Weekend read - Ontologies and technologies: knowledge representation or misrepresentation

This article reads like a prophecy, does anyone know Kieron?

O'Hara, Kieron. "Ontologies and technologies: knowledge representation or
misrepresentation." *ACM SIGIR Forum*. Vol. 38. No. 2. New York, NY, USA:
ACM, 2004.

  The development of the Semantic Web (SW) raises a number of difficult and
interesting technical issues. Less often remarked, however, are the social
and political debates that it will engender, if and when the technologies
become widely accepted. As the SW is a technology for transferring
information and knowledge efficiently and effectively, then many of these
questions have an epistemological base. In this paper I want to focus
especially on the epistemological underpinnings of these social issues, to
think about the interaction between the epistemological and the political.
How does technology affect the social networks in which it is embedded? How
can technology be successfully transplanted into a new context? And,
perhaps most importantly for us, how is technology affected by its context?
In particular, I want to look at how our decisions about how we treat
knowledge can impact quite dramatically on the technologies we produce. Let
us begin with a familiar diagram, the layered view of the SW developed in
the very early stages by Tim Berners-Lee and colleagues (Figure 1).
Knowledge of different types is separated out (ontologies appearning in the
middle). And by making this separation, we can see how the semanticity of
the Semantic Web comes about. The technical machinery of Unicode and URIs
appears at the bottom; XML can be used to tell the computer what objects
this code is concerned with. RDF tells us that those objects are related in
various ways. Ontologies give a context for the object types and relation
types. We then need logic and inference engines to make contextsensitive
inferences over our knowledge bases. Proof theory tells us the properties
of our inference engines – soundness and completeness etc. And finally,
there’s no point being proven innocent if no one trusts the method of
proof. Figure 1: The layered view of the SW O’Hara Ontologies and
Technologies 2 Hence, ontologies are central to the vision of information
aggregation and manipulation underlying the Semantic Web.
Ontology-mediation within services enables much to happen; they can steer
the knowledge acquisition process, choreograph the integration of
information from diverse sources and representational formats, assemble
retrieved information into customised packages, and hence present
information to the right people in the right form, bring intelligence to
the search process, undercutting the direct human input to the drudge work.
An example of an ontology-mediated system which exploits the new
expressivity that the SW allows is the winner of the 2003 Semantic Web
Challenge, CS AKTive Space (Shadbolt et al 2004), which gives an up-to-date
snapshot of the state of the discipline of computer science in Britain. But
the point of view of the snapshot is determined by the user. In a massive
knowledge acquisition exercise, various different technologies assemble
giant quantities of research-related information from the websites of
computer science departments, the EPSRC and so on. This is all collected on
a regular basis, and converted into RDF. The RDF store contains tens of
millions of triples, and the space enables the user to query that store to
ask specific questions about British computer science. Who is working on
what? Who is working with whom? Where are the geographical concentrations
of researchers? Which areas are receiving the funding? Who are the top
people in different areas? Ontology mediation is central to this
flexibility in the presentation and selection of information; how does this
impinge on the social aspects of the SW? There are two ways in which the
Semantic Web can be seen as social. The first is that, as we ascend Figure
1, the layers become increasingly socially rooted. In other words, even
though all of these layers are amenable to technical solutions (even trust,
a social phenomenon par excellence, can be addressed technically, as we
see, of course, in the agents community, and increasingly in Semantic Web
services), as we move up, each layer is more beholden to social phenomena.
No matter how good the technical solutions provided, as we move along the
direction of the arrow, those solutions must respect more, and more
complex, social phenomena. They can be criticised by pointing out that, no
matter how impressive the solution, it fails to capture some vital aspect
of reality

https://dl.acm.org/doi/pdf/10.1145/1041394.1041397?casa_token=V8xeXiRSUGUAAAAA%3AZ-nHUlU8AXV58LtRUqKYhN7uiX8iS0l5DDn94BmqTaEM4QlqcnZFfeEeVnOzAc1I21i9cmp-hduu5P4

Received on Sunday, 27 June 2021 01:09:39 UTC