# Episode 1 Part I  |  Artificial Neuron - The Gate Keeper

## Метаданные

- **Канал:** Hedu AI by Batool Haider
- **YouTube:** https://www.youtube.com/watch?v=6pvMyNVqSy4

## Содержание

### [0:00](https://www.youtube.com/watch?v=6pvMyNVqSy4) Segment 1 (00:00 - 05:00)

in the depths of an ancient ocean there was silence and then life force spoke through delicate Alchemy signals emerged this is one of the first forms of communication but what do these signals mean are other life forms calling for friendship or declaring animosity make decisions from complex information something special was needed a brain and that is how one of the first nerve cells were born electric Messengers carried information from the external realm through branches akin to data highways called the synaptic connections when a neuron has heard enough of these signals it awakens and pronounces a degree foreign of the many ways intelligence emerged in the animal kingdom can we give this ability to a machine let us try is an artificial neuron let us call him Neil just like a biological neuron this too has an input connection and an output connection as our Neo is existing happily in dark in silence it gets a message from the external world an input involving sounds of tiny chemical bursts it's Neil only understands numbers everything needs to be expressed at cell let's Express sound in decibels Neo hence is able to take in the input do some mental processing and try to come with an output a prediction now in our small binary Matrix simulation Neo can make only two predictions zero or one in human language let us say want means safety while zero means danger rightfully predicting these is a matter of Neo's survival so what should happen in Neo's mind for accurate predictions primarily two things weighing and proceeding first with scaling the signals that the cosmos stands around can be anything Sound Energy Force and that can vary from positive Infinity such as the massive imploding pull of a black hole to negative Infinity such as the explosive push force of a supernova now if we squashed our infinite scale to something more manageable so that extremely small values in our data set become zero and really large values become one wouldn't that simplify our math and reduce the computational expense not just that scaling has a few more benefits one being the introduction of doubt new at the moment is quite arrogantly confident about his opinion decreeing either a sweeping zero or one true Intelligence on the other hand is always doubtful and uncertain about the potential unknowns suppose you were born centuries ago and you saw the sun go around the Earth every day this made you a hundred percent confident that Earth was the center of our solar system but then a little doubt trickles in and attempts you to question your assumption you create a tool and your first Peak inside it completely involves your reality therefore rather than an arrogant yes or no we need to equip Neo with an uncertainty scale that is a probability scale if his output is greater than say a threshold of 0. 5 new can be more certain that the input indicates safety if less than the threshold there is a higher probability that he might be in

### [5:00](https://www.youtube.com/watch?v=6pvMyNVqSy4&t=300s) Segment 2 (05:00 - 08:00)

danger now depending on Neil's environment it is possible to make him either more optimistic or risk-averse for instance if we increase the threshold we make new more alert to Danger this is true for humans too children raised in war zones or amidst domestic abuse are at a higher risk of deep-rooted emotional scores sadly the words color hair is not metaphorical if left untreated stress can actually rewire young brains resulting in disproportionate responses to situation that may not even be real threats coming back to the main discussion how can we obtain this probability scale well have you ever pondered over the shape of an s-curve it has a beginning that a period of accelerated gradient and finally a flattening slope interestingly many phenomenon in our world follow the same trajectory such as the age-old question does money buy happiness a few cents won't make many people happy well at least not in an expensive country like the US more money however can drastically increase the quality of your life and increase your happiness in this after being enormously Rich however more money may not have that big of an impact same goes for the spread of a rumor or a viral video first only a few people know about it then suddenly everyone seems to be talking about it and it blows up finally the news becomes old and boring and people move on to the next big thing same applies to the rise and fall of vampires or the growth trajectory of a startup now there is a name for this curve it is called a sigmoid curve and the good news is that it comes with a math equation that is almost magical any number you throw at this equation gets squashed to a scale of zero and one this is also called an activation function because its person artificial neuron to life akin to a biological it determines the neuron's destiny based on whether the threshold is surpassed or not so let us applaud a sigmoid curve to our neuron and just like that any input it takes gets converted to a probability great so far we've discussed the inputs the outputs and the scaling but you are probably still wondering where is the intelligence part

---
*Источник: https://ekstraktznaniy.ru/video/44642*