Can an ANN Use a SR-Latch to Store Fuzzy Floating Point Variables?

AI Thread Summary
The discussion revolves around the use of an artificial neural network (ANN) to store a fuzzy floating point variable, inspired by the concept of a Set Reset Latch. The objective is to determine if input signals are consecutively low or high. The ANN has been modified to calculate error through the difference between the first and second derivatives of the end layer's nodes, aiming for convergence with minimal changes. An activation function has been adjusted to create three convergence areas: -1, 0, and 1, represented visually in a video demonstrating the ANN's output with guitar notes. The conversation touches on the potential profitability of the ANN, suggesting that its success depends on solving a marketable problem. However, the thread concludes with a critique of the initial presentation and a call for clearer questions to facilitate more productive discussions.
ADDA
Messages
67
Reaction score
2
I'm attempting to use a artificial neural network to store a fuzzy floating point variable. While writing the code, I became somewhat creative. I used the idea of an Set Reset Latch and statistically translated the hardware SR-latch into an artificial neural network. My mathematical goal is to determine if input is consecutively low or high. To achieve this, I've modified the neural network idea slightly. Since I did not have a way to calculate error (also a reason for thinking this up), I pointed the error vector down in terms of the difference of the second derivative and first derivative of the end layer's nodes, so that the network would converge on less and less change of nodes. Also I've modified this activation function:

1.0 / (1.0 + e^(-x))

to allow for three convergence areas: -1, 0, 1, as seen in the following output of kmPlot:

Screenshot_20170927_163157.png
 
Computer science news on Phys.org
Okay. It is a little hard to read. But. What is your question, I do not seem to see one?
Could you please post a graphic image which would help us older citizens read your output?
 
I guess, I was wondering if my idea is valid mathematically. Or is it just creative?

I'm unsure how to post media... other than how I have already. Would you like the equation and explanation, jim mcnamara?
 
I've made the following video to show output of my ANN:

The dot pairs represent notes on a guitar: open to the left, low E on the bottom. I start with an open D. The colors represent convergence areas: red (-1), green (0), blue (1). The top dot is one SRANN's output, corresponding to the input of a octave overtone; the bottom dot, an octave fifth overtone. There is an input multiplier in the lower right corner. Magnitude is represented by the alpha channel. A convergence of 0 (green) discards a bright note as being an overtone. The goal is to distinguish between, let's say D3 and D4, or E3 and E4, as played halfway through the video.
 
ADDA said:
Does this ANN make Benjamins?
That depends on if you are selling a solution to a problem someone is willing to pay for. There’s a rule-of-thumb that says if you aren't profitable after doing something for a year, it’s just a hobby.
 
stoomart said:
That depends on if you are selling a solution to a problem someone is willing to pay for. There’s a rule-of-thumb that says if you aren't profitable after doing something for a year, it’s just a hobby.

Literally, I was being cute. My original title was "Does this ANN make sense." Then pun'd on cents to bills to benjamins.
 
FWIW - Since I thought you we trying to make a joke I let it slide. Rather than locking the thread that was basically nonsense. When you expect an answer ask a clear question. You will notice your funny question got no useful answers. This thread is going nowhere. Closed. Please consider asking a clear question.
Thanks.
 
  • Like
Likes Evo
Back
Top