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 • WMT (EMNLP) 2020 • Peter Stanchev, Weiyue Wang, Hermann Ney
An important aspect of machine translation is its evaluation, which can be achieved through the use of a variety of metrics.
no code implementations • IWSLT 2016 • Wilfried Michel, Zoltán Tüske, M. Ali Basha Shaik, Ralf Schlüter, Hermann Ney
In this paper the RWTH large vocabulary continuous speech recognition (LVCSR) systems developed for the IWSLT-2016 evaluation campaign are described.
no code implementations • IWSLT 2016 • Jan-Thorsten Peter, Andreas Guta, Nick Rossenbach, Miguel Graça, Hermann Ney
This work describes the statistical machine translation (SMT) systems of RWTH Aachen University developed for the evaluation campaign of International Workshop on Spoken Language Translation (IWSLT) 2016.
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 • 27 Jan 2025 • Zijian Yang, Vahe Eminyan, Ralf Schlüter, Hermann Ney
In this work, we apply classification error bounds to study the relationship between the error mismatch and the Kullback-Leibler divergence in machine learning.
no code implementations • 24 Oct 2024 • David Thulke, Yingbo Gao, Rricha Jalota, Christian Dugast, Hermann Ney
This paper explores the rapid development of a telephone call summarization system utilizing large language models (LLMs).
1 code implementation • 1 Oct 2024 • Robin Schmitt, Albert Zeyer, Mohammad Zeineldeen, Ralf Schlüter, Hermann Ney
We sometimes observe monotonically decreasing cross-attention weights in our Conformer-based global attention-based encoder-decoder (AED) models, Further investigation shows that the Conformer encoder reverses the sequence in the time dimension.
1 code implementation • 17 Jan 2024 • David Thulke, Yingbo Gao, Petrus Pelser, Rein Brune, Rricha Jalota, Floris Fok, Michael Ramos, Ian van Wyk, Abdallah Nasir, Hayden Goldstein, Taylor Tragemann, Katie Nguyen, Ariana Fowler, Andrew Stanco, Jon Gabriel, Jordan Taylor, Dean Moro, Evgenii Tsymbalov, Juliette de Waal, Evgeny Matusov, Mudar Yaghi, Mohammad Shihadah, Hermann Ney, Christian Dugast, Jonathan Dotan, Daniel Erasmus
To increase the accessibility of our model to non-English speakers, we propose to make use of cascaded machine translation and show that this approach can perform comparably to natively multilingual models while being easier to scale to a large number of languages.
no code implementations • 18 Oct 2023 • Frithjof Petrick, Christian Herold, Pavel Petrushkov, Shahram Khadivi, Hermann Ney
Finally, we explore language model fusion in the light of recent advancements in large language models.
no code implementations • 11 Oct 2023 • Zijian Yang, Wei Zhou, Ralf Schlüter, Hermann Ney
In this work, we investigate the effect of language models (LMs) with different context lengths and label units (phoneme vs. word) used in sequence discriminative training for phoneme-based neural transducers.
1 code implementation • 4 Oct 2023 • Daniel Mann, Tina Raissi, Wilfried Michel, Ralf Schlüter, Hermann Ney
We investigate recognition results and additionally Viterbi alignments of our models.
no code implementations • 25 Sep 2023 • Zijian Yang, Wei Zhou, Ralf Schlüter, Hermann Ney
Empirically, we show that ILM subtraction and sequence discriminative training achieve similar effects across a wide range of experiments on Librispeech, including both MMI and minimum Bayes risk (MBR) criteria, as well as neural transducers and LMs of both full and limited context.
no code implementations • 15 Sep 2023 • Mohammad Zeineldeen, Albert Zeyer, Ralf Schlüter, Hermann Ney
We study a streamable attention-based encoder-decoder model in which either the decoder, or both the encoder and decoder, operate on pre-defined, fixed-size windows called chunks.
