HackTheValley
Hakathon
Idea Description
One of the unique characteristics of the online world is its that one can interact with thousands of people yet remain relatively anonymous. Such anonymity is seen to create alter personalities. When people do not feel the consequences of their actions affecting them personally this tends to create “toxic”/harassing environments, as is proof from YouTube’s comment section.
H-Rase is an application that runs as a service in conjunction with and complimenting online communication platforms such as YouTube and Twitter. H-Rase can filter out harassing comments or discussions, such that the user using H-Rase will not be able to see comments, posts or tweets that are deemed at a high-level risk of harassment by H-Rase.
We hope that people of all ages and backgrounds will use H-Rase, but more specifically we have a focused interest on pushing the service to a younger demography; child demography all the way to early twenties.
We created the app as a measure to curtail the negative affects of online harassment may have on the psychological health of people, more specifically children, teens and people in early twenties. One of the glaring characteristics of online conversations on platforms such as YouTube, which is one of the most popular websites in the world, is that conversations can quickly turn into arguments and ultimately lead to harassment. In one sense, arguably, it has become a part of online culture to use abusive language to communicate. Evidently online harassment is an ever-increasing reality. H-Rase for the first time brings a real-time harassment filter which allows the user to browse the web without the threat of entertaining abusive language or facing harassment.
How we hope to implement it: Lets take the example of Youtube and how H-Rase would work with YouTube. Once a webpage loads H-Rase extension for chrome or H-Rase application on android/IOS quickly reads all the YouTube comments posted on that video, using the deep the learning algorithms hosted on Microsoft Azure it can understand the make-up and construction of a post that can be deemed harassing. The application gives each post a certain level on the harassment scale, high level posts will be filtered out and not loaded on the browser.
We aim to have a versatile app, we will have different levels of filtration. A user can choose from low level filtration to high level of filtration, such as adult filtration and child filtration.
If a certain post is not filtered out and a user finds it offensive, they have the opportunity to take a screen capture of the post and submit it to the Microsoft Azure servers, where the post is first deconstructed to see if it qualifies as a post which could be potentially harassing. If it matches the criteria, it is then fed into the deep learning algorithm such that this post or a post of similar construction are filtered out next time.
The incentive we give to users to report a post is through a point system and reputation system, like the one on Reddit. We have a leaderboard functionality in our app that shows the users with most reputation and greatest points. Users with most points and reputation will have their posts show at the top of the comments section. Some ways to earn points is by reporting harassing posts, it is by maintaining a good standard of one’s personal posts. Reputation is earned by the upvotes received for a post and the keeping a good historical standard of posting.
Log in or sign up for Devpost to join the conversation.