Real Fake

Real Fake Real Fake

Is It Fake or Real?

In this paper, we describe a new methodology to separate photographs and computer generated images . We introduce the Progressive Randomization (PR) technique that captures the statistical properties of each one of these classes. Using only statistical descriptors of the least significant bit (LSB) occurrences, our method already performs as well or better than some comparable existing techniqu...

Real Fake DOMs: Verified Browser Rendering

In this writeup we present Real Fake DOMs, the first Coq development to our knowledge to model Document Object Models or the graphics primitives necessary to model browser rendering algorithms in a shallow embedding. As a first foray into verifying algorithms in this area, we model simplified DOMs with representative backgroundcolor, width, height, position, top, left, and overflow attributes. ...

Privacy through Fake yet Semantically Real Traces

Camouflaging data by generating fake information is a wellknown obfuscation technique for protecting data privacy. The effectiveness of this technique in protecting users’ privacy highly depends on the resemblance of fake information to reality, such that an adversary cannot easily filter such fake information out. In this paper, we focus on a very sensitive and increasingly exposed type of dat...

Fake News, Real Consequences: Recruiting Neural Networks for the Fight Against Fake News

The Fake News Challenge (FNC-1) is a public competition that aims to find automatic methods for detecting fake news. The dataset for the challenge consists of headline-body pairs, with the objective being to classify the pairs as unrelated, agreeing, disagreeing, or discussing. We developed four neural network models for FNC-1, two using a feed-forward architecture and two using a recurrent arc...