Project Analogy

To make the world’s organized information more consumable

Vishant Prajapati

Following Follow project

Liked Like project

Become a Hackaday.io member

Just one more thing

To make the experience fit your profile, pick a username and tell us what interests you.

OK, I'm done! Skip

Similar projects worth following

0 comments

0 followers

0 likes

View Gallery

Description

So the idea falls under the category of text summarization.
A lot of internet companies have been doing this to enhance the user experience of their products.
The idea is not exactly summarizing but simplifying. For example, if I am reading a white paper on blockchain technology, it is very difficult for me to understand it in the first go or maybe not easy at all for a layman to understand.
Another example taken could be of poems. In school, we never would have really understood what the poet wanted to say until and unless it was explained by our teachers.
So the main idea is to build a tool which can simplify very complicated domain-specific records and publish it in a way which is understandable by a layman.
Something very similar to Two Minute Papers but for text and completed automated.

Details

Similar existing use cases:

A couple of real worlds examples are Google Play app reviews and customer reviews on Amazon. What they have done is, they have read all the comments mentioned and summarized it saying that these are the reasons why this app is good and these are the reasons why it is bad. A really good use case.

Use cases of the idea:

One use case I have figured is in the education domain where students can understand complicated topics within minutes.

How can it be built? Definitely through machine learning. Sequence-to-sequence learning model is the model that needs to be trained.

For our POC we can make a model taking the data from Two Minute Papers and see if it gives good results.

Challenges I managed to find out one big challenge till now and that is transferring the learning from one domain to the other as the Vocabulary is not shared. Meaning, a model built for the medical domain could be really tough to be used for engineering domain.