DEV Community: Sandesh

1943 - Deep Learnings inception & visual history.

Sandesh — Sat, 18 Jul 2020 06:47:04 +0000

The modern world fancy tech stack Deep Learning seems to be gaining popularity day by day nowadays. But is this tech new to the world ? hold ON !

The reality is deep learning existed for over 60 years now.

The main reason why deep learning flourished in recent years is :
1 : Availability of data.
2 : Modern techniques to store the big data.
3 : Modern computers to process the data with speed.

Given below is the short history of Deep Learning since inception.

Thanks for reading !

The Game of Modern AI - Glance

Sandesh — Fri, 03 Jul 2020 10:14:42 +0000

Hello Readers ,
For a modern Data scientist , if the boss asks us to build a forecasting/predictive machine learning model , the only job we have ( Besides collecting/cleaning data ) is to just import a python library and do model.fit() irrespective of any algorithms we use.

Now, lets have a look all the major machine learning/deep learning algorithms at one glance.

This image above will give you an idea on major fields in data science and some major algorithms who are unseen heroes in the backend of a successful AI product.

They also segregate various tasks that fall under specific type of an AI application.

Conclusion

As a newbie if you wonder what the hell is computer vision , natural language processing about , You now know some insight on various opportunities and fields available in AI world.

Thanks for Reading !

What Makes Us A Failed Data Scientist !

Sandesh — Fri, 19 Jun 2020 05:30:19 +0000

Data science was once termed sexiest job of 20th century. With the development of user friendly modules like scikit-learn, tensorflow etc. implementing a machine learning project became a kids job.

With all these hype and rush to learn and be a data scientist ,we have to remember these basic protocols so that we don't end up being unsuccessful despite our hard work.

So , these steps will definitely make you a FAILED DATA SCIENTIST which one has to refrain from when starting with DS :

● learning data science libraries before learning coding basics

● learning ML algorithms before learning how to preprocess your data

● learning deep learning before machine learning.

● learning data visualisation before understanding the basics of statistical inference

So , what can be done right ?

● You have to know coding basics before you can even debug the implementations of your DS/ML libraries

● You have to know how to preprocess your data before applying machine learning methods accurately.

● You have to know statistical inference before you make sense of your visualisation.

Conclusion :
Remember to do things in a right way. Because , we may have to begin all over again in case we get our first steps wrong.

Thanks For Reading!