no code implementations • 8 Aug 2023 • Peter Vieting, Ralf Schlüter, Hermann Ney
In this work, we study its capability to replace the standard feature extraction methods in a connectionist temporal classification (CTC) ASR model and compare it to an alternative neural FE.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
no code implementations • 8 Jun 2023 • Christian Herold, Yingbo Gao, Mohammad Zeineldeen, Hermann Ney
The integration of language models for neural machine translation has been extensively studied in the past.
no code implementations • 8 Jun 2023 • Christian Herold, Hermann Ney
Document-level context for neural machine translation (NMT) is crucial to improve the translation consistency and cohesion, the translation of ambiguous inputs, as well as several other linguistic phenomena.
no code implementations • 8 Jun 2023 • Christian Herold, Hermann Ney
On the other hand, in most works, the question on how to perform search with the trained model is scarcely discussed, sometimes not mentioned at all.
no code implementations • 28 May 2023 • Wei Zhou, Eugen Beck, Simon Berger, Ralf Schlüter, Hermann Ney
Modern public ASR tools usually provide rich support for training various sequence-to-sequence (S2S) models, but rather simple support for decoding open-vocabulary scenarios only.
no code implementations • 14 Apr 2023 • David Thulke, Nico Daheim, Christian Dugast, Hermann Ney
This paper summarizes our contributions to the document-grounded dialog tasks at the 9th and 10th Dialog System Technology Challenges (DSTC9 and DSTC10).
no code implementations • 11 Jan 2023 • Christoph Lüscher, Jingjing Xu, Mohammad Zeineldeen, Ralf Schlüter, Hermann Ney
By further adding neural speaker embeddings, we gain additional ~3% relative WER improvement on Hub5'00.
no code implementations • 7 Dec 2022 • Zijian Yang, Wei Zhou, Ralf Schlüter, Hermann Ney
Compared to the N-best-list based minimum Bayes risk objectives, lattice-free methods gain 40% - 70% relative training time speedup with a small degradation in performance.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
no code implementations • 11 Nov 2022 • Wei Zhou, Haotian Wu, Jingjing Xu, Mohammad Zeineldeen, Christoph Lüscher, Ralf Schlüter, Hermann Ney
Detailed analysis and experimental verification are conducted to show the optimal positions in the ASR neural network (NN) to apply speaker enhancing and adversarial training.
1 code implementation • 9 Nov 2022 • Baohao Liao, David Thulke, Sanjika Hewavitharana, Hermann Ney, Christof Monz
We show: (1) [MASK]s can indeed be appended at a later layer, being disentangled from the word embedding; (2) The gathering of contextualized information from unmasked tokens can be conducted with a few layers.
1 code implementation • 31 Oct 2022 • Nico Daheim, David Thulke, Christian Dugast, Hermann Ney
In this work, we present a model for document-grounded response generation in dialog that is decomposed into two components according to Bayes theorem.
1 code implementation • 26 Oct 2022 • Albert Zeyer, Robin Schmitt, Wei Zhou, Ralf Schlüter, Hermann Ney
We restrict the decoder attention to segments to avoid quadratic runtime of global attention, better generalize to long sequences, and eventually enable streaming.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
no code implementations • 26 Oct 2022 • Peter Vieting, Christoph Lüscher, Julian Dierkes, Ralf Schlüter, Hermann Ney
Unsupervised representation learning has recently helped automatic speech recognition (ASR) to tackle tasks with limited labeled data.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
no code implementations • 24 Oct 2022 • Christoph Lüscher, Mohammad Zeineldeen, Zijian Yang, Tina Raissi, Peter Vieting, Khai Le-Duc, Weiyue Wang, Ralf Schlüter, Hermann Ney
Language barriers present a great challenge in our increasingly connected and global world.
