Machine Learning- It's Not Just Algorithms===>The Beginner questions !!!

Let's Start with AI...

In computer science the term artificial intelligence has played a very prominent role. The term has become more popular due to recent advances in Artificial Intelligence and Machine Learning.

One of the finest example is ChatGPT ,machine started to give solution to many of the human problems.But how it is happening? Because Machine started to find the patterns inside the give data

Machine learning is the area of artificial intelligence where machines are responsible for completing daily tasks and are believed to be smarter than humans.

They are known to learn, adapt and perform much faster than humans and are programmed to do so. Robotics and integration with IoT devices have taken machines to think and work to a new level where they out-perform humans in their cognitive abilities.The vast amount of data generation from multiple sources everyday is also one of the reason for these AI Evolution.

What is Machine Learning?

Is it the set of Algorithms or neural networks to predict the patterns inside the given data? 

Yes..But it is not enough to learn just algorithms.It needs Statistics,Probabilities and Linear Algebra to add the value to your result.

Is Machine Learning  Statistics and Math?Again Yes

Experts saying  in order to correctly evaluate the powerful impact and potential of machine learning methods, it is important to first dismantle the misguided notion that modern developments in artificial intelligence are nothing more than age-old statistical techniques with bigger computers and better dataset

You need to be Skilled in Statistics to become Data Scientist or ML Engineer? No Need

You can train the Regression or classification model without using any statistic knowledge.you will able to read and understand a relevant Algorithm and train the model with better accuracy.

But...

If anyone ask ,how you calculate the variance of a population, or Did you face outlier issue or how to define marginal probability or what method you prefer to handled skewed data,you likely would have gotten blank stares.

If machine learning is a subsidiary of statistics, how could someone with no background in stats develop a deep understanding of cutting-edge ML concepts?

Experts would certainly advise anyone interested in becoming a Data Scientist or Machine Learning Engineer should get a deep intuition of statistical concepts at some extend.None of this is to say that ML never uses or builds on statistical concepts either, but that doesn’t mean they’re the same thing.But Engineers believe that the evolution of Neural Network will takeout the usage of Statistics in future.

Examples of machine learning algorithms:

Linear Regression

Logistic Regression

Decision Tree

Artificial Neural Network

k-Nearest Neighbors

k-Means

If you see indepth of each algorithm above, each has its own statistical and mathematical concepts inside like Probability,Vector space,linear equations,calculas,Eucledian distance etc..

I am sure,these concepts are nothing new to us,everything is learned in our schools.But it was not taught to us how to implement it in AI.

Apart from all this questions and confusions about Machine Learning we conclude with one general definition that :

Machine learning is a computational algorithms which iteratively “learn” an approximation to some function. 

Pedro Domingos, a professor of computer science at the University of Washington, laid out three components that make up a machine learning algorithm:

1) Representation: Transform the input from one form to other form and portray the data in the form where model can understand and give better accuracy.For example:Convert string value to vectors using Encoding Techniques,Standardization and Normalization techniques.

2) Evaluation: How well the model is trained with the given input.Usually we use some 'Loss function' to evaluate the given data id predicted with the accurate output.Example:Accuracy,Confusion Matrix,Precision.

3) Optimization: Optimize the representation function in order to improve your evaluation metric and increase the accuracy of the model by keep training the model with different hyperparameters like learning Rate,Cross validation methods.For example:Gradient descent method to keep track of error and fix the parameters which provides bottom most error value as highly accurate model.

Conclusion:

Machine Learning is not just Math or Statistics. Since these fields are not mutually exclusive, but that does not make them the same. Statistics is not invaluable in machine learning research and many statisticians are at the forefront of that work.

But still evolving neural networks and the possibility to train the model with millions of parameter with advancement of CNN,LSTM,Transfer Learning kind of technologies which have led to advances in several domains,particularly in sequential decision making and computational perception. It has found and made use of incredibly efficient optimization algorithms, taking advantage of automatic differentiation and running in parallel on blindingly fast and cheap GPU technology.

MACHINE LEARNING IS NOT JUST REGRESSION OR CLASSIFICATION ALGORITHM.....

Will Discuss further topics in upcoming  posts.

Stay tuned ! Keep Learning

Please do comments for any topics required.