The Ontology of Big Data: A Complex Realist Perspective

--> -->
Many thanks to everyone at the ODYCCEUS Project for the opportunity to present my ideas in Venice, January 29-30, 2018 -- in particular, Eckehard Olbrich, Massimo Warglien, and Petter Törnberg.  It was a great symposium!

My presentation, The Ontology of Big Data: A Complex Realist Approach, comes from the first section of my forthcoming SAGE book with Rajeev Rajaram, titled appropriately enough, Data Mining Big Data: A Complex and Critical Perspective
My argument is as follows:

1. Vis-à-vis the issue of big data, I think it useful to recognize that we are actually dealing with multiple forms of ontology – both in terms of the existence of which we are speaking and in terms of the form such an ontology takes

2. These multiple ontologies include: (a) philosophical ontology (i.e., realism, existentialism, positivism, etc); (b) information science ontology (a formal naming and definition of the types, properties, and interrelationships of the entities that really exist in a particular domain of discourse; (c) the web of information science ontologies, the world-over, which constitute a world-wide complex system of hundreds of thousands of colliding ontological systems, grounded in our global socio-cybernetic infrastructure; (d) the ontological structure of big data (velocity, variety, volume, etc); and (4) the ontological assumptions of the methods used to data mine big data (e.g., machine intelligence, cluster analysis, etc). 

3.  Together, these four intersecting ontologies creates a highly complex system of ontological systems – which are redefining the big data reality in which we presently exist.

4. Based on MacKenzie’s Machine Learners (2017) and Konys’s Ontology-based approaches to big data analytics, I would envision a new transdisciplinary field of study, something akin to the archeology of big data ontologies, focused on an archaeological investigation of the layers of coding and programs though which our big data globalized world is emerging – from machine language up – and the various intersections (or lack thereof) of their corresponding ontologies.

5. This complex system, in my mind, forms the new world(s) in which we live.  As such, specific attention needs to be given to: (a) how these complex structures form the colliding and conflicted ontological boundaries (or limits) of our digital and bio-ecological life; (b) and, in turn, how these contradictory and incomplete collisions (and the power relations upon which they are based and produce) define our epistemological understanding of contemporary digital and bio-ecological world(s) in which we live.  And all of this is important, because, in this new age of socio-ecological-cybernetic existence, everything is at stake!

6. A few possible ways to engage in such an “archaeology of big data ontologies” would include:

(a) embracing the framework of general complexity and, more specifically, complex realism, which treats all forms of ontology (philosophical, methodological, big-data, and information science) in complex systems terms;

(b) employing thenew methodological techniques of the computational and complexity sciences (e.g., case-based complexity, artificial neural nets, genetic algorithms, complex networks, agent-based modeling, etc);

(c)  and, to borrow an excellent phrase Adrian MacKenzie coined at our symposium, changing the social life of data analysis – which includes challenging the passive role of users in employing the tools and techniques of data mining and statistical analysis, as both the ODYCCEUS Project and my team’s development of the COMPLEX-IT App seek to do.

  • For a quick overview of critical realism, CLICK HERE.
  • For a quick overview of complex realism, read Byrne, D. (2004). Complex and contingent causation-the implications of complex realism for quantitative modelling. Making Realism Work: Realist Social Theory and Empirical Research. B. Carter and C. New. London, Routledge. CLICK HERE. 
  • See also, Williams, M., and Dyer, W. (2004). Complex Realism in Social Research. CLICK HERE
  • To learn more about OWL and its usage to develop information science ontologies, CLICK HERE

No comments:

Post a Comment