Project wasannounced at a pressconference of the St. Petersburg RegionalInformation Center TASS.Nikolai Tesla, SeniorResearcher of the Laboratory of Integrated AutomationSystems at St. PetersburgFederal Research Center of the Russian Academy of Sciences (SPC RAS) spokeabout the development of solutions for the project “Pushkin Digital” based on artificial intelligence.
On May 23, other team members of the project “PushkinDigital” also spokeat the press conference:Valentin Golovin,Director of the Institute of Russian Literature(Pushkin House) of the Russian Academy of Sciences, Svetlana Nikolova, Deputy Director, curator of the project, Gavriil Belyak,Rsearcher at the Pushkin House, project participant, and Polina Kolozaridi, Head of the Educational Program"Digital Methods inHumanitarian Studies" at ITMO University. Theytalked about the projectobjectives and implementation stages, about developing a resource that concerns the literary heritage of AlexanderPushkin, and itscapacities.
Alexander Pushkin (1799-1837) left behinda hugebody ofwritten works, aswell as drafts, letters,books, notes andother materials.Researchers of IRLI RAS, who have an access to a significantpart of Pushkin's heritage,initiated the interdisciplinary project "PushkinDigital" – a multimediaresource about the lifeand work of this prominent author. Its objectiveis to create a new toolkit intended for transferring humanitarian knowledge, since the traditional forms of representation(say, a collection of works), although they have not lost theirscientific value, donot use the potential of modern digitalmeans of communication.
“Starting this fall the “Pushkin Digital”portal will be available.We assumethat October19th will be the right launch point. And that the work on the resource will not end: it will be further continued to get this resource replenished and complicated.In turn,users will have a chance to plunge into Pushkin's manuscripts, observe thecreative process anddeeper comprehend it, and eventually produce their own interpretations,” explained Valentin Golovin.
As Svetlana Nikolovaspecified at a pressconference that the resource’s basis was composed out of digitized materialsfrom the IRLI RAS funds,combined into a singledatabase. In particular,today over15 thousandmanuscripts of Alexander Pushkin have been digitized along with almost 2million pages of commentsand scientificliterature. In addition, about20 interactivecommentary plots have been developed for the educational part of the project.
“The “Pushkin Digital” project is aimed at solvingtwo maintasks: scientific anduser-oriented,” markedSvetlana Nikolova. According to her, the first one is being solved because the researcherswith the help of the resource developed can quickly access the entiredata array at once, avoid errorsand informationlosses.
“The user effect is considered through an attempt to surmount the “loss of scientific knowledge”, when what comes out of the pen of scientists either does not reach or does not draw the interest of a new generation. To compensate for this effect, we have provided formats for submitting data under the scientific and popular science angles to attract a wider audience. So that even non-professional can bridge the gap between simply knowing Pushkin's last name, to learning more about his personality and work, and maybe even become interested in philology in the future,” Svetlana Nikolova added.
It is important to note that the Pushkin Digital resource rather offers the development of a complex and extensive structure for presenting information about works whose texts are linked by common elements within the entire data set, than simply creating a digital online library where any reader can quickly find the necessary information by keywords. In order to bring the solution to this problem to life, the researchers of the National Research Center of the Russian Academy of Sciences have arrived to a decision to join the project
“Colleagues from the Pushkin Houseestimated that a simple reading of the entire volume of Pushkin's heritage materials would require about 58 years. It would take even longertime toanalyze them manually.However, if we applymodern text processingmethods and methods of search for links betweentexts tothis dataarray, then the taskfits into a significantly shorter period.Since there existed alot of heterogeneous materials, building a common model turned into a major challenge. The model combinedboth the initial understanding of Pushkin's texts and the possibility of building newlinks betweenthese materials.Moreover, not onlysimple links fromone workto another,but also the allocation of new correspondencesthat hadnot been thought about before, had not been noticed or putaside,” Nikolai Teslya said at a press conference.
To achieve thisgoal, the SPC RAS researchers have developed a specialneural network model thatsearches for common entitiesin the materials of a large-scale arraydigitized by researchers fromIRLI RAS.
Then Polina Kolozariditold about the main goals of the visual side of the “Pushkin Digital” resource. “ITMO University is developing an interface, that is, how the information essences of Pushkin's heritage will be implementedthere. After all, it is important notonly to compress the58 yearsnecessary to comprehendPushkin's heritage, as colleagues have spoken about, butalso to allow for convenientand efficient working with this material,” she noted.
A recording of the press conference is available by the link по ссылке.