With the rapid development of Large Language Models (LLMs), it is crucial to have benchmarks which can evaluate the ability of LLMs on different domains. One common use of LLMs is performing tasks on scientific topics, such as writing algorithms, querying databases or giving mathematical proofs. Inspired by the way university students are evaluated on such tasks, in this paper, we propose SciEx - a benchmark consisting of university computer science exam questions, to evaluate LLMs’ ability on solving scientific tasks. SciEx is (1) multilingual, containing both English and German exams, and (2) multi-modal, containing questions that involve images, and (3) contains various types of freeform questions with different difficulty levels, due to the nature of university exams. We evaluate the performance of various state-of-the-art LLMs on our new benchmark. Since SciEx questions are freeform, it is not straightforward to evaluate LLM performance. Therefore, we provide human expert grading of the LLM outputs on SciEx. We show that the free-form exams in SciEx remain challenging for the current LLMs, where the best LLM only achieves 59.4% exam grade on average. We also provide detailed comparisons between LLM performance and student performance on SciEx. To enable future evaluation of new LLMs, we propose using LLM-as-a-judge to grade the LLM answers on SciEx. Our experiments show that, although they do not perform perfectly on solving the exams, LLMs are decent as graders, achieving 0.948 Pearson correlation with expert grading.
Direct speech translation (ST) models often struggle with rare words. Incorrect translation of these words can have severe consequences, impacting translation quality and user trust. While rare word translation is inherently challenging for neural models due to sparse learning signals, real-world scenarios often allow access to translations of past recordings on similar topics. To leverage these valuable resources, we propose a retrieval-and-demonstration approach to enhance rare word translation accuracy in direct ST models. First, we adapt existing ST models to incorporate retrieved examples for rare word translation, which allows the model to benefit from prepended examples, similar to in-context learning. We then develop a cross-modal (speech-to-speech, speech-to-text, text-to-text) retriever to locate suitable examples. We demonstrate that standard ST models can be effectively adapted to leverage examples for rare word translation, improving rare word translation accuracy over the baseline by 17.6% with gold examples and 8.5% with retrieved examples. Moreover, our speech-to-speech retrieval approach outperforms other modalities and exhibits higher robustness to unseen speakers. Our code is publicly available.
With the number of scientific papers published every year growing and current large language models (LLMs) showing state-of-the-art performance on natural language processing (NLP) tasks, we ask the question if LLMs could be utilized to answer questions on scientific papers.We investigate how well state-of-the-art large language models (LLMs) can answer questions on scientific paper by experimenting with long-context versions of the LLaMA 2 model and evaluating and training on the Qasper dataset.We analyze how well the LLMs handle longer papers and questions that can only be answered by accessing information from far out paragraphs. During our experiments, we see that the performance of these LLMs drops with growing length and position of relevant information.We employ different measures from simple prompts to chain-of-thought prompts and zero-shot usage to fine-tuning with QLoRA.While we still observe a performance loss with increased context length, our measures reduce the effects of this flaw, and we can achieve F1 scores similar to bigger models like GPT-4.
Western Armenian is a low-resource language spoken by the Armenian Diaspora residing in various places of the world. Although having content on the internet as well as a relatively rich literary heritage for a minority language, there is no data for the machine translation task and only a very limited amount of labeled data for other NLP tasks. In this work, we build the first machine translation system between Western Armenian and English. We explore different techniques for data collection and evaluate their impact in this very low-resource scenario. Then, we build the machine translation system while focusing on the possibilities of performing knowledge transfer from Eastern Armenian. The system is finetuned with the data collected for the first Western Armenian-English parallel corpus, which contains a total of approximately 147k sentence pairs, whose shareable part of 52k examples was made open-source. The best system through the experiments performs with a BLEU score of 29.8 while translating into English and 17 into Western Armenian.
