Replicating or Reproducing the Future: An Ethnographic Approach Towards Machine Learning Forecasting
|Event Name||Replicating or Reproducing the Future: An Ethnographic Approach Towards Machine Learning Forecasting|
|Start Date||30th Sep 2020 11:30am|
|End Date||30th Sep 2020 12:30pm|
On 30 September, we welcome Antonio Ballesteros (STIS, Edinburgh) as part of the SKAPE Seminar Series 2020-21.
Quantitative tools (QTs), like forecasting used for policy, base some of their authority on the idea of replicability as a form of transparency. This seminar will discuss the results of a participant observation of the Violence Early Warning System (ViEWS). This tool is aimed to forecast the probability of political conflict within the next 36 months and of climate-related conflicts within 100 years. While part of its core pillars, for its producers, a combination of unrecorded mundane activities and an existing infrastructure inequality make replicability an unachievable goal. Instead, they recognise that reproducibility might be the best possible outcome. This acknowledgement implies that ViEWS could be done elsewhere without a full understanding of the rationale behind its production. The distinction between replicability and reproducibility highlights the lack of a conscious understanding of quantitative tools used for policymaking.
The seminar will take place online via Zoom (for link details, please email the Co-Directors - if you are already a member/associate member of SKAPE, then you will receive a Zoom link two days before the start of the event).