Semantics of Empire: Machine Translation of Ottoman Turkish into English
Tekgürler, Merve
Stanford University, United States of America
HTML XMLThis paper focuses on Ottoman Turkish-English machine translation to create first-pass translations of primary sources. Translation allows incorporating non-English historical sources in education, thereby rendering the histories of underrepresented groups visible. We trained a neural machine translation model and tested large languages models for this task.
Manhattan, Euclidean, and their Siblings: Exploring Exotic Similarity Measures in Text Classification
Eder, Maciej (1); Ochab, Jeremi (2)
1: Institute of Polish Language (Polish Academy of Sciences), Poland; 2: Jagiellonian University, Poland
HTML XMLIn this paper, we empirically assess several similarity measures that share mathematical properties with two classical distance metrics, namely Manhattan and Euclidean. In a systematic experiment, we discover the relation between a given distance and the number of input features.
A Home without a “House”: Modelling Domestic Space in Nineteenth-Century British Fiction
Sherman, Alexander (1); Guhr, Svenja (2); Monaco, Jessica (1); Warner, Matt (1); Lamar, Annie K. (1); Algee-Hewitt, Mark (1)
1: Stanford University, United States of America; 2: Technical University of Darmstadt
HTML XMLWe introduce a trained BERT classifier that can detect which passages from Victorian novels are set in domestic spaces, even absent explicit spatial references. Our model helps understand how colonial, classed, and gendered discourses of domesticity are tied to literary space, as well as the literary history of space.