CENTAL at SemEval-2016 Task 12: a linguistically fed CRF model for medical and temporal information extraction

Abstract

In this paper, we describe the system developed for our participation in the Clinical TempEval task of SemEval 2016 (task 12). Our team focused on the subtasks of span and attribute identification from raw text and proposed a system that integrates both statistical and linguistic approaches. Our system is based on Conditional Random Fields with high-precision linguistic features.

LIMSI-COT at SemEval-2017 Task 12: Neural Architecture for Temporal Information Extraction from Clinical Narratives

“…The task has been offered by SemEval over the past two years. Concerning the first group of subtasks, different approaches have been implemented by the participants including Conditional Random Fields (CRF) (AAl Abdulsalam et al, 2016;Caselli and Morante, 2016;Chikka, 2016;Cohan et al, 2016;Grouin and Moriceau, 2016;Hansart et al, 2016) and deep learning models (Fries, 2016;Chikka, 2016;Li and Huang, 2016). Similarly, CRF and neural networks models have been used for the second group of subtasks (AAl Abdulsalam et al, 2016;Cohan et al, 2016;Lee et al, 2016).…”

SemEval-2016 Task 12: Clinical TempEval

“…CDE-IIITH (Chikka, 2016) submitted 2 runs for each phase, the first based on deep learning models, and the second based on conditional random fields and support vector machines. Cental (Hansart et al, 2016) submitted 1 run for phase 1, based on conditional random fields and lexical resources. GUIR (Cohan et al, 2016) submitted 2 runs for phase 1 and 1 run for phase 2, based on conditional random fields and logistic regression with lexical, morphological, syntactic, dependency, and domain specific features, combined with pattern matching rules.…”

Automatic annotation of French medical narratives with SNOMED CT concepts

“…Then, the system looked up the dictionaries (both monolingual and bilingual) and proceeded to the SCT code attribution. Terms in medical terminologies can be affected by syntagmatic and paradigmatic variation to different degrees, or may be too precise or complex to actually be used in electronic health records [15]. By providing syntactic analysis and a proper recognition of collocations, the parser can detect concepts regardless of the specific morphological or syntactic form under which they appear in the text.…”