Understanding machine learning algorithms and artificial intelligence

By Jayadeep Reddy

Posted on July 24, 2017

Machine learning and artificial intelligence are the buzz words of the tech world, every day I get to see a new startup utilizing current AI and machine learning capabilities to solve some real world problems.

I recently chanced up on to use Amazon machine learning algorithm for one of the apps we are developing. The idea is to get outcomes based on the user purchase behavior and predict his next purchase. A prediction like this will be of great benefit for e-commerce companies, This will not only boost the sales of the product but also align the marketing strategy based on user behavior.

Now Imagine, One fine day you visit your favorite e-commerce store and the home page flashes you the product you are looking to buy on that particular day. That is going to definitely surprise you. Analyzing a large set of data is not a humanly possible task and with machines playing a crucial role, millions of records can be analyzed in few minutes with the highest level of accuracy. This kind of analysis throws light on the areas you could have never imagined.

Steps Involved in using machine learning for your business.

All you need is right data. ML Model

Data is the key recipe to make this machine learning possible. Computers are dumb and they will not have any logic to make a decision. The logic will have to be created with the use of data which is called ” Machine learning model”. ML model is the first step to create some sort of intelligence with your data.

There are different types models for the different type of solutions. Like Linear Model, Multi dimensional model, Regression Models etc. The complexity of these models depends on how complex the data is.

Train your model

Your first model is known to have less accuracy for prediction, but this is just the beginning, every time the model is used it keeps getting better on accuracy. This is called machine learning “You teach your model to keep getting better every time that is put to use”. This is also commonly known “reinforced learning” in artificial intelligence.  This keeps making your model intelligent and better every day. Some models might take a year to get matured enough.

Machine Learning VS AI

AI is more like a reward based algorithm where the program is baited with rewards to succeed and they are set to become faster and better every time they win a reward. AI keeps in memory of every event that leads to failure. This mechanism of learning is just like how we humans learn things from success and failures.

AI on other side is a superior set to multiple machine learning models. AI once mature can technically do any task that is done by a human. Like driving a car, Shopping for you, cooking food etc. We are not far away from cyborgs.

Where machine learning  can be used

Machine learning will be a solution to many business problems, Here are some examples.
1. Predict when will your next employee leave the organization.

2. Predicting disease well before doctor diagnosis. Scan millions of patient records to give clinical decision support.

3.  What is the best time in a day to send promotional messages?

4. Which mutual fund is going to give you maximum returns?

5. What is the best time to post this blog to get maximum user attention and virality?

6. Predicting the occurrence of crime.

7. What is the right time to avoid traffic congestion?

Technically any action that has attributes will have a possible outcome, machine learning makes it possible to predict and act in advance while AI makes the machines more intelligent.

Disadvantages of machine learning and AI

While machine learning and AI is essential to solving some real world problems, in a wider view algorithms are set to steal your jobs. Any job which is not highly intelligent will be replaced with AI and automation in next few years.  Recently Tesla CEO Elon Musk said “AI is a fundamental existential risk for human civilization, and I don’t think people fully appreciate that.”  and urged regulators to have laws in place to control AI practices across the globe.

NOTE: I am building super smart tech team ping me on linkedin if you’re interested. 

Read Artifical Intelligence startups in India

Understanding machine learning algorithms and artificial intelligence