1 code implementation • 24 Oct 2022 • Viet Anh Khoa Tran, David Thulke, Yingbo Gao, Christian Herold, Hermann Ney
Currently, in speech translation, the straightforward approach - cascading a recognition system with a translation system - delivers state-of-the-art results.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
no code implementations • 21 Oct 2022 • Yingbo Gao, Christian Herold, Zijian Yang, Hermann Ney
Encoder-decoder architecture is widely adopted for sequence-to-sequence modeling tasks.
no code implementations • 21 Oct 2022 • Yingbo Gao, Christian Herold, Zijian Yang, Hermann Ney
Checkpoint averaging is a simple and effective method to boost the performance of converged neural machine translation models.
no code implementations • 26 Jun 2022 • Mohammad Zeineldeen, Jingjing Xu, Christoph Lüscher, Ralf Schlüter, Hermann Ney
In this work, we investigate various methods for speaker adaptive training (SAT) based on feature-space approaches for a conformer-based acoustic model (AM) on the Switchboard 300h dataset.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
no code implementations • 22 Apr 2022 • Wei Zhou, Wilfried Michel, Ralf Schlüter, Hermann Ney
In this work, we propose an efficient 3-stage progressive training pipeline to build highly-performing neural transducer models from scratch with very limited computation resources in a reasonable short time period.
1 code implementation • 16 Dec 2021 • David Thulke, Nico Daheim, Christian Dugast, Hermann Ney
This paper summarizes our submission to Task 2 of the second track of the 10th Dialog System Technology Challenge (DSTC10) "Knowledge-grounded Task-oriented Dialogue Modeling on Spoken Conversations".
no code implementations • 11 Nov 2021 • Zijian Yang, Yingbo Gao, Alexander Gerstenberger, Jintao Jiang, Ralf Schlüter, Hermann Ney
Compared to our previous work, the criteria considered in this work are self-normalized and there is no need to further conduct a correction step.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
no code implementations • 5 Nov 2021 • Mohammad Zeineldeen, Jingjing Xu, Christoph Lüscher, Wilfried Michel, Alexander Gerstenberger, Ralf Schlüter, Hermann Ney
The recently proposed conformer architecture has been successfully used for end-to-end automatic speech recognition (ASR) architectures achieving state-of-the-art performance on different datasets.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
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 • 18 Oct 2021 • Felix Meyer, Wilfried Michel, Mohammad Zeineldeen, Ralf Schlüter, Hermann Ney
We show on the LibriSpeech (LBS) and Switchboard (SWB) corpora that the model scales for a combination of attentionbased encoder-decoder acoustic model and language model can be learned as effectively as with manual tuning.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+5
no code implementations • 13 Oct 2021 • Wei Zhou, Zuoyun Zheng, Ralf Schlüter, Hermann Ney
In this work, we study various ILM correction-based LM integration methods formulated in a common RNN-T framework.
no code implementations • ACL 2021 • Evgeniia Tokarchuk, David Thulke, Weiyue Wang, Christian Dugast, Hermann Ney
Data processing is an important step in various natural language processing tasks.
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 • 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 2021 • Weiyue Wang, Zijian Yang, Yingbo Gao, Hermann Ney
The neural hidden Markov model has been proposed as an alternative to attention mechanism in machine translation with recurrent neural networks.
1 code implementation • ACL (dialdoc) 2021 • Nico Daheim, David Thulke, Christian Dugast, Hermann Ney
For the second subtask, we use a cascaded model which grounds the response prediction on the predicted span instead of the full document.
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.
