A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is
A single abstract from the DHd-2016 Book of Abstracts.doi:10.5281/zenodo.4645167 fatcat:755ubzzbfnembhvggx3wb2hswm
Zur Durchführung unserer Experimente nutzen wir JeSemE 2 (Hellrich u. Hahn 2017b) . ... Grund für diese Entscheidung ist dessen hohe Reliabilität, wohingegen mit word2vec durchgeführte Experimente nur bedingt wiederholbar sind (Hellrich u. Hahn 2017a) . ...doi:10.5281/zenodo.4622410 fatcat:hc4qybsqfvhptdvxool4ktudni
Diachronic research focuses on the English Fiction part, with the exception of some work relating to German data (Hellrich and Hahn, 2016b) . ... SGNS is preferred in general, yet SGHS showed slight benefits in some reliability scenarios in our prior investigations (Hellrich and Hahn, 2016a) . ...dblp:conf/coling/HellrichH16 fatcat:x5aoeihwpnb7xnxxppooaehva4
(Hellrich and Hahn, 2017) . ... qualitative conclusions since they typically involve random processes that severely limit the reliability of resultsrepeated experiments differ in which words are deemed most similar with each other (Hellrich ...dblp:conf/dihu/HellrichH17 fatcat:cjfmpvmynvc5nkvnekcsvj3auy
A single abstract from the DHd-2017 Book of Abstracts.doi:10.5281/zenodo.4622679 fatcat:pmizjbha3rffpkfowlc5brfvpq
Its development owes credits to the interdisciplinary Graduate School "The Romantic Model" at Friedrich-Schiller-Universität Jena (Germany). 1 http://jeseme.org 2 https://github.com/hellrich/JeSemE 2 Related ... Most work is centered around word2vec (e.g., Kim et al. (2014) ; Kulkarni et al. (2015) ; Hellrich and Hahn (2016b)), whereas alternative approaches are rare, e.g., Jo (2016) using GloVe (Pennington ...doi:10.18653/v1/p17-4006 dblp:conf/acl/HellrichH17 fatcat:a63lxgvnjjdynjqltnheok5wnq
Today's widely used annotation tools were designed for annotating typically short textual mentions of entities or relations, making their interface cumbersome to use for long(er) stretches of text, e.g, sentences running over several lines in a document. They also lack systematic support for hierarchically structured labels, i.e., one label being conceptually more general than another (e.g., anamnesis in relation to family anamnesis). Moreover, as a more fundamental shortcoming of today'sdoi:10.18653/v1/w19-4025 dblp:conf/acllaw/LohrKLHKSH19 fatcat:3k22cqrujjendehqiykjvfapc4
more »... they provide no continuous quality control mechanisms for the annotation process, an essential feature to intrinsically support iterative cycles in the development of annotation guidelines. We alleviated these problems by developing WAT-SL 2.0, an open-source web-based annotation tool for long-segment labeling, hierarchically structured label sets and built-ins for quality control.
used point-wise mutual information with singular value decomposition (Levy et al, 2015; SVDPPMI) to measure word similarity, since it turned out to be superior for DH applications in previous work (Hellrich ...dblp:conf/dihu/BuechelHH17 fatcat:oxcrxabobrd5le7k7hby4fgefe
Hellrich et al. / Sharing Models and Tools for Processing German Clinical Texts ... Hellrich et al. / Sharing Models and Tools for Processing German Clinical Texts Table 1. ...pmid:25991250 fatcat:n76hyqwexvejbkd2nerpx6u3gu
Stability is critical for studies examining the underlying semantic space as a more advanced form of corpus linguistics, e.g., tracking lexical change (Kim et al., 2014; Kulkarni et al., 2015; Hellrich ... We show how the choice of down-sampling strategies, a seemingly minor detail, leads to major differences in the characterization of SVD PPMI in recent studies (Hellrich and Hahn, 2017; Antoniak and Mimno ...arXiv:1808.06810v2 fatcat:jmms3ulkzffzld6vgtkd372bnq
We here describe the set-up for the system from the Jena University Language & Information Engineering (JULIE) Lab which participated in the CLEF-ER 2013 Challenge. The task of this challenge was to identify hitherto unknown translation equivalents for biomedical terms from several parallel text corpora. The languages being covered are English, German, French, Spanish and Dutch. Our translation system enhanced, in a realistic scenario, the French, German, Spanish and Dutch parts of adblp:conf/clef/HellrichH13 fatcat:wzp7rdesfvbvpiuym2ialndz2y
more »... d terminology with 4k to 15k new entries each. Based on expert assessment of the new German translations about 76% of these were judged as plausible term enhancements.
An earlier state of our own work was decribed by Hellrich & Hahn. 22 This contribution supersedes that former version in numerous technical and methodological respects. ...pmid:25954371 pmcid:PMC4419887 fatcat:kt7x4paa6vdefj5vdqf2qcr47y
We devised annotation guidelines for the de-identification of German clinical documents and assembled a corpus of 1,106 discharge summaries and transfer letters with 44K annotated protected health information (PHI) items. After three iteration rounds, our annotation team finally reached an inter-annotator agreement of 0.96 on the instance level and 0.97 on the token level of annotation (averaged pair-wise F1 score). To establish a baseline for automatic de-identification on our corpus, wedoi:10.3233/shti190212 pmid:31437914 fatcat:qxfits7s3jfavnuyvpvkqadp3e
more »... d a recurrent neural network (RNN) and achieved F1 scores greater than 0.9 on most major PHI categories.
We here describe the evolution of annotation guidelines for major clinical named entities, namely Diagnosis, Findings and Symptoms, on a corpus of approximately 1,000 German discharge letters. Due to their intrinsic opaqueness and complexity, clinical annotation tasks require continuous guideline tuning, beginning from the initial definition of crucial entities and the subsequent iterative evolution of guidelines based on empirical evidence. We describe rationales for adaptation, with focus ondoi:10.3233/shti200116 pmid:32570340 fatcat:5pbaysmnszc7fnhn437y5gvzsq
more »... everal metrical criteria and task-centered clinical constraints.
The first release of JESEME (Hellrich and Hahn, 2017b) mainly provides timevariant diachronic lexical semantic information. ... Compared to Hellrich and Hahn (2017b) , our current version also reduces the database size from approximately 120GB to 40GB. ...arXiv:1807.04148v2 fatcat:he45emfaw5fivonydaz3ktfa7q
« Previous Showing results 1 — 15 out of 34 results