Understanding post-editing for machine translation: challenges and opportunities
Federico Gaspari, Macerata, Italy
This talk will dispel some of the widespread myths surrounding machine translation (MT) and post-editing (PE), with the aim of stimulating a prejudice-free and realistic reflection on their current impact and future prospects, especially for professional translators, editors, revisers, localisation specialists, project managers and language service providers working with English. The presentation will explore the complex interplay between clients’ quality expectations of MT and the effectiveness of PE, illustrating the main challenges and opportunities involved in a rapidly evolving market.
I will explain the main technical differences between the currently competing paradigms of statistical and neural MT, pointing out their strengths and weaknesses, particularly those with implications for PE. I will then describe the resources required to build state-of-the-art MT systems and provide an overview of the main approaches to MT quality evaluation, including both manual techniques and the most commonly used automatic metrics.
The focus will then shift to PE for MT, a service that is increasingly required by translation buyers and for which there is an apparent skills shortage. Following a discussion of the key criteria that can be used to estimate the actual effectiveness of PE of MT output as a productivity tool without compromising on the translation quality, the talk will review the core skills desirable for post-editors, emphasising that these are different from those required to translate from scratch (e.g. using translation memory software or other computer-assisted translation tools) and to revise translations done by human translators.
The professionals attending the talk will be guided to reflect on the opportunities presented by PE, depending on their working languages and on the types of clients and texts they typically deal with, so that they can consider whether they should add PE – and more broadly services related to MT – to their portfolio of professional activities.
teaches English linguistics and translation studies at the University for Foreigners “Dante Alighieri” of Reggio Calabria, and is a post-doctoral researcher at Dublin City University, where he works on European research projects focusing on machine translation evaluation. He earned a PhD in machine translation from the University of Manchester and has held lecturing positions at the Universities of Manchester, Salford, Bologna at Forlì and Macerata. His teaching and research interests include translation technologies (especially machine translation evaluation and post-editing), technical and specialised translation, and corpus linguistics. He is a regular presenter at international conferences on these topics.