1 code implementation • 31 May 2021 • Albert Zeyer, Ralf Schlüter, Hermann Ney
The peaky behavior of CTC models is well known experimentally.
no code implementations • 21 Apr 2021 • Yingbo Gao, David Thulke, Alexander Gerstenberger, Khoa Viet Tran, Ralf Schlüter, Hermann Ney
As the vocabulary size of modern word-based language models becomes ever larger, many sampling-based training criteria are proposed and investigated.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
no code implementations • 19 Apr 2021 • Wei Zhou, Mohammad Zeineldeen, Zuoyun Zheng, Ralf Schlüter, Hermann Ney
Subword units are commonly used for end-to-end automatic speech recognition (ASR), while a fully acoustic-oriented subword modeling approach is somewhat missing.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
no code implementations • 13 Apr 2021 • Wei Zhou, Albert Zeyer, André Merboldt, Ralf Schlüter, Hermann Ney
With the advent of direct models in automatic speech recognition (ASR), the formerly prevalent frame-wise acoustic modeling based on hidden Markov models (HMM) diversified into a number of modeling architectures like encoder-decoder attention models, transducer models and segmental models (direct HMM).
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
no code implementations • 12 Apr 2021 • Mohammad Zeineldeen, Aleksandr Glushko, Wilfried Michel, Albert Zeyer, Ralf Schlüter, Hermann Ney
Attention-based encoder-decoder (AED) models learn an implicit internal language model (ILM) from the training transcriptions.
no code implementations • 12 Apr 2021 • Nick Rossenbach, Mohammad Zeineldeen, Benedikt Hilmes, Ralf Schlüter, Hermann Ney
We achieve a final word-error-rate of 3. 3%/10. 0% with a hybrid system on the clean/noisy test-sets, surpassing any previous state-of-the-art systems on Librispeech-100h that do not include unlabeled audio data.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+4
no code implementations • 9 Apr 2021 • Peter Vieting, Christoph Lüscher, Wilfried Michel, Ralf Schlüter, Hermann Ney
With the success of neural network based modeling in automatic speech recognition (ASR), many studies investigated acoustic modeling and learning of feature extractors directly based on the raw waveform.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
2 code implementations • 7 Apr 2021 • Albert Zeyer, André Merboldt, Wilfried Michel, Ralf Schlüter, Hermann Ney
We present our transducer model on Librispeech.
Ranked #33 on
Speech Recognition
on LibriSpeech test-clean
(using extra training data)
no code implementations • 30 Mar 2021 • Albert Zeyer, Ralf Schlüter, Hermann Ney
We compare several monotonic latent models to our global soft attention baseline such as a hard attention model, a local windowed soft attention model, and a segmental soft attention model.
1 code implementation • 9 Feb 2021 • David Thulke, Nico Daheim, Christian Dugast, Hermann Ney
This paper summarizes our work on the first track of the ninth Dialog System Technology Challenge (DSTC 9), "Beyond Domain APIs: Task-oriented Conversational Modeling with Unstructured Knowledge Access".
no code implementations • COLING 2020 • Yingbo Gao, Baohao Liao, Hermann Ney
Soft contextualized data augmentation is a recent method that replaces one-hot representation of words with soft posterior distributions of an external language model, smoothing the input of neural machine translation systems.
no code implementations • COLING 2020 • Zhihong Lei, Weiyue Wang, Christian Dugast, Hermann Ney
Named entity recognition is a key component in various natural language processing systems, and neural architectures provide significant improvements over conventional approaches.
no code implementations • Asian Chapter of the Association for Computational Linguistics 2020 • Zijian Yang, Yingbo Gao, Weiyue Wang, Hermann Ney
Attention-based encoder-decoder models have achieved great success in neural machine translation tasks.
no code implementations • Asian Chapter of the Association for Computational Linguistics 2020 • Yingbo Gao, Weiyue Wang, Christian Herold, Zijian Yang, Hermann Ney
In order to combat overfitting and in pursuit of better generalization, label smoothing is widely applied in modern neural machine translation systems.
no code implementations • 24 Nov 2020 • Parnia Bahar, Tobias Bieschke, Ralf Schlüter, Hermann Ney
Direct speech translation is an alternative method to avoid error propagation; however, its performance is often behind the cascade system.
no code implementations • 24 Nov 2020 • Parnia Bahar, Christopher Brix, Hermann Ney
Neural translation models have proven to be effective in capturing sufficient information from a source sentence and generating a high-quality target sentence.
