Today, speech technology is only available for a small fraction of the thousands of languages spoken around the world because traditional systems need to be trained on large amounts of annotated speech audio with transcriptions. Obtaining that kind of data for every human language and dialect is almost impossible.
Wav2vec works around this limitation by requiring little to no transcribed data. The model uses self-supervision to push the boundaries by learning from unlabeled training data. This enables speech recognition systems for many more languages and dialects, such as Kyrgyz and Swahili, which don’t have a lot of transcribed speech audio. Self-supervision is the key to leveraging unannotated data and building better systems.
Textuality is often thought of in linguistic terms; for instance, the talk and writing that circulate in the classroom. In this paper I take a multimodal perspective on textuality and context. I draw on illustrative examples from school Science and English to examine how image, colour, gesture, gaze, posture and movement—as well as writing and speech—are mobilized and orchestrated by teachers and students, and how this shapes learning contexts. Throughout the paper I discuss the issues raised by a multimodal perspective for the conceptualization of text and learning context, and how this approach can contribute to learning and pedagogy more generally. I suggest that attending to the full ensemble of communicative modes involved in learning contexts enables a richer view of the complex ways in which curriculum knowledge (and policy) is mediated and articulated through classroom practices.
DocumentCloud is a catalog of primary source documents and a tool for annotating, organizing and publishing them on the web. Documents are contributed by journalists, researchers and archivists. We're helping reporters get more out of documents and helping newsrooms make their online presence more engaging.
Tesla (an acronym for Text Engineering Software Laboratory), is a Java-based open-source framework for computational linguistics, developed by the department of Computational Linguistics at the University of Cologne, Germany.
CATMA integrates three functional, interactive modules: a tagger, a query-builder and an analyzer. The analyzer module contains most of the text analytical functions known to users of TACT
A. Nenkova, und R. Passonneau. Proceedings of the Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics: HLT-NAACL 2004, Seite 145--152. Boston, Massachusetts, USA, Association for Computational Linguistics, (2004)
H. Seidel. Informatik 2009: Im Focus das Leben, Beiträge der 39. Jahrestagung der Gesellschaft für Informatik e.V. (GI), Volume P-154 von GI Edition: Lecture Notes in Informatics (LNI) -- Proceedings, Seite 5--14. Bonn, Gesellschaft für Informatik, (2009)