Hi there. I am a neuroscientist who is looking into continuing education in neural networks. My info gathering has shown me that algorithms are a large part of ANN. So my question is would it be advised to learn about algorithms first? And if so - what are algorithms I basis of. My highest math understanding is precalulus..... ...so wha tpath should I take - learn about algorithms first and then neural networks? Thanks let me know any help you have!
Apparently no exchange of wanted information actually takes place here. Anyone have anything useful to post?
Okay if you want information. What exactly do you want to achieve? Afterall the basis of Neural networking for the past 10 years has pretty much been theory into how biological cognative systems work and how the theory can be applied to the artificial systems to similate certain preportions of biological cognative. Not to forget to mention the better theory for networking process and storage systems (researching a more ergonomical method). (Although all my understanding of neural networks is really derived from the discussions and information I occasionally read about Artificial intelligence programming) If your really interested in the utilisation of neural networking you might want to try and find a copy of "an Introduction to Neural Networks" by Igor Aleksander and Helen Morton (It's a very good book), If you have a look through it, then you should be able to establish which direction to turn.
I was attempting to understand any precurser information I might need to bridge the gap. I have found many recommended books on intro which are expensive, but I thought perhaps I should understand alogirithms first. Anyhow - I have found a giant online intro which should help
Can you post a link to that intro I want to see it as well. BTW why don't you take a few college courses if you're really serious?
Specialist..... .....I am currently IN college. There are no college courses on neural networks offered in nearly any college in the country. They're only graduate level courses. Here's the guide..... http://www.statsoftinc.com/textbook/stneunet.html
If you really want an "exchange of wanted knowledge", it's a bad idea to start out by lying. If you are IN college, then you are a student. You are NOT a "neuroscientist". By the way, not knowing how to change a tire may not disqualify one as a mechanic, but not being able to find the carburetor surely would! Calculus and differential equations are requisite for a "neuroscientist".
I didn't lie little child. I am a neuroscientist because I spend 15 hours in the lab each week studying electrical impulses in rats. I remove their skull cap, stimulate certain section - then place a remote controled stimulator in their brain and seal the skull with a temporary cap. I study many instances here and go through roughly a dozen rats a week. Yes I am a neuroscientist you feeble minded twit.
And no calc is NOT requires. Learn to read a graduate level curriculum you high school drop out welfare poster child.
LMAO!!! Halls of Ivy gets told to learn how to read graduate level preceeded by a preschool level sentence.
I read in a book that the brain has the possibility to create around 10^2000 interconnections. If that's true I have a hard time seeing how AI is ever going to be feasible Please Register or Log in to view the hidden image!
that is a HUGE number that I find hard to believe. (it may be true.) but i find it HARD to believe. -AntonK
Understanding algorithms is the most important thing in most scenarios. I guess what you should study depends on what you are trying to do however. NNS are often a pointless academic over-complication, almost any organic cognition process can be successfully and robustly emulated in a digital system. I've worked in commercial expert and cognitive systems for a relatively long time and I've never needed to use NN tech for anything.. ever. Once you have spent time refining your algorithms to their simplest forms, they almost invariably mirror the organic system. The management layers are the tricky bit, but in time you will see what works. IMHO studying simple insect and fish behaviour is best when trying to comprehend the logic behind complex animal cognition and mnemonic processes. Its also a lot less hassle for the lab rats Please Register or Log in to view the hidden image! Again, it really depends on how you intend to apply your research in the real world.
In fact. Calculus is required. One of the major things to push forward artificial neural networks last century was the idea of back-propogation. Back propogation is the reworking of synapse weights based on the resulting outputs for a given input. Back propogation took a while to come into play because of the calculus behind it. http://www.faqs.org/faqs/ai-faq/neural-nets/part2/section-3.html That should help. Though I doubt you'll understand, as it is that you are not a scientist but simply a lab assistant. I had friends doing this at age 13 in my middle school. Your experience (or lack there of) in biology couldn't possibly help you in artificial neural networks at this level. You need at least a year or 2 of calculus background and at least a year or two in programming and basic algorithm techniques as well as some training in parallel computing (since neural networks are parallel computation machines). Please step off your high horse or you're going to get bucked off. -AntonK