no code implementations • Findings of the Association for Computational Linguistics 2020 • Baohao Liao, Yingbo Gao, Hermann Ney
Mutual learning, where multiple agents learn collaboratively and teach one another, has been shown to be an effective way to distill knowledge for image classification tasks.
no code implementations • 30 Oct 2020 • Wei Zhou, Simon Berger, Ralf Schlüter, Hermann Ney
To join the advantages of classical and end-to-end approaches for speech recognition, we present a simple, novel and competitive approach for phoneme-based neural transducer modeling.
no code implementations • WMT (EMNLP) 2020 • Jingjing Huo, Christian Herold, Yingbo Gao, Leonard Dahlmann, Shahram Khadivi, Hermann Ney
Context-aware neural machine translation (NMT) is a promising direction to improve the translation quality by making use of the additional context, e. g., document-level translation, or having meta-information.
no code implementations • 20 May 2020 • Jingjing Huo, Yingbo Gao, Weiyue Wang, Ralf Schlüter, Hermann Ney
After that, we apply the best norm-scaling setup in combination with various margins and conduct neural language models rescoring experiments in automatic speech recognition.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+4
no code implementations • 20 May 2020 • Wilfried Michel, Ralf Schlüter, Hermann Ney
This is compared to a global renormalization scheme which is equivalent to applying shallow fusion in training.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
1 code implementation • 19 May 2020 • Albert Zeyer, André Merboldt, Ralf Schlüter, Hermann Ney
We compare the original training criterion with the full marginalization over all alignments, to the commonly used maximum approximation, which simplifies, improves and speeds up our training.
1 code implementation • 19 May 2020 • Mohammad Zeineldeen, Albert Zeyer, Wei Zhou, Thomas Ng, Ralf Schlüter, Hermann Ney
Following the rationale of end-to-end modeling, CTC, RNN-T or encoder-decoder-attention models for automatic speech recognition (ASR) use graphemes or grapheme-based subword units based on e. g. byte-pair encoding (BPE).
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
no code implementations • 15 May 2020 • Tina Raissi, Eugen Beck, Ralf Schlüter, Hermann Ney
In this work, we address a direct phonetic context modeling for the hybrid deep neural network (DNN)/HMM, that does not build on any phone clustering algorithm for the determination of the HMM state inventory.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+2
no code implementations • 11 May 2020 • Harald Hanselmann, Hermann Ney
The term fine-grained visual classification (FGVC) refers to classification tasks where the classes are very similar and the classification model needs to be able to find subtle differences to make the correct prediction.
Ranked #11 on
Fine-Grained Image Classification
on Stanford Cars
no code implementations • ACL 2020 • Christopher Brix, Parnia Bahar, Hermann Ney
Sparse models require less memory for storage and enable a faster inference by reducing the necessary number of FLOPs.
no code implementations • EAMT 2020 • Yunsu Kim, Miguel Graça, Hermann Ney
This paper studies the practicality of the current state-of-the-art unsupervised methods in neural machine translation (NMT).
no code implementations • 2 Apr 2020 • Wei Zhou, Wilfried Michel, Kazuki Irie, Markus Kitza, Ralf Schlüter, Hermann Ney
We present a complete training pipeline to build a state-of-the-art hybrid HMM-based ASR system on the 2nd release of the TED-LIUM corpus.
