Filters








324 Hits in 4.2 sec

Seshat: A tool for managing and verifying annotation campaigns of audio data [article]

Hadrien Titeux
2021 arXiv   pre-print
In addition, it includes procedures for checking the content of annotations following specific rules that can be implemented in personalised parsers.  ...  Finally, we propose a double-annotation mode, for which Seshat computes automatically an associated inter-annotator agreement with the γ measure taking into account the categorisation and segmentation  ...  Campaign manager can customise the distance used by γ by inserting a custom distance along their own parser (See short snippet of code for a parser of French Phonetics with the SAMPA alphabet in Algorithm  ... 
arXiv:2003.01472v2 fatcat:j4niqtv3ebalzi5tpf6daygzva

Using Lexicon-Grammar Tables for French Verbs in a Large-Coverage Parser [chapter]

Elsa Tolone, Benoît Sagot
2011 Lecture Notes in Computer Science  
In this paper, we describe the integration of Lexicon-Grammar tables for French verbs in the large-coverage FRMG parser and the evaluation of the resulting parser.  ...  We compare the results of the FRMG parser on the EASy reference corpus depending on whether it relies on the verb entries of the Lefff or those of the converted Lexicon-Grammar verb tables.  ...  The relevance of the resulting lexicon is confirmed by its use for parsing the evaluation corpus of the French parsing evaluation campaign EASy. given for the whole EASy corpus and for only a sample of  ... 
doi:10.1007/978-3-642-20095-3_17 fatcat:345ngk2l65f6nhtolxluj55ybu

Semi-supervised SRL System with Bayesian Inference [chapter]

Alejandra Lorenzo, Christophe Cerisara
2014 Lecture Notes in Computer Science  
This approach is evaluated on French and English.  ...  In both cases, it achieves very good performance and outperforms a strong supervised baseline when only a small number of annotated sentences is available and even without using any previously trained  ...  Acknowledgments This work has been partially funded by the French ANR project ContNomina.  ... 
doi:10.1007/978-3-642-54906-9_35 fatcat:ti4aqhm7tzf4vauoniwutm5s3e

Overview of the SPMRL 2013 Shared Task: A Cross-Framework Evaluation of Parsing Morphologically Rich Languages

Djamé Seddah, Reut Tsarfaty, Sandra Kübler, Marie Candito, Jinho D. Choi, Richárd Farkas, Jennifer Foster, Iakes Goenaga, Koldo Gojenola Galletebeitia, Yoav Goldberg, Spence Green, Nizar Habash (+11 others)
2013 Workshop on Statistical Parsing of Morphologically Rich Languages  
The task features data sets from nine languages, each available both in constituency and dependency annotation.  ...  We report on the preparation of the data sets, on the proposed parsing scenarios, and on the evaluation metrics for parsing MRLs given different representation types.  ...  We thank Alon Itai and MILA, the knowledge center for processing Hebrew, for kindly making the Hebrew treebank and morphological analyzer available for us, Anne Abeillé for allowing us to use the French  ... 
dblp:conf/acl-spmrl/SeddahTKCCFFGGG13 fatcat:5ngltl7jr5hu5o72goudyy4tuq

A Graph-based Approach to Cross-language Multi-document Summarization

Florian Boudin, Stéphane Huet, Juan-Manuel Torres-Moreno
2011 POLIBITS Research Journal on Computer Science and Computer Engineering With Applications  
We evaluate our method on a manually translated subset of the DUC 2004 evaluation campaign.  ...  Cross-language summarization is the task of generating a summary in a language different from the language of the source documents.  ...  We performed both automatic evaluation of content and manual evaluation of readability on a subset of the DUC 2004 data set made of 16 randomly selected clusters. 1) Automatic Evaluation: The majority  ... 
doi:10.17562/pb-43-16 fatcat:6y4sfnmgwbbzfa2bujyfky4seu

Error mining in parsing results

Benoît Sagot, Éric de la Clergerie
2006 Proceedings of the 21st International Conference on Computational Linguistics and the 44th annual meeting of the ACL - ACL '06  
We introduce an error mining technique for automatically detecting errors in resources that are used in parsing systems.  ...  We were thus able to identify missing and erroneous information in these resources.  ...  EASy corpus : This is the 40,000-sentence corpus that has been built for the EASy parsing evaluation campaign for French (Paroubek et al., 2005) .  ... 
doi:10.3115/1220175.1220217 dblp:conf/acl/SagotC06 fatcat:hcveb2tqx5b3fbatxfjtjo4m3a

CreatingZombilingo, a game with a purpose for dependency syntax annotation

Karën Fort, Bruno Guillaume, Hadrien Chastant
2014 Proceedings of the First International Workshop on Gamification for Information Retrieval - GamifIR '14  
This paper presents the design of Zombilingo, a Game With A Purpose (GWAP) that allows for the dependency syntax annotation of French corpora.  ...  The development will start mid-2014 and the game is to be made available by the end of the year. The created language resource will be freely and continuously available on the game Web site.  ...  Lafourcade for his inputs and feedback on the project.  ... 
doi:10.1145/2594776.2594777 dblp:conf/ecir/FortGC14 fatcat:ciz3d2axjnez3ipmhpfpxak5ma

