Learning machines and teaching humans
Year: 2017
Type: article
Abstract: Developments in artificial intelligence (AI) and deep learning (DL) have led to increasing effective machine-translation systems, most recently evident in the measurable improvements achieved by neural machine translation (NMT) over existing statistical machine translation (SMT) systems. Although NMT is still at the R&D stage, initial test results on both in-domain (e.g. Junczys-Dowmunt et al. 2016) and out-of-domain (e.g. Wu et al. 2016) performance have been promising. New AI and DL technologies are set to change the face of translation as well the nature and the distribution of translators’ and other language mediators’ primary tasks. The Translation Automation User Society (TAUS), for instance, is predicting Fully Automatic Useful Translation (FAUT) by about 2030 (Massardo et al. 2016). Yet, although theoretical and practical courses on computer-aided and machine translation are widespread in language mediator education, less attention has been directed towards how recent advances in AI and DL might feasibly and coherently be accommodated in competence profiles, development models and curricula. This presentation addresses the challenge to language mediator education posed by AI in general and NMT in particular. From the perspective of competence-oriented curriculum development, it considers the ways in which students can gain the knowledge and toolkits to learn when and how new technologies should be embraced, and proposes a framework for when and how the added value of human intuition, creativity and ethics might be deployed in a future “re-humanized” approach to language mediator education.
Authors: Gary Massey, Maureen Ehrensberger‐Dow
Language: English
Subfield: Language and Linguistics
Field: Arts and Humanities
Domain: Social Sciences
SDG: Quality education
Open Access status: green
OpenAlex ID: W2794613673