1 code implementation • 19 Dec 2019 • Nick Rossenbach, Albert Zeyer, Ralf Schlüter, Hermann Ney
We achieve improvements of up to 33% relative in word-error-rate (WER) over a strong baseline with data-augmentation in a low-resource environment (LibriSpeech-100h), closing the gap to a comparable oracle experiment by more than 50\%.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+5
no code implementations • 20 Nov 2019 • Parnia Bahar, Tobias Bieschke, Hermann Ney
Recent advances in deep learning show that end-to-end speech to text translation model is a promising approach to direct the speech translation field.
no code implementations • EMNLP (IWSLT) 2019 • Parnia Bahar, Albert Zeyer, Ralf Schlüter, Hermann Ney
This work investigates a simple data augmentation technique, SpecAugment, for end-to-end speech translation.
no code implementations • 20 Nov 2019 • Parnia Bahar, Albert Zeyer, Ralf Schlüter, Hermann Ney
Attention-based sequence-to-sequence models have shown promising results in automatic speech recognition.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
1 code implementation • 17 Nov 2019 • Harald Hanselmann, Hermann Ney
The task of fine-grained visual classification (FGVC) deals with classification problems that display a small inter-class variance such as distinguishing between different bird species or car models.
Ranked #20 on
Fine-Grained Image Classification
on FGVC Aircraft
no code implementations • EMNLP (IWSLT) 2019 • Yingbo Gao, Christian Herold, Weiyue Wang, Hermann Ney
Prominently used in support vector machines and logistic regressions, kernel functions (kernels) can implicitly map data points into high dimensional spaces and make it easier to learn complex decision boundaries.
1 code implementation • WS 2019 • Yunsu Kim, Duc Thanh Tran, Hermann Ney
Document-level context has received lots of attention for compensating neural machine translation (NMT) of isolated sentences.
no code implementations • IJCNLP 2019 • Yunsu Kim, Petre Petrov, Pavel Petrushkov, Shahram Khadivi, Hermann Ney
We present effective pre-training strategies for neural machine translation (NMT) using parallel corpora involving a pivot language, i. e., source-pivot and pivot-target, leading to a significant improvement in source-target translation.
1 code implementation • IJCNLP 2019 • Yingbo Gao, Weiyue Wang, Hermann Ney
The preprocessing pipelines in Natural Language Processing usually involve a step of removing sentences consisted of illegal characters.
no code implementations • WS 2019 • Peter Stanchev, Weiyue Wang, Hermann Ney
Over the years a number of machine translation metrics have been developed in order to evaluate the accuracy and quality of machine-generated translations.
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 • 1 Jul 2019 • Eugen Beck, Wei Zhou, Ralf Schlüter, Hermann Ney
LSTM based language models are an important part of modern LVCSR systems as they significantly improve performance over traditional backoff language models.
no code implementations • 1 Jul 2019 • Wilfried Michel, Ralf Schlüter, Hermann Ney
This allows for a direct comparison of lattice-based and lattice-free sequence discriminative training criteria such as MMI and sMBR, both using the same language model during training.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
no code implementations • WS 2019 • Miguel Graça, Yunsu Kim, Julian Schamper, Shahram Khadivi, Hermann Ney
Back-translation - data augmentation by translating target monolingual data - is a crucial component in modern neural machine translation (NMT).
no code implementations • 14 Jun 2019 • Markus Kitza, Pavel Golik, Ralf Schlüter, Hermann Ney
Further, i-vectors were used as an input to the neural network to perform instantaneous speaker and environment adaptation, providing 8\% relative improvement in word error rate on the NIST Hub5 2000 evaluation test set.
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.
1 code implementation • ACL 2019 • Yunsu Kim, Yingbo Gao, Hermann Ney
Transfer learning or multilingual model is essential for low-resource neural machine translation (NMT), but the applicability is limited to cognate languages by sharing their vocabularies.
Cross-Lingual Transfer
Low Resource Neural Machine Translation
+4
no code implementations • 10 May 2019 • Kazuki Irie, Albert Zeyer, Ralf Schlüter, Hermann Ney
We explore deep autoregressive Transformer models in language modeling for speech recognition.
no code implementations • 9 May 2019 • Tobias Menne, Ilya Sklyar, Ralf Schlüter, Hermann Ney
In a more realistic ASR scenario the audio signal contains significant portions of single-speaker speech and only part of the signal contains speech of multiple competing speakers.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
2 code implementations • 8 May 2019 • Christoph Lüscher, Eugen Beck, Kazuki Irie, Markus Kitza, Wilfried Michel, Albert Zeyer, Ralf Schlüter, Hermann Ney
To the best knowledge of the authors, the results obtained when training on the full LibriSpeech training set, are the best published currently, both for the hybrid DNN/HMM and the attention-based systems.
