In this document the researchers propose a novel method to detect coordinated inauthentic behavior in Facebook pages and groups. This method modifies a recently published technique, with the intention of establishing a firm statistical foundation for defining “simultaneous link sharing” behavior.
A new wave of scholarship has made major advances in how we understand the politics of civilizational identity by drawing powerfully from conceptual tools developed over the years to study other forms of identity.
In this work, the researchers test the performance of two bidirectional transformer-based language models, BERT and SpanBERT, on predicting directionality in causal pairs in the textual content. The preliminary results show that predicting direction for inter-sentence and implicit causal relations is more challenging. And, SpanBERT performs better than BERT on causal samples with longer span length. The researchers also introduce CREST which is a framework for unifying a collection of scattered datasets of causal relations.
Home page - events - title
Monday, May 3, 2021 - 12:30pm to Tuesday, May 4, 2021 - 2:00pm
The Institute for Data, Democracy & Politics and the Social Science Research Council’s Media & Democracy program invite you to join them for The Conservative Dilem
Wednesday, March 24, 2021 - 1:00pm
AI Fairness and Transparency addressed the importance of independent, academic research into algorithms. Dr. Nicol Turner Lee of the Brookings Institution and Professor Robert Brauneis of the GW Law School discussed recent developments in the algorithmic transparency, equity and justice and digital civil rights.
Wednesday, February 17, 2021 - 1:00pm
As the Biden administration and the new Congress begin to grapple with how to bring transparency and accountability to massive digital platforms like Facebook, Google, TikTok, and Twitter, they may seek ideas and guidance from other countries and jurisdictions where regulatory work has been moving more rapidly.
Now it’s all about the ground game. With more than 2.7 million Massachusetts residents fully vaccinated and the pace of COVID-19 shots slowing markedly, state officials are stepping up efforts to reach the holdouts.
One of the strongest underlying drivers that determines whether individuals or groups are vaccine confident or vaccine hesitant is the level of trust—or distrust—in the individuals and institutions that discover, develop, and deliver vaccines.
With effective COVID-19 vaccines in hand, we must now address the spread of information on social media that might encourage vaccine hesitancy.