As humans we are always curious about the things surrounding us. We always want to know something or the others for example weather, movie review, name of the device, etc. Internet has become a vital player in quenching our curiosity. But the problem with internet is the amount of noise it creates. At times it might lead to over of the same information or covering the other details we are looking for. So we wanted to build an application which feeds in content from the internet, process it and churns out the relevant novel information about the topic we are looking for.
What it does
Our solution takes in the search result contents for the given search keyword from the internet and select the best diverse search results.
How we built it
We used python to develop it. Used Google Collab, Anaconda distribution to build it. We also leveraged on techniques like word embeddings, rankSVM, similarity measures etc.
Challenges we ran into
We started our solution late. So we have to complete the end to end solution in less than 4 weeks. We also tried out different techniques before finalizing the current solution approach.
Accomplishments that we're proud of
Creating an working end to end solution in lesser time to address one of the real world problems.
What we learned
We learnt to use Tiger Graph. Usage of network embeddings from the graphs.
What's next for News Noise Reduction
We can optimize the application and make it process near real time. This application can act like a plugin and give you the relevant search results.
Log in or sign up for Devpost to join the conversation.