Ranked #30 on
Speech Recognition
on LibriSpeech test-other
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+4
no code implementations • EACL 2017 • Yunsu Kim, Julian Schamper, Hermann Ney
We address for the first time unsupervised training for a translation task with hundreds of thousands of vocabulary words.
no code implementations • WS 2016 • Yunsu Kim, Andreas Guta, Joern Wuebker, Hermann Ney
This work systematically analyzes the smoothing effect of vocabulary reduction for phrase translation models.
no code implementations • EMNLP 2018 • Yunsu Kim, Jiahui Geng, Hermann Ney
Unsupervised learning of cross-lingual word embedding offers elegant matching of words across languages, but has fundamental limitations in translating sentences.
no code implementations • EMNLP 2018 • Jan-Thorsten Peter, Eugen Beck, Hermann Ney
Training and testing many possible parameters or model architectures of state-of-the-art machine translation or automatic speech recognition system is a cumbersome task.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
1 code implementation • EMNLP 2018 • Parnia Bahar, Christopher Brix, Hermann Ney
This work investigates an alternative model for neural machine translation (NMT) and proposes a novel architecture, where we employ a multi-dimensional long short-term memory (MDLSTM) for translation modeling.
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.
no code implementations • WS 2018 • Christian Herold, Yingbo Gao, Hermann Ney
Embedding and projection matrices are commonly used in neural language models (NLM) as well as in other sequence processing networks that operate on large vocabularies.
no code implementations • WS 2018 • Miguel Gra{\c{c}}a, Yunsu Kim, Julian Schamper, Jiahui Geng, Hermann Ney
This paper describes the unsupervised neural machine translation (NMT) systems of the RWTH Aachen University developed for the English ↔ German news translation task of the \textit{EMNLP 2018 Third Conference on Machine Translation} (WMT 2018).
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.
no code implementations • WS 2018 • Tamer Alkhouli, Gabriel Bretschner, Hermann Ney
This work investigates the alignment problem in state-of-the-art multi-head attention models based on the transformer architecture.
no code implementations • ACL 2018 • Weiyue Wang, Derui Zhu, Tamer Alkhouli, Zixuan Gan, Hermann Ney
Attention-based neural machine translation (NMT) models selectively focus on specific source positions to produce a translation, which brings significant improvements over pure encoder-decoder sequence-to-sequence models.
no code implementations • 19 Jun 2018 • Tobias Menne, Ralf Schlüter, Hermann Ney
The proposed adaptation approach is based on the integration of the beamformer, which includes the mask estimation network, and the acoustic model of the ASR system.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
1 code implementation • CVPR 2018 • Necati Cihan Camgoz, Simon Hadfield, Oscar Koller, Hermann Ney, Richard Bowden
SLR seeks to recognize a sequence of continuous signs but neglects the underlying rich grammatical and linguistic structures of sign language that differ from spoken language.
Ranked #11 on
Sign Language Translation
on RWTH-PHOENIX-Weather 2014 T
3 code implementations • ACL 2018 • Albert Zeyer, Tamer Alkhouli, Hermann Ney
We compare the fast training and decoding speed of RETURNN of attention models for translation, due to fast CUDA LSTM kernels, and a fast pure TensorFlow beam search decoder.
14 code implementations • 8 May 2018 • Albert Zeyer, Kazuki Irie, Ralf Schlüter, Hermann Ney
Sequence-to-sequence attention-based models on subword units allow simple open-vocabulary end-to-end speech recognition.
