Amalgam

The relationship between media and power is older than the information age, in certain ways History is itself a mass media construction. History has always been mediated, however the magnitude in which governments, political parties and society itself depends on this mediatization became the defining signature of the last century. This media-power relationship was severely threatened by the advent of social media. The shift from the one to many construction, which is the natural way of mass media dissemination, to the peer to peer construction of social media created a false feeling of democratization of information, especially news.

However the "public space" that social networks created for self organizing is anything but public. Social Networks are billion dollar companies that protect and direct their own interest and agendas. Hence we haven't seen the shift from the monolithic media construction to a democratic one. News and social media present us a construction of events that favors their own interest. As the past is being stored physically in the present this situation creates a very grim space for social justice, fairness, and equality.

In addition there is one more invisible mechanism of control that is being applied to us in the information age: Filter Bubbles. Filter bubbles are algorithmically created systems that sift information for us. The best example is Facebook's time-line; each time-line is created by filtering the hundred of posts a user could see based on the interests and activity of that user. These filter bubbles have eliminated the richness of opposing views and moved us into a world where everybody agree with us; taking us away from the search of consensus and pushing us to extreme versions of our own biases.

Amalgam is the product of a very long research process, it all started with a very simple question: It is possible to remove the narrative elements from News and just keep the factual data? Can computation intervene in the space of subjectivity? This questions can spawn lengthy and complex discussions about the what is factual data and how the nature of computers, as something man made, makes them unable to be unbiased. The main elements of the natural language field used in this research are: Coreference Resolution, Named Entity Recognition, Sentence Breaking, Word Sense Disambiguation and Corpus Linguistics. All the iterations of this process were matched against humans using the Amazon Mechanical Turk.

Leveraging Social Media’s methodology to naturally fight against News Media biases by increasing the amount of different opinions, Amalgam aggregates bias in media to contradict itself. Amalgam retrieves article URLs and extracts clean text from the HTML file. The clean text is analyzed for Named Entities using the DBpedia Spotlight and the Alchemy API. Besides the database and the exposing mechanism, the system is also composed of 4 autonomous processes: Article retrieving, Information extraction, Metadata matching and a Temporal Aggregator.