Providing quality scores along with Machine Translation (MT) output, so-called reference-free Quality Estimation (QE), is crucial to inform users about the reliability of the translation. We propose a model-specific, unsupervised QE approach, termed kNN-QE, that extracts information from the MT model’s training data using k-nearest neighbors. Measuring the performance of model-specific QE is not straightforward, since they provide quality scores on their own MT output, thus cannot be evaluated using benchmark QE test sets containing human quality scores on premade MT output. Therefore, we propose an automatic evaluation method that uses quality scores from reference-based metrics as gold standard instead of human-generated ones. We are the first to conduct detailed analyses and conclude that this automatic method is sufficient, and the reference-based MetricX-23 is best for the task.
Large language models (LLMs) have demonstrated considerable success in various natural language processing tasks, but open-source LLMs have yet to attain state-of-the-art performance in Neural Machine Translation (NMT). Nevertheless, their significant performance in tasks demanding a broad understanding and contextual processing shows their potential for translation. To exploit these abilities, we investigate using LLMs for MT and explore recent parameter-efficient fine-tuning techniques. Surprisingly, our initial experiments found that fine-tuning with Q-LoRA for translation purposes led to performance improvements in terms of BLEU but degradation in COMET compared to in-context learning. To overcome this, we propose an alternative approach: adapting LLMs as Automatic Post-Editors (APE) rather than direct translators. Building on the ability of the LLM to handle long sequences, we also propose extending our approach to document-level translation. We show that leveraging Low-Rank-Adapter fine-tuning for APE can yield significant improvements across both sentence and document-level metrics while generalizing to out-of-domain data. Most notably, we achieve a state-of-the-art accuracy rate of 88.7% on the ContraPro test set, which assesses the model’s ability to resolve pronoun ambiguities when translating from English to German. Lastly, during manual post-editing for document-level translation, the source sentences are iteratively annotated, which can be used to refine further translations in the document. Here, we demonstrate that leveraging human corrections can significantly reduce the number of edits required for subsequent translations.
Finetuning pretrained models on downstream generation tasks often leads to catastrophic forgetting in zero-shot conditions. In this work, we focus on summarization and tackle the problem through the lens of language-independent representations. After training on monolingual summarization, we perform zero-shot transfer to new languages or language pairs. We first show naively finetuned models are highly language-specific in both output behavior and internal representations, resulting in poor zero-shot performance. Next, we propose query-key (QK) finetuning to decouple task-specific knowledge from the pretrained language generation abilities. Then, after showing downsides of the standard adversarial language classifier, we propose a balanced variant that more directly enforces language-agnostic representations. Moreover, our qualitative analyses show removing source language identity correlates to zero-shot summarization performance. Our code is openly available.
This paper reports on the shared tasks organized by the 21st IWSLT Conference. The shared tasks address 7 scientific challenges in spoken language translation: simultaneous and offline translation, automatic subtitling and dubbing, speech-to-speech translation, dialect and low-resource speech translation, and Indic languages. The shared tasks attracted 17 teams whose submissions are documented in 27 system papers. The growing interest towards spoken language translation is also witnessed by the constantly increasing number of shared task organizers and contributors to the overview paper, almost evenly distributed across industry and academia.
Large Language Models (LLMs) are currently under exploration for various tasks, including Automatic Speech Recognition (ASR), Machine Translation (MT), and even End-to-End Speech Translation (ST). In this paper, we present KIT’s offline submission in the constrained + LLM track by incorporating recently proposed techniques that can be added to any cascaded speech translation. Specifically, we integrate Mistral-7B into our system to enhance it in two ways. Firstly, we refine the ASR outputs by utilizing the N-best lists generated by our system and fine-tuning the LLM to predict the transcript accurately. Secondly, we refine the MT outputs at the document level by fine-tuning the LLM, leveraging both ASR and MT predictions to improve translation quality. We find that integrating the LLM into the ASR and MT systems results in an absolute improvement of 0.3% in Word Error Rate and 0.65% in COMET for tst2019 test set. In challenging test sets with overlapping speakers and background noise, we find that integrating LLM is not beneficial due to poor ASR performance. Here, we use ASR with chunked long-form decoding to improve context usage that may be unavailable when transcribing with Voice Activity Detection segmentation alone.
