Second—“The paths that gets used the most strengthen, those that are less useful weakens over time” seems good to include, but you don’t make any mention of how that translates into ANNs. Maybe that comes later? Apart from that, I prefer the short version.
Third—Both discussions of threshold functions are kind of awkward. Long version—the first part makes it sound like the activation function is Heaviside: either 0 or 1. You attempt to correct that impression, but you don’t say what an activation function is, so it doesn’t work. Short version—depending on your audience, “sigmoid” and “hyperbolic tangent” need to be defined; and the second part again seems to suggest Heaviside contraray to the first part.
I think I’d recommend you just leave the Heaviside impression in place (maybe correct it in a footnote). “Artificial neurons imitate this behaviour using a threshold function. Each neuron weights each of its inputs, adds them together, and propagates the signal if the combined input is above some level.”
I am pro-[pic], depending on the pic.
Fourth—you can’t leave the “short” version there; you’re opening a plot thread and then not closing it. If you’re going somewhere with it (discussing/mentioning different ANN models, or how this model can be made manageable, or something), then great, but you might want to not start it until the point where you take it somewhere. If not, you should probably just cut it.
“The first to be considered and implemented” has similar problems, but less strongly.
Thanks for the lenghty analysis, yes I’m aware that these are the first drafts of a first draft, and I don’t even know if they will retain this form, but your comments made me realize that a shorter, faster pace is the way to go. Thank you very much!
Both versions need proofreading. That aside:
First paragraph—short version.
Second—“The paths that gets used the most strengthen, those that are less useful weakens over time” seems good to include, but you don’t make any mention of how that translates into ANNs. Maybe that comes later? Apart from that, I prefer the short version.
Third—Both discussions of threshold functions are kind of awkward. Long version—the first part makes it sound like the activation function is Heaviside: either 0 or 1. You attempt to correct that impression, but you don’t say what an activation function is, so it doesn’t work. Short version—depending on your audience, “sigmoid” and “hyperbolic tangent” need to be defined; and the second part again seems to suggest Heaviside contraray to the first part.
I think I’d recommend you just leave the Heaviside impression in place (maybe correct it in a footnote). “Artificial neurons imitate this behaviour using a threshold function. Each neuron weights each of its inputs, adds them together, and propagates the signal if the combined input is above some level.”
I am pro-[pic], depending on the pic.
Fourth—you can’t leave the “short” version there; you’re opening a plot thread and then not closing it. If you’re going somewhere with it (discussing/mentioning different ANN models, or how this model can be made manageable, or something), then great, but you might want to not start it until the point where you take it somewhere. If not, you should probably just cut it.
“The first to be considered and implemented” has similar problems, but less strongly.
Thanks for the lenghty analysis, yes I’m aware that these are the first drafts of a first draft, and I don’t even know if they will retain this form, but your comments made me realize that a shorter, faster pace is the way to go.
Thank you very much!