no code implementations • EMNLP (insights) 2021 • Jan Rosendahl, Christian Herold, Frithjof Petrick, Hermann Ney
In this work, we conduct a comprehensive investigation on one of the centerpieces of modern machine translation systems: the encoder-decoder attention mechanism.
no code implementations • IWSLT (ACL) 2022 • Frithjof Petrick, Jan Rosendahl, Christian Herold, Hermann Ney
After its introduction the Transformer architecture quickly became the gold standard for the task of neural machine translation.
no code implementations • IWSLT 2017 • Parnia Bahar, Jan Rosendahl, Nick Rossenbach, Hermann Ney
This work describes the Neural Machine Translation (NMT) system of the RWTH Aachen University developed for the English$German tracks of the evaluation campaign of the International Workshop on Spoken Language Translation (IWSLT) 2017.
no code implementations • Findings (ACL) 2022 • Christian Herold, Jan Rosendahl, Joris Vanvinckenroye, Hermann Ney
The filtering and/or selection of training data is one of the core aspects to be considered when building a strong machine translation system. In their influential work, Khayrallah and Koehn (2018) investigated the impact of different types of noise on the performance of machine translation systems. In the same year the WMT introduced a shared task on parallel corpus filtering, which went on to be repeated in the following years, and resulted in many different filtering approaches being proposed. In this work we aim to combine the recent achievements in data filtering with the original analysis of Khayrallah and Koehn (2018) and investigate whether state-of-the-art filtering systems are capable of removing all the suggested noise types. We observe that most of these types of noise can be detected with an accuracy of over 90% by modern filtering systems when operating in a well studied high resource setting. However, we also find that when confronted with more refined noise categories or when working with a less common language pair, the performance of the filtering systems is far from optimal, showing that there is still room for improvement in this area of research.
no code implementations • EMNLP (IWSLT) 2019 • Jan Rosendahl, Viet Anh Khoa Tran, Weiyue Wang, Hermann Ney
In this work we analyze and compare the behavior of the Transformer architecture when using different positional encoding methods.
no code implementations • 18 Oct 2021 • Nils-Philipp Wynands, Wilfried Michel, Jan Rosendahl, Ralf Schlüter, Hermann Ney
Lastly, it is shown that this technique can be used to effectively perform sequence discriminative training for attention-based encoder-decoder acoustic models on the LibriSpeech task.
Automatic Speech Recognition Automatic Speech Recognition (ASR) +2
no code implementations • 27 Sep 2021 • Evgeniia Tokarchuk, Jan Rosendahl, Weiyue Wang, Pavel Petrushkov, Tomer Lancewicki, Shahram Khadivi, Hermann Ney
Pivot-based neural machine translation (NMT) is commonly used in low-resource setups, especially for translation between non-English language pairs.
no code implementations • ACL (IWSLT) 2021 • Evgeniia Tokarchuk, Jan Rosendahl, Weiyue Wang, Pavel Petrushkov, Tomer Lancewicki, Shahram Khadivi, Hermann Ney
Complex natural language applications such as speech translation or pivot translation traditionally rely on cascaded models.
no code implementations • NAACL 2021 • Christian Herold, Jan Rosendahl, Joris Vanvinckenroye, Hermann Ney
While we find that our approaches come out at the top on all three tasks, different variants perform best on different tasks.
no code implementations • WS 2019 • Jan Rosendahl, Christian Herold, Yunsu Kim, Miguel Gra{\c{c}}a, Weiyue Wang, Parnia Bahar, Yingbo Gao, Hermann Ney
For the De-En task, none of the tested methods gave a significant improvement over last years winning system and we end up with the same performance, resulting in 39. 6{\%} BLEU on newstest2019.
no code implementations • WS 2019 • Yunsu Kim, Hendrik Rosendahl, Nick Rossenbach, Jan Rosendahl, Shahram Khadivi, Hermann Ney
We propose a novel model architecture and training algorithm to learn bilingual sentence embeddings from a combination of parallel and monolingual data.
no code implementations • WS 2018 • Nick Rossenbach, Jan Rosendahl, Yunsu Kim, Miguel Gra{\c{c}}a, Aman Gokrani, Hermann Ney
We use several rule-based, heuristic methods to preselect sentence pairs.
1 code implementation • WS 2018 • Julian Schamper, Jan Rosendahl, Parnia Bahar, Yunsu Kim, Arne Nix, Hermann Ney
In total we improve by 6. 8{\%} BLEU over our last year{'}s submission and by 4. 8{\%} BLEU over the winning system of the 2017 German→English task.