In honor of National Bullying Prevention Month, Securly has partnered with an expert in pattern recognition to develop an app that alerts parents and school administrators about incidents of bullying and harmful behavior.
WHAT YOU GET
We have developed a prototype for the industry’s first monitoring engine capable of capturing sentiments and emotions behind posts on social networks like Facebook and Twitter, and searches performed on sites like Google and Bing.
Much like Apple’s Siri interprets a user’s conversational request, our sentiment analysis engine uses natural language processing to interpret the keyword patterns in a message. Our machine learning algorithms then compare the keyword patterns seen with a database of flagged patterns from a training data set.
This allows us to flag any suspicious activity buried within millions of social posts and online searches. We then automatically send the parents an alert directly to their mobile device for serious issues.
These powerful algorithms enable us to distinguish the casual sentiment in “I hate this sweater” from the strong negative sentiment in “Everyone at schools hates me”, without the need for parents or school admins to set specific keyword alerts.
We then classify these suspicious activities into behavioral categories such as bullying, grief, and violence.
We also make it easy to get a complete trail of your child’s activity surrounding a specific post or tweet.
EASY TO SETUP
Parents can install Securly in one of three easy ways:
- If your child belongs to one of the thousands of schools already using our service, simply sign-up at securly.com.
- If you need to monitor your child across all devices at home, install our mobile app that makes a simple update to your home router.
- And finally, for monitoring use on mobile phones, all you need is our mobile app!
WHO'S WITH US
We are deeply committed to our new Sentiment Analysis offering. Within a week of beta-testing, we have already helped a customer school preempt a serious incident. But we need help. We are solving a very serious problem, and the solution involves advanced research and technology. We've turned to Kickstarter to raise awareness and secure the financial support necessary to push forward with the development of our technology.
We are grateful for the support of Associate Professor Yi Zhang, our new Kickstarter community and others like the Tyler Clementi Foundation
Thank you for your support!
The "Wall of Encouragement"
Backers of $50 or more can choose one of our custom shirts designed to spread the word about the issue of bullying in schools.
Risks and challenges
The ability to first perform natural language processing of free-form posts, searches and tweets, then performing machine learning on keyword relationships seen among previously classified data, and finally applying these learned patterns/relationships to newer data in order to classify the data into specific emotions is a hard problem to solve.
One of those challenges is gathering a broad enough data set to be representative of the many nuances and subtleties of sarcasm and hateful speech. Verifying that the results are accurate and providing the AI engine with honest feedback is critical to sorting this vast data set quickly and precisely. We plan to solve this challenge by 1) using our installed based of hundreds of thousands of students and staff to collect *anonymized* posts/searches from participating schools, 2) contracting with language specialists (under NDA) to classify our initial set of annoymized training data, 3) building reporting and alerting into the product UI directly, then 4) constantly reviewing the quality of our alerts from participating schools to make sure the alerts fall within acceptable levels of sensitivity and confidence.
Our founders with their network security expertise, our team with a dedicated NLP+Machine Learning engineer and our adviser Yi Zhang who is an industry expert in this space together form the right team to tackle this hard problem.
While we accept there will always be a few false alarms, it is vital that we do everything to avoid the tragedy of a missed opportunity.Learn about accountability on Kickstarter
- (30 days)