The University of Alcalá (UAH), one of KantanMT’s Academic Partners used the platform to teach final year undergraduate students Custom Machine Translation during the 2015-2016 academic year.
KantanMT.com was used in the course ‘Machine Translation and Post-editing,’ which was taught for the first time in the ‘Degree in Modern Languages Applied to Translation’ in UAH. English and Spanish were the main languages used during this course.
We caught up with two students from the course taught by Professor Cristina Toledo Báez, where they speak about their experience using KantanMT.com. To see a Lecturer’s perspective on teaching MT, read our interview with Professor Toledo.
Alice Hammonds, Degree in Modern Languages Applied to Translation, UAH
KantanMT: What made you choose this course, and how to do you think it will help your career?
Alice: As an English student studying Spanish and Portuguese, I have encountered, and used Machine Translation throughout my degree. The option to explore MT further interested me since I began to post-edit translations. For the future, I think what I have learnt on this course could aid me in finding a job as a translator. I have also learnt a lot about tools I had never experienced before, such as KantanMT.
KantanMT: Had you any previous experience of Machine Translation, if not how did you find using the KantanMT system?
Alice: I had no previous experience of MT, and at first KantanMT was completely alien to me. However, with the instructional videos they provide on their website and the instructions provided by our teacher, I found using the MT engines relatively problem free.
KantanMT: Can you describe your assignment and thoughts about using the KantanMT, what did you find good and do you have any suggestions for improvement?
Alice: Firstly, in order to draw comparisons between the two, we used KantanMT to create translations with and without parallel texts. We also added elements such as glossaries to see if these strengthened the translations provided by the engine. For the final task, we created a Translation Memory (I used the website wordfast) based on parallel texts to see if this provided a better translation. I found it interesting to note the difference that adding these texts made.
The only improvement I would suggest is to make the engine building process faster. Each element on KantanMT, from building an engine to creating a translation, takes a fair amount of time. Therefore, it takes more time to complete the tasks.
KantanMT Responds:
Thank you Alice for your feedback. We have allocated a set number of servers for our academic partners, this sometimes means delays in building in engines when groups of students are using the platform simultaneously.
However, when working in a real-world project with KantanMT, the servers are scaled up according to the translation needs of our clients, and we are proud to be one of the most scalable and fastest Custom Machine Translation solution providers in the world, We can build engines and translate at the minimum rate of 6 million words per hour.
If a client requires to build an MT engine quickly, they can also use one of our pre-built engines in the KantanFleet, to get their projects up and running. You can read more about the pre-built engines for Financial and Medical domain, as well as Legal domain on our News and Events page.
Read our interview with Professor Cristina Toledo Báez to know more about how CMT was used during the course.
KantanMT: What is your impression of the translation industry, and in your opinion, what do you think the industry will look like in the future?
Alice: My impression of the translation industry has definitely changed. Prior to undertaking this course, I thought the majority of translations were done manually, especially for the translation of literature, subtitles etc. However, I now realise that Machine Translation is not as much of a ‘cheat’ as usually insinuated in some schools and universities.
Furthermore, the translation provided by MT can be of high quality. Machine Translation can facilitate a number of jobs, including that of a post-editor. I think in the future, MT will continue to improve and the awareness surrounding it will grow, making it less of a ‘cheat’ and more of an ‘aid’.
Raymonda Nodis Cesarela, Degree in Modern Languages Applied to Translation, UAH
KantanMT: What made you choose this course, and how to do you think it will help your career?
Raymonda: These days, the knowledge regarding the use of Machine Translation is indispensable in the translation field. Therefore, I decided to join this course in order to improve my translation skills and at the same time, to learn more about automated translation. I would say that thanks to this course – the theoretical part and the practical exercises – I have expanded my abilities in translation and I feel more prepared for my future career.
KantanMT: Did you have any previous experience of Machine Translation, if so how does KantanMT.com compare?
Raymonda: During my Erasmus stay in Paris last year, I had the opportunity to work with some particular tools related to MT, especially with the free versions of CAT tools such as Systran, OmegaT, WordFastPro, Wordfast Anywhere, LF Aligner etc. However, I used KantanMT for the first time during the current course and I can say that it was a rewarding experience. Moreover, I consider that this MT tool is different from the others because it is easy to use and also quite fast, intuitive and reliable.
KantanMT: Can you describe your assignment and thoughts about using the KantanMT, what did you find good and do you have any suggestions for improvement?
Raymonda: We have used KantanMT in several assignments for this course. First, we learnt the basics of KantanMT. We got familiar with the web page, we watched the demo videos and we created our accounts etc. Then we translated a text without training data and after, the same text was translated, but this time we used training data (glossaries, parallel texts, Translation Memories).
After downloading the results provided by KantanMT, we had to commit them in an .xlsx file while analysing the adequacy, fluency and also the differences of the translation output (both without and with training data). As conclusion, I would say that the results provided by KantanMT improved after uploading the training data.
I noticed that the translation process is quite time-consuming and in some occasions, an error message appears and the process has to be restarted from the beginning. However, this may be due to the format of this system, which depends on the internet connection.
KantanMT Responds:
Raymonda, thank you for your feedback. We are sorry to hear you encountered error messages while working with translation jobs. Sometimes, this can happen if the right training data is not uploaded, or if they are uploaded incorrectly. We always try to improve our MT platform based on these suggestions.
KantanMT is a cloud-based solution and requires and internet connection to access, but all jobs are carried out on the cloud, so once a request is sent the performance of the platform should not be affected by the internet connection.
KantanMT: What is your impression of the translation industry, and in your opinion, what do you think the industry will look like in the future?
Raymonda: Nowadays, translation is a very important industry in the globalized world. Despite its relatively short existence, in recent years, Machine Translation has constantly developed and improved and it seems that it will become an indispensable tool for the professional translator.
Thanks to the technological progress in this field, translation has improved in quality, speed and productivity. For this reason, MT should be considered an essential tool, which is able to provide a high-quality output when used in combination with human translators.
About the University of Alcalá
The University of Alcalá is a UNESCO world heritage site and one of the longest standing European universities which dates back to May 20, 1293. Today the University of Alcalá is a modern institution which covers all fields of knowledge, from humanities to engineering, and from social science to experimental and biomedical science. An estimated 28,000 students are currently pursuing their degrees at the University of Alcalá.