Question

sir i need a problem statement about unsupervised representation learning..

Answer 1

Please leave a LIKE and comment for doubts

--------

You have just asked for the problem statement hence I am assuming that you know all about unsupervised learning and is confident enough that you will solve the problem by yourself but still, let me explain you something before giving you the problem statement

Representation learning is a way of extracting features from data that are unlabeled by training a neural network.

You must know that even after making so much progress in the field of machine and deep-learning still unsupervised learning is a largely unsolved problem.

In the unsupervised learning algorithm, there is no need for supervision and it can learn all the useful representation which is needed in order to solve any problem with very limited label information.

In the case of unsupervised learning, the algorithm will try to discover the latest features which describe the data set structure under certain assumptions that can be implicit or explicit.

These algorithms are can derive insights directly from the data itself and then work to a group or summarize the data so that the insights can be used to make any data-driven decision.

Problem Statement
There is a bank name FED which wants to change its way of operation and wants to make it more efficient.

In order to do so, they want to divide their customers into different groups so that they can recommend them the right products.

For example, the bank FED has a customer base of around 10K, and 2K of the customer is a youngster and about 6K is professional and rest are older people and they also want to divide them among gender to make it more efficient.

Approach
In order to do the most efficient approach is to follow data-driven way and based on their ages they can segment the customer and after that from these segments, they can derive the insights which will help the bank to make better product recommendations for their customers, and in this way, it will increase customer satisfaction.

Let me give you one more problem statement

In this problem statement, you need to organize a photo gallery lets just say there are more than 2000 images in my smartphone and when I am going through these images every third image turns out to be useless and unnecessary.

So I will need an app which can organize my photos in a way that I can go through most of the images and can
look at it if I want to and this would help me to understand the different kinds of photos I have on my phone.

Approach
If you will go through your pics you will definitely be able to arrange them in a different category like some of them are memes, some of them are family photos while some of them are work photos and so on

• So you can arrange them on the basis of time

• or you can arrange them on the basis of location

• or we can possibly do is that we can extract the semantic meaning from the photos and then can use it to define my collection.

This will the unsupervised way of solving problems

I won't explain much further because I think I have already explained more than what you asked

I hope this will be useful.