions on data it has never seen before by using the knowledge gained during training.

However, you do have to be careful when training a neural network. For example, there is a story concerning the U.S. Military, where they attempted to use an NN to identify tanks in the battlefield as either friend or foe. To do this they input photos of American and Russian tanks, and used these to train the network. It was 100% successful at identifying the targets when tested on the training data, but in practice it was completely useless when used on any photos outside of the training set. After some analysis, it was noticed that all the American tanks had been photographed on a sunny day, and the Russian tanks had been photographed on a cloudy day. The NN was simply identifying the different weather conditions.

Neural nets can also provide an indication of whether there is any correlation to be found within a dataset. A good example is to take a database of recent Lottery numbers and use these as training data. The first thing you discover is that the network is unable to converge, especially if the dataset is quite large. The hypersurface does not contain a true global minimum, and even if you stop the training process once the network shows no further improvement in the network error, the predicted result for next week's Lottery will not make you a rich man. This is because the Lottery results bear no correlation to the previous results, despite what many so-called Lottery prediction programs would have you believe.

So, no doubt you would find some correlation in the metal detector data, but is it worth the effort?

Hi Qiaozhi,

nice weather detection example. Brilliant. The NN did work. I like it.
This is a good example of one of the pitfalls in NNs.

Is it worth at all?
Metal detection: No, metal detection is trivial and a well-known science. You know, how to process the data.

But there are many many non-trivial problems that might be challenged. If you don't know, how to solve a problem. If you don't know, what parameters do affect the problem. If you don't know, whether a problem can be solved or not. In this case, the NN is a very very useful data mining tool.

Nevertheless, it causes a lot of work. And it isn't justifying the lots of effort in many cases.

Cheers,
Aziz

[Qiaozhi]

Yes, there is a reason.
How about fairness I asked recently?

Aziz

I suppose that's your prerogative.

Neural networks (NN) could theoretically be used in metal detectors, but whether the added complication would be a step-function increase in performance is open for debate. Dave Johnson obviously considers the idea to be of little practical value, and I tend to agree. I have some small experience using back-propagation neural nets, and I can see that the concept is do-able, but only of academic interest. Neural networks are very useful for finding patterns in imprecise data, which you could argue is the case for buried metal targets, as there is some correlation between the signals received by the metal detector and the actual target, and a trained network can also make decisions on data it has never seen before by using the knowledge gained during training.

However, you do have to be careful when training a neural network. For example, there is a story concerning the U.S. Military, where they attempted to use an NN to identify tanks in the battlefield as either friend or foe. To do this they input photos of American and Russian tanks, and used these to train the network. It was 100% successful at identifying the targets when tested on the training data, but in practice it was completely useless when used on any photos outside of the training set. After some analysis, it was noticed that all the American tanks had been photographed on a sunny day, and the Russian tanks had been photographed on a cloudy day. The NN was simply identifying the different weather conditions.

Neural nets can also provide an indication of whether there is any correlation to be found within a dataset. A good example is to take a database of recent Lottery numbers and use these as training data. The first thing you discover is that the network is unable to converge, especially if the dataset is quite large. The hypersurface does not contain a true global minimum, and even if you stop the training process once the network shows no further improvement in the network error, the predicted result for next week's Lottery will not make you a rich man. This is because the Lottery results bear no correlation to the previous results, despite what many so-called Lottery prediction programs would have you believe.

So, no doubt you would find some correlation in the metal detector data, but is it worth the effort?

[Aziz]

Is there any reason why you don't post it here?

Yes, there is a reason.
How about fairness I asked recently?

Aziz

[Qiaozhi]

Hi all,

if you are interested in the implementation of the NN application (Backpropagation type NN), see in the other forum for the fully suplied source code and binaries (C/C++ Win32 application).
(It's Christmas)

(You have to be a member there to download it I think. Guests can not see pictures and download attachments I think. Search for the "NEURAL NETWORKS" topic.)

