|
Three representative scenarios illustrate the use of Papyrus; a scholar with an interest in the internationalization of technological infrastructures in Europe, a group of university students preparing a paper on the history of artificial intelligence, and a journalist seeking to contextualize her coverage of European biomedical technology and science policy.
Matteo Mancini is a professor in the Department of Philosophy and History of Science. He is currently working on a paper related to the internationalization of the technological infrastructures in the Universita di Bologna. He is deeply interested in the history of energy and has a strong concern on communication infrastructure. Although he has a set of sources he usually consults (as library books and magazines) to find the latest historical facts, the most natural and immediate source for such information is news.
For example, Matteo was recently looking for information on automobile transportation. By logging into the Papyrus portal he can carry out his search in the following easy steps, without even knowing which news archives are consulted and in which way:
1. Query Formulation. He provides his research question to Papyrus or alternatively provides keywords he believes are relevant, as well as a limiting time period for the results, if necessary.
2. Query Analysis and Enrichment. His query is directly analyzed. The Papyrus historical domain ontology is used to precisely define the historical context of the query and enrich it with related terms (synonyms, hypernyms, hyponyms, and related concepts). This enrichment process will be valuable in the context of archival material, as terminology changes significantly with the passage of time. For example the word “computer” was used to refer to human workers before the emergence of electronic devices!
3. Query Fine Tuning. When the query has been analyzed and enriched, Matteo may review it. He has the option of viewing his initial query along with the related terms in a schematic manner or having it presented in a more textual form. If necessary, Matteo may add or remove concepts to this model or leave it as it is and proceed to the next step.
4. Query Execution. At this stage, the enriched query is used by Papyrus to locate relevant documents in all available news archives by automatically ‘translating’ the query, based on the corresponding concepts of the news domain ontologies. These documents may be textual but can also include multimedia elements, as images or short video, which have been indexed off line following the Papyrus news domain ontology. The Papyrus indexing process, although being automatic, has proved to be quite accurate as it makes use of the news content annotations which have been manually added by the content creators in an unstructured way. This way Papyrus manages to maximize the precision as well as the recall of the results.
5. Result Presentation. After the execution of the query, Matteo has several options at his disposal. The simple one is to view the resulting documents one by one, as a ranked list with the relevant sections highlighted. He may also view the results grouped into clusters, according to the query interpretation. Furthermore, the Papyrus advanced text analysis and extraction methods can also produce a synthesis of the relevant results into one document. This will offer Matteo an overview of the relevant news content in a single document where only the relevant sections of the news items have been preserved and grouped into hyperlinked clusters of chronologically ordered material. Moreover, the information extraction may result in new instances added to the ontology, enriching thus the initial knowledge model. If Matteo wishes to investigate more into the same subject, he may return to “Step 3: Query Fine Tuning” and further filter the query.
6. Result Storage. After each research, Matteo has the option to save his query, the results (list of the document titles, source, etc) or the synthesized document along with its’ semantic annotation in his personal archive. He may also execute the same query at a later date and observe the differences in the results, which will be highlighted by Papyrus together with the causes for these differentiations (new content that appeared in the archive sources, evolution of the domain ontologies or modifications in the reasoning models.
The use of Papyrus provides Matteo with an efficient and dynamic way of checking and re-contextualizing his work by utilizing a seamless access to news archives. Matteo can constantly enrich his research on automobile transportation by studying in parallel the results on other methods of transportation, stored in the ontology and enriched with further queries on the archive material. Instead of offering, simply, a set of results, the Papyrus library entries also take the form of an annotated narrative, easily read or copied for possible use in a draft research paper.
By following the aforementioned steps, Papyrus can also serve educational purposes. For example, students who are researching the historical evolution of technology will be given the means to compare current technologies to all past inventions and technological changes.
Our students are enrolled in the Department of Computer Engineering in Columbia University of New York, and currently work on an assignment in the history of artificial intelligence. Sammie, David, Linda, Nickolas, Anja and Karen, imaginative as they may be in choosing keywords for a web search, still have no way to reach decent secondary sources or even any primary source.
Papyrus in this case will be of much use. Anja proceeds by posing her query on “artificial intelligence” as a researcher would do. Sammie poses questions in his queries on “electronic brain issues” and Linda enriches Sammies’ questions by filling in historical dates. Papyrus processes the queries, translating them into the corresponding terminology commonly met in the news domain and returns a set of results, following the same steps as described above.
The students end up with a summary of the most related terminology found in news items along with an organized list of relevant historical events mined from distributed European news archives. They can view the firsthand sources by being directly linked to the news articles so that they can better understand how people perceived technological progress in their time, look at their expectations and the significance they attributed to new technologies.
Having examining the first output, Karen can use Papyrus to limit the number of the results by proposing expressions that could further orient the research. For example, she can relate “electronic brain” with the term “progress”. Additionally, David can form his personal digital library, which will be dynamically reconfigured as Papyrus approaches a deeper understanding of David’s intentions.
In whole, Papyrus not only will provide students with valuable information for their paper’ preparation, but will also offer them functional experience with primary sources and an easy and educative manner of becoming familiar with the terminology of the domain they are investigating.
Likewise, Papyrus can also be used by professional journalists, for example by Marie-Alice Fleuette. She covers rapidly evolving scientific fields for La Recherche, a Science and Technology magazine, addressed to the wider public. She covers a whole set of fields, periodically returning to them for publishing new stories.
Today Marie is working on an article that refers to stem-cell policy, a field involving technological and scientific events in the key area of biomedical technology and science.
Papyrus will allow her to produce digital library entries for each field, dynamically updated with new entries or with reconfiguration of old ones, based on new findings. The Papyrus ontology, will allow her to periodically check not only a much broader number of fields but also their interrelationship.
Using Papyrus and automatically producing a draft story is very important to a busy journalist like Marie. Hitting article deadlines is imperative, but effective gathering of all her necessary information through Papyrus is a mighty task.
|