Profiling information sources and services for discovery

Profiling information sources and services for discoveryMathieu d’Aquin - @mdaquinInsight Centre - NUI Galway

Let’s start with a quote

“access billions of heterogeneous data sources, stored or real-time, locally or globally, from personal and public terminals, mobile, wearable, Internet of Things, and cyber-physical systems [...] present relevant information services in a meaningful and personalised way”

-- Fahim Kawsar, Christophe Diot, Carles Sierra and Nozha Boujemaa, Discovery & Identification Technologies for the Next Generation Internet

Let’s start with a quote

“access billions of heterogeneous data sources, stored or real-time, locally or globally, from personal and public terminals, mobile, wearable, Internet of Things, and cyber-physical systems [...] present relevant information services in a meaningful and personalised way”

-- Fahim Kawsar, Christophe Diot, Carles Sierra and Nozha Boujemaa, Discovery & Identification Technologies for the Next Generation Internet

More concretely: A scenario

Let’s think about an autonomous, physical mobile agents, a robot, coming into a new environment (building, city, organisation, event) where data sources are available:

1. How does it find the data sources relevant to its task?

2. How does it use the data sources relevant to its task?

3. How does it contribute to data sources relevant to its task?

Metadata and content

Discovery currently handled through data cataloguing.

Should focus on the content of data: In which context is it valid, and what can be done with it?

Example: What question can it answer?

Example: What question can it answer?

Service code

Area

Restaurant

Organisation

isa

population

deprivation

locatedIn

rating

employee

The population of Walnut Tree is 4096

What is the population of Walnut Tree?

Example: How good is it? Does it fit what I know?

Consensus(agreement)

Controversy

Disagreement (negative consensus)

Example: Do I actually have the right to use it?

Conclusion

With information sources expanding in all directions, we need mechanisms by which their consumption and use can emerge dynamically through autonomous agents being able to make sense of new data spaces.

Requires profiling which is based on the “meta-analysis” of data, to extract:

- Scope, validity (in time, space, rights, etc.)- Competence, inter-relatedness- Views, biases, quality issues- Exploitability

@mdaquin - mdaquin.net