Syntactic concordancing and multi-word expression detection

V. Seretan, E. Wehrli
2013 International Journal of Data Mining Modelling and Management  
We also provide relevant performance evaluation results for the main system components, focusing on the comparison between syntax-based and syntax-free approaches.  ...  Also called key word in context (KWIC), these tools are nowadays indispensable in the work of lexicographers, linguists, and translators.  ...  The authors would like to thank the three anonymous reviewers, whose comments and suggestions helped improve the article.  ... 
doi:10.1504/ijdmmm.2013.053694 fatcat:fqlfeitlzzcx7leqebca4idrlu

Optimality Theory as a Framework for Lexical Acquisition [chapter]

Thierry Poibeau
2014 Lecture Notes in Computer Science  
This paper re-investigates a lexical acquisition system initially developed for French.  ...  We show that, interestingly, the architecture of the system reproduces and implements the main components of Optimality Theory.  ...  It is a relatively accurate parser, e.g. it obtained the best precision and F-measure for written French text in the first EASY evaluation campaign (2006).  ... 
doi:10.1007/978-3-642-54906-9_2 fatcat:qbyuz54m3zfsxn6odoncgmxmqu

Optimality Theory as a Framework for Lexical Acquisition [article]

Thierry Poibeau
2014 arXiv   pre-print
This paper re-investigates a lexical acquisition system initially developed for French.We show that, interestingly, the architecture of the system reproduces and implements the main components of Optimality  ...  However, we formulate the hypothesis that some of its limitations are mainly due to a poor representation of the constraints used.  ...  It is a relatively accurate parser, e.g. it obtained the best precision and F-measure for written French text in the first EASY evaluation campaign (2006).  ... 
arXiv:1405.6682v1 fatcat:ro5x7lgu7zgefagxpz22ahkoli

Yaafe, An Easy To Use And Efficient Audio Feature Extraction Software

Benoît Mathieu, Slim Essid, Thomas Fillon, Jacques Prado, Gaël Richard
2010 Zenodo  
YAAFE has already been used in Quaero project internal evaluation campaigns for the music/speech discrimination and musical genre recognition tasks.  ...  Marsyas performance clearly suffers from writing 16 column of zeros. For the evaluated task, the CPU times in Table 1 show that YAAFE tends to be faster than Marsyas.  ... 
doi:10.5281/zenodo.1418320 fatcat:wlhttc62gjhhnf5fugunxxmche

Coupling an annotated corpus and a lexicon for state-of-the-art POS tagging

Pascal Denis, Benoît Sagot
2012 Language Resources and Evaluation  
We also conduct experiments on datasets and lexicons of varying sizes in order to assess the best trade-off between annotating data vs. developing a lexicon.  ...  We find that the use of a lexicon improves the quality of the tagger at any stage of development of either resource, and that for fixed performance levels the availability of the full lexicon consistently  ...  which have been evaluated during the GRACE evaluation campaign. 16 Although a direct comparison is difficult, given the differences in terms of reference corpus and tagsets, it is worth mentioning that  ... 
doi:10.1007/s10579-012-9193-0 fatcat:nlourtdczvd5xeseo3orztjfxq

Bibliography [chapter]

2016 Collaborative Annotation for Reliable Natural Language Processing  
[FOR 12c] FORT K., FRANÇOIS C., GALIBERT O. et al., "Analyzing the impact of prevalence on the evaluation of a manual annotation campaign", International Conference on Language Resources and Evaluation  ...  ., "Training and evaluation of POS taggers on the French MULTITAG corpus", Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08), Marrakech, Morocco, 28-30, European  ... 
doi:10.1002/9781119306696.biblio fatcat:vfya5l7vgbgnlhcktom5zoyudu

Disambiguating Discourse Connectives for Statistical Machine Translation

Thomas Meyer, Najeh Hajlaoui, Andrei Popescu-Belis
2015 IEEE/ACM Transactions on Audio Speech and Language Processing  
This improvement is demonstrated here when translating from English to four target languages -French, German, Italian and Arabic -using several test sets from recent MT evaluation campaigns.  ...  The translation of connectives is improved significantly, between 0.7% and 10% as measured with the dedicated ACT metric.  ...  ACKNOWLEDGMENTS The authors are grateful for the funding of this work to the Swiss National Science Foundation (SNSF) under the COMTIS and MODERN Sinergia Projects (CRSI22 127510 and CRSII2 147653, see  ... 
doi:10.1109/taslp.2015.2422576 fatcat:3k6j26yd3zfarju3nelavzrtwu

Performance Comparison of Bootstrapped Statistical Taggers on Urdu Tweets

Amber Baig, Mutee U Rahman, Sehrish Abrejo, Khalid H Mohamadani, Ahsanullah Baloch
2021 International Journal of Scientific and Research Publications (IJSRP)  
At the end of each iteration, the performance of taggers was evaluated against the development set and automatically tagged, manually corrected 100 tweets were added in the training set to retrain both  ...  Finally, at the end of last iteration, tagger performance was evaluated against test set. Stanford tagger achieved an accuracy of 93.8% Precision, 92.9% Recall and 93.3% F-Measure.  ...  All faculty members of Department of Computer Science, Isra University are acknowledged for their help and support throughout the course of this study.  ... 
doi:10.29322/ijsrp.11.07.2021.p11559 fatcat:2awecd6mwzgt3clwkcd6ml65ku
« Previous Showing results 1 — 15 out of 324 results