It is widely acknowledged by the multimedia community that progress on new media applications has leapfrogged over the last decade. This significant trend can be traced back to the adoption of deep-learning technologies within main-stream scientific innovation. This trend has also led to the development of critical application including, among many others, generating text that describes images, automatic translation across multiple languages, automatic tagging of semantic objects in images and

more »

... he ability to perform highly accurate face recognition and person detection. Since machine learning and more specifically deep-learning models aim to emulate the cognitive functions of humans, there is strong evidence that derived technologies and applications are often subjected to misuse. Instances of such technology misuse can be found in the pervasive propagation of misinformation, fake pictures and videos and fake news, where the validity of the information distributed is not properly authenticated. Despite efforts launched by platform owners such as Facebook, Google, Twitter and others to combat the spread of misinformation, the capabilities provided by automated multimedia signal processing technology to combat such challenges are often underexploited. Parallel and complementary to the scientific advancements in the field of machine learning and deeplearning, developments in the field of big-data systems have also enabled software application