Filters








6 Hits in 3.2 sec

INEX Tweet Contextualization task: Evaluation, results and lesson learned

Patrice Bellot, Véronique Moriceau, Josiane Mothe, Eric SanJuan, Xavier Tannier
2016 Information Processing & Management  
This motivated the proposal of a new track at CLEF INEX lab of Tweet Contextualization.  ...  The evaluation measures for automatic summarization designed in DUC or MUC were not appropriate to evaluate tweet contextualization, we explain why and depict in detailed the LogSim measure used to evaluate  ...  Among these few, there is the IRIT [32] system that reached best readability scores in Table 7 and also in 2014 as reported in [9] .  ... 
doi:10.1016/j.ipm.2016.03.002 fatcat:7h64buwylvhefbcarhxeusjd5q

IRIT at TREC Temporal Summarization 2015

Rafik Abbes, Bilel Moulahi, Abdelhamid Chellal, Karen Pinel-Sauvagnat, Nathalie Hernandez, Mohand Boughanem, Lynda Tamine, Sadok Yahia
unpublished
This paper describes the IRIT lab participation to the TREC 2015 Temporal Summarization track.  ...  The goal of the Temporal Summarization track is to develop systems that allow users to efficiently monitor information about events over time.  ...  K f use is tuned using the TREC 2014 TS track data, and is set to 30. The rank fusion model parameters σ and ϵ are set to 300 and 10, respectively.  ... 
fatcat:7mmwsf4fyrdw3mrdz7nb4ccovu

Open Archive TOULOUSE Archive Ouverte (OATAO)

Moulahi, Chellal, Abdelhamid, Pinel-Sauvagnat, Karen And Hernandez, Nathalie, Boughanem, Tamine, Lynda, Ben Yahia
2015 Sadok IRIT at TREC Temporal Summarization 2015. (2015) In: Text REtrieval Conference (TREC 2015)   unpublished
This is an author-deposited version published in : http://oatao.univ-toulouse.fr/ Eprints ID : 15470 The contribution was presented at :  ...  K f use is tuned using the TREC 2014 TS track data, and is set to 30. The rank fusion model parameters σ and ǫ are set to 300 and 10, respectively.  ...  These values were learned according to experiments carried out on 2014 TREC TS filtered dataset. The parameters of each submitted run are shown in table 2.  ... 
fatcat:ptvvq3jlbvasjgzvffsmj2zt5m

Shallow features as indicators of English–German contrasts in lexical cohesion

Kerstin Kunz, Ekaterina Lapshinova-Koltunski, José Manuel Martínez Martínez, Katrin Menzel, Erich Steiner
2017 Languages in Contrast: International Journal for Contrastive Linguistics  
We avoided the distinction, at the syntactic level, between Complement and Adjunct.  ...  The annotation of primary versus marked variation has been evaluated at 9.5 K-Cohen agreement (Gagliardi 2014).  ...  Social media messages also have at least a creation time as temporal context. This implicit spatio-temporal (ST) metadata is not currently heavily exploited by modern NLP methods.  ... 
doi:10.1075/lic.16005.kun fatcat:p7pa2crm6rgkrfmsdf7m5265ea

Salience Estimation and Faithful Generation: Modeling Methods for Text Summarization and Generation

Christopher Kedzie
2021
At a high level, we divide these tasks into two categories: content selection, or "what to say" and content realization, or "how to say it" (McKeown, 1985).  ...  Somewhat against the prevailing trends, we eschew end-to-end training of an abstractive summarization model, and instead break down the text summarization problem into its constituent tasks.  ...  Frey and Dueck variations to the 2014 TREC Temporal Summarization shared-task (Aslam et al., 2015).  ... 
doi:10.7916/d8-61n8-mg23 fatcat:sztaye4ftvhubjibjd3ty5dqy4

EMBnet.journal 18 Suppl. B

EMBnet Journal
2012 EMBnet journal  
Segagni et al, presented at the DILS 2012 conference in Washington DC.  ...  The topology summarizes the relationships among the organisms, while branch lengths summarize the expected changes along a given section of them (Felsenstein, 2004) .  ...  We also aim to provide a building block in an analysis pipeline that can be used to look at temporal reconstruction problems that assume an already (partially) ordered dataset (Ramakrishnan, 2010; .  ... 
doi:10.14806/ej.18.b.592 fatcat:wlwsmbdlfzbjtk7vyhiabdov6q