Proof of upload (and submission) to App Store
Our AI algorithm reads about 15 articles on the topic and generates a summary that only includes the information that all sources reported.
At the end of the short summary, the direct links to the articles used to create the summary is presented to the user in card format.
This is one reference article that was collected by the spider and formatted in the app.
References in this specific article that a user can click on. Tags that are associated with the article are also included.
When reading news articles, we're aware that the writer has bias that affects the way they build their narrative. Throughout the article, we're constantly left wondering—"What did that article not tell me? What convenient facts were left out?"
What it does
News Report collects news articles by topic from over 70 news sources and uses natural language processing (NLP) to determine the common truth among them. The user is first presented with an AI-generated summary of approximately 15 articles on the same event or subject. The references and original articles are at your fingertips as well!
How we built it
First, we find the top 10 trending topics in the news. Then our spider crawls over 70 news sites to get their reporting on each topic specifically. Once we have our articles collected, our AI algorithms compare what is said in each article using a KL Sum, aggregating what is reported from all outlets to form a summary of these resources. The summary is about 5 sentences long—digested by the user with ease, with quick access to the resources that were used to create it!
Challenges we ran into
We were really nervous about taking on a NLP problem and the complexity of creating an app that made complex articles simple to understand. We had to work with technologies that we haven't worked with before, and ran into some challenges with technologies we were already familiar with. Trying to define what makes a perspective "reasonable" versus "biased" versus "false/fake news" proved to be an extremely difficult task. We also had to learn to better adapt our mobile interface for an application that’s content varied so drastically in size and available content.
Accomplishments that we're proud of
We’re so proud we were able to stretch ourselves by building a fully functional MVP with both a backend and iOS mobile client. On top of that we were able to submit our app to the App Store, get several well-deserved hours of sleep, and ultimately building a project with a large impact.
What we learned
We learned a lot! On the backend, one of us got to look into NLP for the first time and learned about several summarization algorithms. While building the front end, we focused on iteration and got to learn more about how UIScrollView’s work and interact with other UI components. We also got to worked with several new libraries and APIs that we hadn't even heard of before. It was definitely an amazing learning experience!
What's next for News Report
We’d love to start working on sentiment analysis on the headlines of articles to predict how distributed the perspectives are. After that we also want to be able to analyze and remove fake news sources from our spider's crawl.