A claim for deeper semantics in a Social and Digital World - Fernanda Baiao

A claim for deeper semantics in a Social and Digital World

1. A claimfor deepersemanticsin a Social and
Digital World
Fernanda Baião
DepartmentofIndustrial Engineering
PUC-Rio, Brazil
Ontobras
November2020

2. Data Science &
Computational Intelligence
Industry 4.0 &
Digital Transformation
Social reality &
Humachineparadigm

3. (1)
Problem
understanding
(2)
Gettingthedata
(3)
Inernalcycleof
research
(4)
Visualizationof
results
(5)
Creatingactions
based onresults
(6)
Gettingfeedback
fromaction
Data Science life cycle
[Shcherbakov, M., Shcherbakova, N., Brebels, A., Janovsky, T., Kamaev, V., 2014, “Lean Data Science
Research Life Cycle: A Concept for Data Analysis Software Development”,
Knowledge-Based Software Engineering, v. 466, pp. 708-716, Springer, 2014
Fayyad, Usama, Gregory Piatetsky-Shapiro, and Padhraic Smyth. "From data mining
to knowledge discovery in databases." AI magazine 17.3 (1996): 37-37.

4. “We-re not passive consumers of data
and technology. We shape the role it
plays in our lives and the way we make
meaning from it. But to do that, we have
to pay as much attention to how we think
as how we code. We have to ask
questions, and hard questions, to move
past counting things to understanding
them”

6. The quality of results from
data science projects is
directly related to the extent
that they reflect important
properties of real-world
entities represented therein
“Foundational ontologies are domain-
independent axiomatic systems of
categories and their ties (e.g. objects,
events, causality, parthood, spatial-
temporal connections, dependencies,
etc) that can be used to articulate the
representation of phenomena in
different material domains”
[Guizzardi, 2007]

7. Claim #1
Data Science should leverage real-world
semantics throughout its lifecycle

8. During data pre-processing
(Data integration)
•ontology design patterns (ODP) well founded on foundational
ontologies may help to identify and/or discard potential alignments
between data from different repositories which presumably refer to
the same real-world entity" (Padilha et al., 2012).
Padilha, N. F., Baião, F. and Revoredo, K. (2012) Ontology alignment for semantic data integration through foundational ontolo- gies.
In International Conference on Conceptual Modeling, 172–181. Springer.

9. During Mining
(classification)
•being aware whether a to-be learned category is a Mixin (spanning multiple
kinds, instead of a specific kind in particular) raises issues for feature
selection, for example
•the use of well-founded ontological design patterns (ODP) may help to either
choose appropriate class labels, select relevant features or improve
performance of classification algorithms
Amaral, G. and Guizzardi, G. (2019) On the application of ontological patterns for conceptual modeling in multidimensional
models. In European Conference on Advances in Databases and Information Systems, 215–231. Springer.

10. During Mining
(clustering)
•quality reifications (JP’keynote) provides a way to cope with different
measurement schemes for qualities
•should be explicitly considered, to avoid leading to misinterpretation of the results from
mining tasks – diffferent values do not always imply different things in reality….
•foundational ontologies (and ontological analysis) serve as a fundamental
support for establishing grouping criteria and similarity calculation, reducing
the possibility of creating groups of objects that do not reflect genuine real-
world regularities.
•identifying similarities in clustering process that are not merely accidental
•leverage ontological distinctions (objects versus events, dependent versus independent
entities, kinds versus roles, …)

11. Claim #1
Data Science should leverage real-world semantics
throughout its lifecycle
•how to “reverse engineer” the real-world semantics of all the
data that is already available?
•how to make foundational ontologies more available for
domain modelers?
•Ongoing work with Glenda Amaral & Giancarlo Guizzardi…

13. Understanding the dynamic nature of the COVID-19
pandemic: An ontological Analysis of Control Measures
•Ongoing work with Glenda Amaral, Cristine Griffo & Giancarlo Guizzardi…

14. Fernanda Baião
[email protected]