WYA? (Where You At)

The acclaimed Netflix documentary The Social Dilemma unveiled hidden machinations behind social media and its control over billions of users. For our generation, the millennials and gen Z, social media forms a very important aspect of our lives. Almost all our experiences make it to Instagram as pictures, or on Twitter as texts, or Facebook Posts or even on LinkedIn. Tristan Harris, the primary subject of the documentary, has spent his career studying how today's major technology platforms have increasingly become the social fabric by which we live and think, giving these platforms dangerous power over our ability to make sense of the world.

Following in his footsteps of uncovering hidden security issues, our team (Ayushman Panda, Kancharla Aditya Hari, P. Sahithi Reddy, Pratishtha Abrol, Snehal Ranjan, and V.J.S. Pranavasri) came up with the idea of this project. Where You At attempts to gauge how a person's location information can be inferred from his posts online. Location forms a vital bridge connecting the virtual world to the real world, and has multitude ways to be misused, which makes it an important aspect to study, keeping in mind the privacy and safety of the users of social media.

For the first and foremost step of the project, it was important to know which, among the wide variety of social media platforms would be the most suitable for the scope of this project.

For this, we had to conduct a survey, the targeted age group being 15 to 30, according to which unsurprisingly, Instagram easily made it to the most influential and most used platform. In fact, on further studying about user engagement on Instagram, we found that close to 100 MILLION posts are made per day on Instagram. This made Instagram the perfect choice for our project.

In another survey conducted by us, on the same participants, we asked them of their preferences about revealing the location of their posts, 43% responded with aversion, close to half of that percentage stating that they do it for privacy concerns. Although, the results of our survey did show that the participants were aware of some of the privacy issues, they were unaware that even a friend in their network, or even a random post with a similar background could be used to identify their location. This reinforced the idea of this project.

 

Implementation

Deep Learning Models formed were used to analyze images and extract features from user's social media posts. These features would later be used to create a comparison with the posts made by the people in the victim's social network. We limited our search to only the people in the network of the target user to reduce computational load on our devices. This limit is easy to expand, by creating a larger image dataset, or even using location tagged images from Google, but would require infrastructure and time, and thus could not be included in the scope of this project. We encourage people to build upon this and carry out further research in the field.

Our model followed the following approach:

 

  • Scraper: The scraper forms the first step in the implementation of this project. It creates a dataset according to each image url input, scraping the target user's network for their friends and their posts. It outputs a file of images, metadata and a list of followees.

  • Filter: The scraped images from the user's network are passed through a filter, to filter out images with no location tags.

  • DELF model: DELF (DEep Local Features) is a TensorFlow model, useful for large-scale instance-level image recognition. IT detects and describes semantic local features which can be geometrically verified between images showing similar object instances. DELF forms the core of the whole model, algorithmically matching images with the same background data, to output the best matches and the probable location information.

 

Result

The picture above showcases an output example. The red dot represents the victim, green, the victim's friend network, blue represents the posts with location information, and finally, the gold ones for the posts with a high similarity score. It can easily be drawn from the result, an image of another friend with the same location as the victim.

 

Click here to see our project explained in a video.


Comments

Post a Comment

Popular posts from this blog

Applications of NLP in Privacy Policies

Image
  Photo by Marija Zaric on Unsplash How many privacy policies have you accepted without knowing what the policy states? We're guessing that number is probably pretty high. And you are not alone. A  survey  of 2100 individuals conducted reported that 87% of individuals accept policies without reading them. Common reasons for this included the length and general readability of the policies. Of the top 500 websites, the average reading score was comparable to that of a high school student , while the average reading level of US adults is that of the 7th grade. Luckily, Natural Language Processing (NLP) can help. For our project, we set out to detect the readability of current privacy policies (spoiler alert: you'd need to have at least completed high school to make sense of what you agree to). We also wanted to see how we could make the readability of the policies better. So what did we do with the privacy policies? We used privacy policies from the  OPP-115 Corpus ...
Read more

"TL;DR: No more an excuse!": On making privacy policies easier to read and interpret

Image
Introduction Many websites collect, share, and use users’ Personally Identifiable Information (PII). Many regulatory bodies around the globe have long enforced requirements on posting privacy policies online. Over the past few decades, however, the collection, sharing, and usage of users' PII have risen to a major privacy concern over the internet, so much so that newer laws have gone into effect to protect user privacy. Prominent examples of such laws are the General Data Protection Regulation (GDPR) in the European Union and the California Consumer Privacy Act (CCPA) in the United States. Research has shown, time and again, that privacy policies are too long and hard to comprehend for their intended users, and therefore users rarely take the time and effort to read them. Privacy Policies are getting longer over time and increasingly difficult to read. To address the poor readability of privacy policies, we have developed tools that leverage machine learning (ML) to automatically s...
Read more

SiPP - Simple Privacy Policy

Image
SiPP: Simple Privacy Policy     SiPP Introduction Privacy Policies are legal documents that are long and complex that enable the initiated to select their preferences and cripple the uninitiated into opting in by default into the conditions by the organizations to utilize their service.  In today's world, technology is an integral and influential part of most lives. While technology provides services, and to provide services, user-specific data that can be interpreted into information is crucial. Privacy policies are the means to collecting this data legally while providing services and monetizing the same for other incentives. [Trivia: in 2017, data became more valuable than Oil] For generalizing the privacy policies, universal guidelines are established.  FTC Privacy Principles OECD Principles Fair Information Practice Principles We introduce SiPP, Simple Privacy Policy, that enables you to take smaller bites into the privacy policies with ease by selecting cat...
Read more