539 research outputs found
Sort by
Annotation guidelines for German verbal synonyms included in SynSemClass Lexicon
No full textThis report presents a guideline for including German synonyms into the multilingual SynSemClass lexicon
SynSemClass 3.5
No full textThe SynSemClass synonym verb lexicon version 3.5 investigates, with respect to contextually-based verb synonymy, semantic ‘equivalence’ of Czech, English and German verb senses and their valency behavior in parallel Czech-English and German-English language resources.
SynSemClass3.5 is a multilingual event-type ontology based on classes of synonymous verb senses, complemented with semantic roles and links to existing semantic lexicons.
Apart of the already used links to PDT-Vallex, EngVallex, CzEngVallex, FrameNet, VerbNet, PropBank, Ontonotes, and English WordNet for Czech and English entries the new links to German language lexical resources are exploited for German verb entries, such as Woxikon, E-VALBU, and GUP.
The German part of the lexicon has been created within the project Multilingual Event-Type-Anchored Ontology for Natural Language Understanding (META-O-NLU) by two cooperating teams - by the team of the Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics, Prague (ÚFAL), Czech Republic and the team of the German Research Center for Artificial Intelligence (DFKI) Speech and Language Technology, Berlin, Germany
CorefUD 0.1
No full textCorefUD is a collection of previously existing datasets annotated with coreference, which we converted into a common annotation scheme. In total, CorefUD in its current version 0.1 consists of 17 datasets for 11 languages
Do UD Trees Match Mention Spans in Coreference Annotations?
No full textOne can find dozens of data resources for various languages in which coreference – a relation between two or more expressions that refer to the same real-world entity – is manually annotated. One could also assume that such expressions usually constitute syntactically meaningful units; however, mention spans have been annotated simply by delimiting token intervals in most coreference projects, i.e., independently of any syntactic representation. We argue that it could be advantageous to make syntactic and coreference annotations convergent in the long term. We present a pilot empirical study focused on matches and mismatches between hand-annotated linear mention spans and automatically parsed syntactic trees that follow Universal Dependencies conventions. 8 datasets for 7 different languages are included in the study
Deliverable D7.2 Report on NLP Technologies Workshop at EUROSAI Congress
No full textThis deliverable reports on the preparation of the LangTools workshop at EUROSAI Congress 2020, aimed at presenting NLP Technologies to supreme audit institution (SAI) representatives
CUNI Neural ASR with Phoneme-Level Intermediate Step for Non-Native SLT at IWSLT 2020
No full textIn this paper, we present our submission to
the Non-Native Speech Translation Task for
IWSLT 2020. Our main contribution is a proposed speech recognition pipeline that consists of an acoustic model and a phoneme-to grapheme model. As an intermediate representation, we utilize phonemes. We demonstrate that the proposed pipeline surpasses
commercially used automatic speech recognition (ASR) and submit it into the ASR track.
We complement this ASR with off-the-shelf
MT systems to take part also in the speech
translation track
COSTRA 1.0: A Dataset of Complex Sentence Transformations
No full textWe present COSTRA 1.0, a dataset of complex sentence transformations. The dataset is intended for the study of sentence-level embeddings beyond simple word alternations or standard paraphrasing. This first version of the dataset is limited to sentences in Czech but the construction method is universal and we plan to use it also for other languages.
The dataset consists of 4,262 unique sentences with an average length of 10 words, illustrating 15 types of modifications, such as simplification, generalization, or formal and informal language variation. The hope is that with this dataset, we should be able to test semantic properties of sentence embeddings and perhaps even to find some topologically interesting ''skeleton'' in the sentence embedding space. A preliminary analysis using LASER, multi-purpose multi-lingual sentence embeddings suggests that the LASER space does not exhibit the desired propertie
Outbound Translation User Interface Ptakopet: A Pilot Study
No full textIt is not uncommon for Internet users to have to produce a text in a foreign language they have very little knowledge of and areunable to verify the translation quality.We call the task “outbound translation” and explore it by introducing an open-sourcemodular system Ptakopˇet. Its main purpose is to inspect human interaction with MT systems enhanced with additional subsystems,such as backward translation and quality estimation. We follow up with an experiment on (Czech) human annotators tasked toproduce questions in a language they do not speak (German), with the help of Ptakopˇet. We focus on three real-world use cases(communication with IT support, describing administrative issues and asking encyclopedic questions) from which we gain insight intodifferent strategies users take when faced with outbound translation tasks. Round trip translation is known to be unreliable for evaluat-ing MT systems but our experimental evaluation documents that it works very well for users, at least on MT systems of mid-range quality
Human or Machine: Automating Human Likeliness Evaluation of NLG Texts
No full textAutomatické vyhodnocení různých kritérií kvality textu vytvořených inteligentními metodami založenými na datech je velmi běžné a užitečné, protože je levné, rychlé a obvykle přináší opakovatelné výsledky. V tomto příspěvku prezentujeme pokus automatizovat hodnocení lidské pravděpodobnosti výstupních textových vzorků pocházejících z metod generování přirozeného jazyka používaných k řešení několika úkolů
Findings of the 2020 Conference on Machine Translation (WMT20)
No full textThis paper presents the results of the news translation task and the similar language translation task, both organised
alongside the Conference on Machine Translation (WMT) 2020. In the news task, participants
were asked to build machine translation systems for any of 11 language pairs, to be evaluated on test sets consisting mainly of news stories. The task was also opened up
to additional test suites to probe specific aspects of translation. In the similar language translation task, participants built
machine translation systems for translating between closely related pairs of languages