The European Parliament Interpreting Corpus (EPIC): implementation and developments
Bibliographic record
Abstract
The call for the creation of corpora in Interpreting Studies that could be queried by means of Corpus Linguistics tools was first made by Shlesinger (1998) over a decade ago. However, only recently has this need started to be met. The European Parliament Interpreting Corpus (EPIC) is one of the first machine-readable corpora to be openly accessible in the field of Interpreting Studies. It was created in 2004/2006 by the Directionality Research Group of the University of Bologna at Forlì, and consists of 9 sub-corpora in total: three sub-corpora of source language speeches (Italian, English and Spanish) and six sub-corpora of simultaneously interpreted speeches, thus comprising all possible directions and combinations of the three languages involved (Monti et al. 2005, Sandrelli et al.. 2010). At present, the corpus includes only a small part of all the recorded material, which is stored in the EPIC Multimedia Archive. \nThe present paper describes the steps undertaken to create the corpus and the ongoing developments to further expand it and improve its structure. Firstly, the methodology used for user-friendly data collection and transcription and for the part-of-speech (POS) tagging and lemmatisation of this open corpus will be described; then, the web-interface developed to carry out simple and advanced queries on-line will be illustrated (see http://sslmitdev-online.sslmit.unibo.it/corpora/corporaproject.php?path=E.P.I.C.). Examples of the corpus-based studies carried out so far will be provided (Russo et al 2006, Bendazzoli et al 2011) and a special emphasis will be placed on the great potential of EPIC as a pedagogical and research tool in interpreter training. Interpreting students can transcribe and analyse part of the recorded material stored in the EPIC Multimedia Archive in their graduation dissertations, thus taking advantage of a unique opportunity to reflect upon real-life professional interpreting performances and upon their own learning process. Finally, ongoing developments and future steps will be discussed: text-to-sound and source text-to-target text alignment procedures are currently being tested, so as to make EPIC a more powerful resource to be explored by the interpreting research community \n \n \nReferences \nBENDAZZOLI, C., SANDRELLI, A. AND M. RUSSO (2011) “Disfluencies in simultaneous interpreting: a corpus-based analysis”, in A. Kruger, K. Walmach and J. Munday (eds.) Corpus-based Translation Studies: Research and Applications, London /New York: Continuum, 282-306. \nMONTI, C., BENDAZZOLI, C., SANDRELLI A. AND M. RUSSO (2005) “Studying Directionality in Simultaneous Interpreting through an Electronic Corpus: EPIC (European Parliament Interpreting Corpus)” paper presented at the International Symposium “Pour une traductologie proactive” organised for the 50° anniversary of META, University of Montreal, 6th-9th April 2005, (vol 50:4). Online: http://www.erudit.org/revue/meta/2005/v50/n4/019850ar.pdf \nRUSSO, M., BENDAZZOLI, C. E A. SANDRELLI (2006) "Looking for Lexical Patterns in a Trilingual Corpus of Source and Interpreted Speeches: Extended Analysis of EPIC (European Parliament Interpreting Corpus)", Forum, vol. 4:1, 221-254. \nSANDRELLI, A., BENDAZZOLI, C. AND M. RUSSO (2010) “European Parliament Interpreting Corpus (EPIC): Methodological issues and preliminary results on lexical patterns in SI”, International Journal of Translation 22 (1-2), 165-203. \nSHLESINGER, M. (1998): “Corpus-based interpreting studies as an offshoot of corpus-based translation studies”, META, 43-4, pp. 486-493.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.027 | 0.044 |
| Meta-epidemiology (narrow) | 0.002 | 0.003 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.008 | 0.009 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.006 | 0.008 |
| Open science | 0.005 | 0.010 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.062 | 0.041 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".