Ranked #51 on
Speech Recognition
on LibriSpeech test-clean
(using extra training data)
no code implementations • WS 2017 • Jan-Thorsten Peter, Hermann Ney, Ond{\v{r}}ej Bojar, Ngoc-Quan Pham, Jan Niehues, Alex Waibel, Franck Burlot, Fran{\c{c}}ois Yvon, M{\=a}rcis Pinnis, Valters {\v{S}}ics, Jasmijn Bastings, Miguel Rios, Wilker Aziz, Philip Williams, Fr{\'e}d{\'e}ric Blain, Lucia Specia
no code implementations • ACL 2017 • Weiyue Wang, Tamer Alkhouli, Derui Zhu, Hermann Ney
Recently, the neural machine translation systems showed their promising performance and surpassed the phrase-based systems for most translation tasks.
no code implementations • CVPR 2017 • Oscar Koller, Sepehr Zargaran, Hermann Ney
This work presents an iterative re-alignment approach applicable to visual sequence labelling tasks such as gesture recognition, activity recognition and continuous sign language recognition.
no code implementations • 5 May 2017 • Patrick Doetsch, Pavel Golik, Hermann Ney
In this work we compare different batch construction methods for mini-batch training of recurrent neural networks.
no code implementations • WS 2015 • Markus Freitag, Jan-Thorsten Peter, Stephan Peitz, Minwei Feng, Hermann Ney
In this paper, we enhance the traditional confusion network system combination approach with an additional model trained by a neural network.
3 code implementations • 2 Aug 2016 • Patrick Doetsch, Albert Zeyer, Paul Voigtlaender, Ilya Kulikov, Ralf Schlüter, Hermann Ney
In this work we release our extensible and easily configurable neural network training software.
no code implementations • WS 2016 • Jan-Thorsten Peter, Tamer Alkhouli, Hermann Ney, Matthias Huck, Fabienne Braune, Alex Fraser, er, Ale{\v{s}} Tamchyna, Ond{\v{r}}ej Bojar, Barry Haddow, Rico Sennrich, Fr{\'e}d{\'e}ric Blain, Lucia Specia, Jan Niehues, Alex Waibel, Alex Allauzen, re, Lauriane Aufrant, Franck Burlot, Elena Knyazeva, Thomas Lavergne, Fran{\c{c}}ois Yvon, M{\=a}rcis Pinnis, Stella Frank
Ranked #12 on
Machine Translation
on WMT2016 English-Romanian
no code implementations • 22 Jun 2016 • Albert Zeyer, Patrick Doetsch, Paul Voigtlaender, Ralf Schlüter, Hermann Ney
On this task, we get our best result with an 8 layer bidirectional LSTM and we show that a pretraining scheme with layer-wise construction helps for deep LSTMs.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
no code implementations • CVPR 2016 • Oscar Koller, Hermann Ney, Richard Bowden
Furthermore, we demonstrate its use in continuous sign language recognition on two publicly available large sign language data sets, where it outperforms the current state-of-the-art by a large margin.
no code implementations • LREC 2014 • Jens Forster, Christoph Schmidt, Oscar Koller, Martin Bellgardt, Hermann Ney
This paper introduces the RWTH-PHOENIX-Weather 2014, a video-based, large vocabulary, German sign language corpus which has been extended over the last two years, tripling the size of the original corpus.
no code implementations • LREC 2012 • Saab Mansour, Hermann Ney
Next, we try a different strategy, where we combine the different segmentation methods rather than the different segmentation schemes.
no code implementations • LREC 2012 • Jens Forster, Christoph Schmidt, Thomas Hoyoux, Oscar Koller, Uwe Zelle, Justus Piater, Hermann Ney
This paper introduces the RWTH-PHOENIX-Weather corpus, a video-based, large vocabulary corpus of German Sign Language suitable for statistical sign language recognition and translation.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+5
no code implementations • NeurIPS 2011 • Simon Wiesler, Hermann Ney
Log-linear models are widely used probability models for statistical pattern recognition.