This paper presents KIT’s submissions to the IWSLT 2024 dialectal and low-resource track. In this work, we build systems for translating into English from speech in Maltese, Bemba, and two Arabic dialects Tunisian and North Levantine. Under the unconstrained condition, we leverage the pre-trained multilingual models by fine-tuning them for the target language pairs to address data scarcity problems in this track. We build cascaded and end-to-end speech translation systems for different language pairs and show the cascaded system brings slightly better overall performance. Besides, we find utilizing additional data resources boosts speech recognition performance but slightly harms machine translation performance in cascaded systems. Lastly, we show that Minimum Bayes Risk is effective in improving speech translation performance by combining the cascaded and end-to-end systems, bringing a consistent improvement of around 1 BLUE point.
Speech translation has witnessed significant progress driven by advancements in modeling techniques and the growing availability of training data. In this paper, we highlight recent advances in two ongoing research directions in ST: scaling the models to 1) many translation directions (multilingual ST) and 2) beyond the text output modality (multimodal ST). We structure this review by examining the sequential stages of a model’s development lifecycle: determining training resources, selecting model architecture, training procedures, evaluation metrics, and deployment considerations. We aim to highlight recent developments in each stage, with a particular focus on model architectures (dedicated speech translation models and LLM-based general-purpose model) and training procedures (task-specific vs. task-invariant approaches). Based on the reviewed advancements, we identify and discuss ongoing challenges within the field of speech translation.
Generating rare words is a challenging task for natural language processing in general and in speech translation (ST) specifically. This paper introduces a test suite prepared for the Offline ST shared task at IWSLT. In the test suite, corresponding rare words (i.e. named entities) were annotated on TED-Talks for English and German and the English side was made available to the participants together with some distractors (irrelevant named entities). Our evaluation checks the capabilities of ST systems to leverage the information in the contextual list of named entities and improve translation quality. Systems are ranked based on the recall and precision of named entities (separately on person, location, and organization names) in the translated texts. Our evaluation shows that using contextual information improves translation quality as well as the recall and precision of NEs. The recall of organization names in all submissions is the lowest of all categories with a maximum of 87.5 % confirming the difficulties of ST systems in dealing with names.
Customizing machine translation models to comply with desired attributes (e.g., formality or grammatical gender) is a well-studied topic. However, most current approaches rely on (semi-)supervised data with attribute annotations. This data scarcity bottlenecks democratizing such customization possibilities to a wider range of languages, particularly lower-resource ones. This gap is out of sync with recent progress in pretrained massively multilingual translation models. In response, we transfer the attribute controlling capabilities to languages without attribute-annotated data with an NLLB-200 model as a foundation. Inspired by techniques from controllable generation, we employ a gradient-based inference-time controller to steer the pretrained model. The controller transfers well to zero-shot conditions, as it is operates on pretrained multilingual representations and is attribute- rather than language-specific. With a comprehensive comparison to finetuning-based control, we demonstrate that, despite finetuning’s clear dominance in supervised settings, the gap to inference-time control closes when moving to zero-shot conditions, especially with new and distant target languages. The latter also shows stronger domain robustness. We further show that our inference-time control complements finetuning. Moreover, a human evaluation on a real low-resource language, Bengali, confirms our findings. Our code is in the supplementary material.
Diffusion models have recently shown great potential on many generative tasks.In this work, we explore diffusion models for machine translation (MT).We adapt two prominent diffusion-based text generation models, Diffusion-LM and DiffuSeq, to perform machine translation.As the diffusion models generate non-autoregressively (NAR),we draw parallels to NAR machine translation models.With a comparison to conventional Transformer-based translation models, as well as to the Levenshtein Transformer,an established NAR MT model,we show that the multimodality problem that limits NAR machine translation performance is also a challenge to diffusion models.We demonstrate that knowledge distillation from an autoregressive model improves the performance of diffusion-based MT.A thorough analysis on the translation quality of inputs of different lengths shows that the diffusion models struggle more on long-range dependencies than other models.
