Uudiste arhiiv
28. jaanuari teisipäevaseminar
Teisipäeval kl 12.00 peab Kirjandusmuuseumi uues seminariruumis ettekande arvutiteadlane Mark Finlayson teemal "Inferring Propp's Functions from Semantically-Annotated Russian Folktales". Mark Finlayson (href="http://www.mit.edu/~markaf) on Massachusettsi Tehnoloogiainstituudi uurija, kes osales nädalavahetusel Tartus Wordneti konverentsil.
Seminar on inglise keeles. Kõik huvilised on väga oodatud!
Lühiülevaade ettekandest ja esinejast: Vladimir Propp's "Morphology of the Folktale" is a classic work in folkloristics, and has inspired much thought in other fields such as cognitive science and artificial intelligence. I describe Analogical Story Merging (ASM), a new machine learning algorithm that is a important step forward in learning Proppian-like morphologies from text. I first describe the data preparation phase of the technique, which involves annotating the "surface semantics" of a set of Russian folktales analyzed in Propp's work. I will present the Story Workbench, a general text annotation tool that can be used by annotators to annotate 20+ layers of syntax and semantics on top of text. I next discuss creation of a gold standard annotation of Propp's system on the Russian folktale texts, and show, for the first time, that Propp's system can actually be reliably recreated by trained annotators. Finally, I show how well the structures learned by the algorithm match with Propp's, and discuss directions for improvement and future research, including automatically extracting morphologies for many different cultures.
Dr. Mark Finlayson is a Research Scientist at the Computer Science and Artificial Intelligence Laboratory at MIT. His research focuses on representing, extracting, and using higher-order semantic patterns in natural language, especially with regard to narrative. He received the B.S.E from the University of Michigan in 1998, and the M.S. and Ph.D. from MIT in 2001 and 2012, respectively, all in Electrical Engineering and Computer Science. He is general chair of the Computational Models of Narrative (CMN) Workshop series, now approaching its fifth meeting, and is lead guest editor of a special issue on Computational Models of Narrative, to be published by the Journal of Literary & Linguistic Computing in 2014.