Aziz

Is there any reason why you don't post it here?

[Aziz]

Hi all,

there are very sceptic subjects around the forums. And evil subjects, which are ridiculing the NNs and myself (PJ ).
(Hey!, if you aren't an expert, just keep quiet and learn from the majik master. *LOL*)

I've posted an application of a NN on the Australian Electronic Gold Prospecting forum. You must be a member there to see the picture I have posted recently.

Cheers,
Aziz

Hi all,

if you are interested in the implementation of the NN application (Backpropagation type NN), see in the other forum for the fully suplied source code and binaries (C/C++ Win32 application).
(It's Christmas)

(You have to be a member there to download it I think. Guests can not see pictures and download attachments I think. Search for the "NEURAL NETWORKS" topic.)

Aziz

[Aziz]

Hi all,

there are very sceptic subjects around the forums. And evil subjects, which are ridiculing the NNs and myself (PJ ).
(Hey!, if you aren't an expert, just keep quiet and learn from the majik master. *LOL*)

I've posted an application of a NN on the Australian Electronic Gold Prospecting forum. You must be a member there to see the picture I have posted recently.

Cheers,
Aziz

[Davor]

There is another point, a proper interface that would enable our own brains to perform neural networks' tasks more successfully. There is actually only so much you can do, as cognitive capacity of our senses is limited.

[Aziz]

Aziz, most of the time I can't make any sense of the things you say here, but that statement (in a field where you seem to have some actual expertise) really hit the nail on the head. You must have laughed for at least 10 minutes after making that post.

I don't often say much about what we do or don't have under development in El Paso USA, but rest assurred we are NOT screwing around with neural networks!

--Dave J.

Hi Dave,

there are many pitfalls one can fall into by using NNs. Precautions must be made and it requires fully understanding to avoid the pitfalls. In principle, it is even very simple to use NNs. One don't have to understand the physics behind the non-trivial problem.

One of the biggest advantages of the NN's is, that one can analyse the quality of the input parameters and whether they contribute to the solution of the non-trivial problem. If one parameter does not contribute to the solution, the associated network weights convergate to zero. It is a very useful data mining tool however.

Metal detection is very trivial task of course. So we don't need NNs as the physics behind it is well-known and the problem can be solved quite easily.

Cheers,
Aziz

[Dave J.]

I should add that there's room in the world for a bit of everything, and that I highly encourage our competitors to go down the neural network road that I've explicitly stated we left open for other rush hour drivers.

--Dave J.

[Dave J.]

Successful implementation of NN will cause a lot of work however.
Aziz

Aziz, most of the time I can't make any sense of the things you say here, but that statement (in a field where you seem to have some actual expertise) really hit the nail on the head. You must have laughed for at least 10 minutes after making that post.

I don't often say much about what we do or don't have under development in El Paso USA, but rest assurred we are NOT screwing around with neural networks!

--Dave J.

[Aziz]

Hi all,

my own thesis was on neural networks (1992/93) and its application on medical analysis (classifying chromosomes and sperms - but not my own ). It worked fine.

There are dozens/hundreds of sources (source codes, binaries, tools) about neural networks (NN). Thousands of interesting publications and its applications. Just have a look if you are interested in the matter.

Successful implementation of NN will cause a lot of work however.
Aziz

[sinclairuser]

VLF or IB may give more usefull infomation than PI(as in what goes to the neural net), you could shift to that format later, if the proof of concept experiment is valid

[johnl]

I saw article for mine detection just curious if it could transferred to a PI project I am just amateur it is over my head
http://ieeexplore.ieee.org/xpl/login...mber%3D5445767

[Aziz]

Hi all,

there is a nice thesis on "Metal Detection Using Neural Networks":

http://www.academia.edu/1468268/Metal_Detection_Using_Neural_Networks

(alternate: http://docsfiles.com/view.php?view=h...20sussex&count= )

(pdf: http://base2designs.com/THESIS2.pdf )

Aziz