Quality Estimation for Machine Translation
This project offers full scholarships (at Home rates) to suitable PhD candidates who wish to work on the interaction between translation and technology, especially with regard to how we can use computers to assess the quality of (automatic) translations. In the context of this PhD assessment of quality covers a wide range of topics including quality estimation, (automatic) post-editing, human evaluation, and even new evaluation metrics.
Start date1 July 2022
Funding sourceUniversity of Surrey
A stipend of £15,609 for 2021/22, which will increase each year in line with the UK Research and Innovation (UKRI) rate, plus home rate fee allowance of £4,500 (with automatic increase to UKRI rate each year). For exceptional international candidates, there is the possibility of obtaining a scholarship to cover overseas fees.
Quality Estimation (QE) for Machine Translation (MT) enables the evaluation of MT systems without needing the reference translations. With the help of the recent advances in the Natural Language Processing (NLP), QE systems have emerged as a viable solution to evaluate the performance of black-box Neural MT systems. Since QE makes it possible to estimate translation quality based on learned correspondences between source and target segments, it is an active field of MT research that offers exciting research opportunities.
We are seeking a motivated candidate who is interested in pursuing a PhD on the topic of how we can use linguistically grounded methods to improve the predictions obtained from QE systems. The research to be carried out is expected to propose new methods which can be applied to automatically obtained translations. New evaluation methods could be built on recent data-driven approaches or propose alternative metrics which rely on eye-tracking or postediting effort to assess the quality. The overall research agenda will also entail applying QE to low-resource languages and efforts towards developing data for multiple language pairs. In the light of this, programming experience is essential and experience with Python, PyTorch or TensorFlow, along with previous experience in NLP is desirable.
The successful project will need to rely on solid, evidence-based, eclectic mixed-methods approach benefiting from cross-fertilization among different disciplines (deep learning, natural language processing, translation studies, eye-tracking, etc.) in order to further the boundaries of QE research. The successful candidate will benefit from excellent technological working conditions, international contacts, and a stimulating interdisciplinary work environment.
Dr Diptesh Kanojia
Prof Constantin Orăsan
Related linksSurrey Institute for People-Centred AI Centre for Vision, Speech and Signal Processing Centre for Translation Studies
All applicants should have (or expect to obtain) a first-class degree in a numerate discipline (mathematics, science or engineering) or MSc with Distinction (or 70% average) and a strong interest in pursuing research in this field. Additional experience which is relevant to the area of research is also advantageous.
How to apply
Applications should be submitted via the Vision, Speech and Signal Processing PhD programme page on the "Apply" tab. Please clearly state the studentship title and supervisor on your application.
Vision, Speech and Signal Processing PhD
Read our studentship FAQs to find out more about applying and funding.