Presenting advanced access related research in PERICLES at QI-14
The concept of advanced access combines automatic indexing, automatic document categorization, document retrieval and information visualization in digital libraries, and is crucial to any progress in digital preservation (DP). Namely, what would be the point of preserving anything for the future if one had no access to safeguarded content?
Lately the temporal aspect of content, its ageing, composition and dynamics, has become a matter of study in several fields, all of which have a lot to learn from physics, the heartland of such considerations. Further, increasing scholarly interest is being paid to the insight that quantum theory – practically quantum mechanics stripped of its physical content and focusing on uncertainty in human decision-making – could be relevant to the dynamics of macroscale systems as well, much beyond the subatomic world of Niels Bohr, Werner Heisenberg and Erwin Schrödinger. This has brought in focus application areas such as concept theory, economics, information retrieval, linguistics and organizational dynamics under the name of quantum interaction, the new testing ground for such ideas.
Quantum interaction uses techniques and models developed in quantum physics and applies them in new areas of informatics. We are interested in the question of how they can be applied to access to semantic content in digital objects, and if parallel research trends in the field justify our own approach. Therefore related work done in PERICLES was presented by University of Borås (HB) representative Sándor Darányi at the 8th International Conference on Quantum Interaction (QI-14) held in Filzbach, Switzerland, from June 30 to July 3, 2014. QI-14 followed previous successful meetings at Stanford University; Oxford University; the German Research Center for Artificial Intelligence in Saarbrücken; the AAAI Symposium in Washington DC; Robert Gordon University in Aberdeen; the Paris School of Economics; and the University of Leicester.
At this point it is important to recall that over the past 20 years, the role of context for the interpretation of semantic content has been emphasized in many scientific disciplines relevant to DP, and has become central to PERICLES too. And, as it happens, contextuality is also the central notion of the QI conferences. So the relevance of having presented our own solutions at QI-14 was threefold. We could learn about new methods on how to handle context and its semantics-modifying effects; receive critical feedback about the combination of ideas tested; and, last but not least, link in DP as a new and important research area for large-scale experimentation with evolving content in this scholarly community.
Of course the big research question is, does evolving semantics of digital collections show quantum-like behaviour over time, or rather, one fully described by classical physics? How to design measurements so as to falsify one of these assumptions? First answers related to space and digital media data investigated by PERICLES are expected next year. For now, using a software tool called Somoclu, we modelled the semantic content of words in a text collection as an evolving vector field (see below). The white dots indicate index terms, the black zones are inactive, stable semantic areas, whereas the red hot zone – just like in a lava field – indicates underlying tensions going back to language use. In fact one can expect new index terms to appear in this area, with the red and yellow boundaries being those fault lines where “content tectonics” will push apart existing content and allow for protruding new content to modify the semantic landscape.
Interesting related readings:
For a quick background, popular science articles about the interactions between quantum theory and wider sciences include the following stories in New Scientist: