1) Artificial Neural Networks (ANNs) & Feedforward Neural Networks

Hello guys this is Saksham Jain and I welcome you 
all to The FutreCamp. So in this video we will be   learning very briefly that what is artificial 
neural network how do we get how did we get   the name artificial neural network what is feed 
forward artificial neural network and how is feed   forward ann or artificial neural network different 
from a regular artificial neural network so all   these things will be covered in this brief video 
so without wasting any time let's get started so   neural network word has been derived from neurons 
so you must be knowing that neurons are also   called our brain cells and these brain cells or 
neurons extend throughout our whole of the nervous   system all these neural networks are connected to 
each other and they communicate with the help of   electrical impulses or signals so we know that our 
brain is made up of neurons we also understand the   working of these neurons how do these neurons work 
so what we had tried to do is what the scientists   tried to do is they tried to replicate these 
neurons they tried to replicate the working of the   of these neurons and since we have replicated 
or i can say simulated this neural network   this network of neurons artificially hence we got 
the name artificial neural networks now let's talk   about feed forward artificial neural network so 
feed forward ann is a part of artificial neural   network so that in the later stages when you will 
be learning about other kinds of neural networks   and many other complex neural networks when you 
will later on learn about cnn's rns and every   other neural network whatever you learn there 
your base is going to be of artificial neural   network especially of feed forward artificial 
neural network so if you don't have this base   if you don't know the concepts of ann you won't 
be able to understand the concept of concepts   of cnn and other neural networks in logistic 
regression we feed the input to our neuron   you must be knowing that in our logistic 
regression we feed our input to our neuron   and we get our output so here this is the working 
of our logistic regulation the inputs goes to this   head and we will get a prediction from this so 
now let us suppose that i insert one more neuron   now both these neurons this is this 
is first neuron this is second neuron   this is one neuron this thing is one 
neuron now this is the second neuron now all these neurons get the same input but 
they surge or i can say they compute different   outputs so this neuron might be looking from uh 
might be looking for some different features or   can be computing some different values and this 
neuron will be looking for some other features   and might be computer computing 
some other value now let us suppose   i insert one more neural neuron here and so 
on i am able to connect a network of neurons   now since there are many neurons or i 
can say many logistic regression neurons   so i can say that this is similar to 
simply multi-class logistic regression   here the work of each neuron is to compute a 
different feature or to compute a different output   or we can say that to look for different features 
and here you can see that i am feeding some input   and i will be receiving some output from the right 
hand side i will be feeding input from the right   uh left hand side and i will be receiving and i 
will be receiving output from the right hand side   since this is going towards this direction 
forward direction i feed the input from the left   and i receive the output in the right so uh hence 
i got the name as feed forward neural network in   feed forward neural networks there is no looping 
between the neurons you know that in brain all the   neural neurons are connected to each other in form 
of loops in forward direction backward direction   every direction they are connected to each other 
so but in feed forward neural network they are   connected in only one direction that is forward 
from left to right here you can see that all these   all the neurons are connected to each other they 
are forming kinds of loops they're forming loops   in the nodes so feed forward neural network 
won't look like this it will some it will   somewhat look like this so this is a perfect 
example of feed forward neural network here   i feed the input from the left hand side and i 
will receive the output from the right hand side   so this is the difference between artificial 
neural network and feed forward neural network   in artificial neural network what i do is i have 
tried to simulate the working of a brain feed   forward neural network is a part of artificial 
neural network that goes from the where the   neurons go from left direction to the 
right that is in the forward direction   so that is the difference between food for neural 
network and artificial neural network so that was   all for today guys i know that this video was very 
brief but uh believe me in the upcoming videos we   will be covering each and every part of artificial 
neural network that is forward propagation   loss functions optimizers back propagation 
hyper parameter tuning and many other things   we will be covering each and every topic in in 
a much detailed way so let us meet in the next   video in the next video we will be discussing 
our learning path our roadmap for artificial   neural networks particularly feed forward enns so 
let us meet in the next video until then bye bye

You May Also Like