Proceedings of the ACL Workshop on Building and Using Parallel Texts
Notice bibliographique
Résumé
The ACL 2005 Workshop on Building and Using Parallel Texts: Data-Driven Machine Translation and Beyond, took place onWednesday, June 29 and Thursday, June 30 in Ann Arbor Michigan, immediately following the 43rd Annual Meeting of the Association for Computational Linguistics. This workshop represented a merger of two workshops that were originally proposed as independent events. Joel Martin, Rada Mihalcea, and Ted Pedersen had proposed a workshop on Building and Using Parallel Texts for Languages with Scarce Resources, which was intended as a follow-up event to the NAACL 2003 Workshop on Parallel Text that had been organized by Mihalcea and Pedersen. At the same time, Philipp Koehn and Christof Monz had proposed a workshop on Exploiting Parallel Texts for Statistical Machine Translation, featuring a shared task on Phrase Based Machine Translation. Given the close relationship between the two proposed topics, the idea of a merger was quickly embraced by all concerned. It was agreed that the workshop would have two tracks, one regarding Parallel Texts for Languages with Scarce Resources (Track 1), and the other focused on Statistical Machine Translation (Track 2). Prior to the workshop, in addition to soliciting relevant papers for review and possible presentation, the organizers of both tracks conducted shared tasks that brought together systems for an evaluation on previously unseen data. Track 1 featured a Word Alignment shared task, where the object was to align parallel text in one or more of the following langauge pairs: Inuktitut-English, Romanian-English, and Hindi-English. Track 2 carried out a shared task on Phrase Based Statistical Machine Translation, where eleven participating teams competed to build machine translation systems for French-English, Spanish-English, German-English, and Finnish-English. The results of the shared tasks were announced at the workshop, and these proceedings also include an overview paper for each shared task that summarizes the results, as well as provides information about the data used and any procedures that were followed in conducting or scoring the task. In addition, there are short papers from each participating team for each shared task that describe their underlying system in some detail. Wednesday, June 29 was dedicated to Track 1. It featured an invited talk by Mike Maxwell of the Linguistic Data Consortium, eight long paper presentations relevant to the topic of building and using parallel texts for languages with scarce resources, six short paper presentations describing systems that participated in the Word Alignment shared task (four additional short papers are included in the proceedings), a shared task overview, and a panel discussion about lessons learned from the shared task. Track 2 was featured on Thursday, June 30. It included an invited talk by Franz Josef Och of Google, six long paper presentations, a shared task overview, and nine shared task system descriptions.
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction machine sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Le volet Gemma est une étiquette directe du modèle pour chaque travail de la base, lue sur la notice réduite au titre. Le volet Codex est un classifieur appris des 10 348 étiquettes directes de Codex et calibré sur les taux pondérés de l'échantillon; les champs sans appui suffisant ne portent aucun appel Codex. Le mode candidate est l'union des deux volets; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont pas des étiquettes humaines.
Scores du classifieur distillé par catégorie (deux têtes)
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,016 | 0,047 |
| Méta-épidémiologie (sens strict) | 0,004 | 0,003 |
| Méta-épidémiologie (sens large) | 0,003 | 0,004 |
| Bibliométrie | 0,006 | 0,005 |
| Études des sciences et des technologies | 0,003 | 0,004 |
| Communication savante | 0,018 | 0,031 |
| Science ouverte | 0,007 | 0,013 |
| Intégrité de la recherche | 0,006 | 0,008 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,124 | 0,072 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule source (Gemma direct ou Codex distillé), pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».