Neural network in the sphere of machine learning will not be just price knowing the algorithm’s technicalities but in addition might be about understanding more about ourselves.
Why Neural Networks?
While getting began on data science, I learned about various core ML stuff, however the one thing that actually caught my attention aside from the deep-rooted foundation in mathematics was after I learned concerning the Artificial Neural Network.
Developing an algorithm that mimics the human brain to finally attempt to create machines that may think and do cognitive evaluation like humans, is something that I discovered price exploring.
🙂
I all the time feel more drawn to a field of study after I could all the time relate it to each day life, or after I’m capable of pick up some life skills and understand more about myself from the subject under study irrespective of how scientific or technical the subject is. So tried writing on this neural network and life analogy.
PS: Also, it could look like a really obvious interpretation of neural networks but yeah I just wanted to write down about it as I liked it 🙂
Neurons and Weights
In neural networks, there may be an input layer and multiple interconnected hidden layers that finally produce an output. This exactly happens in our brain, at any given time limit, there are tens of millions of neurons all the time firing up giving us multiple thoughts and concepts.
Every decision we make is a product of some interconnected thoughts (activated neurons) with weights assigned to them. So ultimately it comes all the way down to that help activate any particular neuron that contributes to the hidden network that works as much as activate a final thought that drives motion.
:
Focus is what drives our dreams and ambitions and helps us to attain them.
Almost every trigger/thought is all the time not easy but might be rooted back to some past, recent happening, or future apprehension. The form of trigger/thought we put more weight on, the brain prompts similar thoughts and tries to form multiple hidden layers between them, assigning weights, and thus bringing our entire attention to that exact thought, which finally results in motion.
So we are able to just have one thought, and the brain will construct a whole kingdom on it. Amazing right?
Back-propagating Errors
Backpropagation is thatinvolves taking the error rate of a forward propagation neural network and feeding this loss backward through the neural network layers to fine-tune the weights.
The neural network learns and gets higher by back-propagating the error toward the contributing neurons and hence modifying the unique weight.
We, humans, are creatures of habits and training. Like all machine learning model that gets higher when trained with good input features, we humans too are products of coaching on any particular skill till we improve at it. And any training process will not be all the time a linear path but stuffed with trial and error till we master that skill.
This brings us to the proven fact that life needs to be lived forward while looking backward, or as precisely described by Steve Jobs’s famous quote: .
We improve at looking back at our mistakes and learning from them by adjusting our beliefs and future actions for upcoming events until we grow to be our greatest version. We ourselves are a machine learning model with training in progress, and the more we leverage this back-propagation of errors and learnings the higher we grow to be at what we train.
The sector of deep learning provides quite a lot of life learnings if we just pause and reflect. The concept of assigning weights gives us a superpower in itself, which is the . If we consciously don’t give much weight(importance) to some neurons which might be triggered, they never contribute to the ultimate equation of motion.
This might be an awesome tool for coping, or for healing some past trauma, because if we give minimal attention to those painful memories, they stop to exist, it’s like they almost never even happened or existed. Magic right?
Our brains are mostly projecting devices spotlighting thoughts based on some backend equation like in neural networks, but unlike artificial neural networks, we humans have the ability of discrimination, where we are able to take into consideration our triggering thoughts, and re-focus our attention.
Probably the most amazing a part of humans is that we’ve got this that’s lacking in these artificial neural network models or every other species. We’ve got the ability to assume, introspect and develop the world of our dreams and never just survive and perish.
And we understand this great thing about our brains a lot that we try to develop one, cause sure as hell, it’s in motion and deserves a celebration, and what higher method to rejoice than devoting a whole field of study towards its replication? And perhaps that’s what also keeps me hooked on this fascinating field of artificial intelligence and machine learning.