Recent advancements in NLP have resulted in models with specialized strengths, such as processing multimodal inputs or excelling in specific domains. However, real-world tasks, like multimodal translation, often require a combination of these strengths, such as handling both translation and image processing. While individual translation and vision models are powerful, they typically lack the ability to perform both tasks in a single system. Combining these models poses challenges, particularly due to differences in their vocabularies, which limit the effectiveness of traditional ensemble methods to post-generation techniques like N-best list re-ranking. In this work, we propose a novel zero-shot ensembling strategy that allows for the integration of different models during the decoding phase without the need for additional training. Our approach re-ranks beams during decoding by combining scores at the word level, using heuristics to predict when a word is completed. We demonstrate the effectiveness of this method in machine translation scenarios, showing that it enables the generation of translations that are both speech- and image-aware while also improving overall translation quality.
Human evaluation is a critical component in machine translation system development and has received much attention in text translation research. However, little prior work exists on the topic of human evaluation for speech translation, which adds additional challenges such as noisy data and segmentation mismatches. We take the first steps to fill this gap by conducting a comprehensive human evaluation of the results of several shared tasks from the last International Workshop on Spoken Language Translation (IWSLT 2023). We propose an effective evaluation strategy based on automatic resegmentation and direct assessment with segment context. Our analysis revealed that: 1) the proposed evaluation strategy is robust and scores well-correlated with other types of human judgements; 2) automatic metrics are usually, but not always, well-correlated with direct assessment scores; and 3) COMET as a slightly stronger automatic metric than chrF, despite the segmentation noise introduced by the resegmentation step systems. We release the collected human-annotated data in order to encourage further investigation.
Automatic Speech Recognition (ASR) can be a valuable tool to document endangered languages. However, building ASR tools for these languages poses several difficult research challenges, notably data scarcity. In this paper, we show the whole process of creating a useful ASR tool for language documentation scenarios. We publish the first speech corpus for Khinalug, an endangered language spoken in Northern Azerbaijan. The corpus consists of 2.67 hours of labeled data from recordings of spontaneous speech about various topics. As Khinalug is an extremely low-resource language, we investigate the benefits of multilingual models for self-supervised learning and supervised learning and achieve the performance of 6.65 Character Error Rate (CER) points and 25.53 Word Error Rate (WER) points. The benefits of multilingual models are further validated through experimentation with three additional under-resourced languages. Lastly, this work conducts quality assessments with linguists on new recordings to investigate the model’s usefulness in language documentation. We observe an evident degradation for new recordings, indicating the importance of enhancing model robustness. In addition, we find the inaudible content is the main cause of wrong ASR predictions, suggesting relating work on incorporating contextual information.
This paper reports on the shared tasks organized by the 20th IWSLT Conference. The shared tasks address 9 scientific challenges in spoken language translation: simultaneous and offline translation, automatic subtitling and dubbing, speech-to-speech translation, multilingual, dialect and low-resource speech translation, and formality control. The shared tasks attracted a total of 38 submissions by 31 teams. The growing interest towards spoken language translation is also witnessed by the constantly increasing number of shared task organizers and contributors to the overview paper, almost evenly distributed across industry and academia.
We present the ACL 60/60 evaluation sets for multilingual translation of ACL 2022 technical presentations into 10 target languages. This dataset enables further research into multilingual speech translation under realistic recording conditions with unsegmented audio and domain-specific terminology, applying NLP tools to text and speech in the technical domain, and evaluating and improving model robustness to diverse speaker demographics.
Many existing speech translation benchmarks focus on native-English speech in high-quality recording conditions, which often do not match the conditions in real-life use-cases. In this paper, we describe our speech translation system for the multilingual track of IWSLT 2023, which focuses on the translation of scientific conference talks. The test condition features accented input speech and terminology-dense contents. The tasks requires translation into 10 languages of varying amounts of resources. In absence of training data from the target domain, we use a retrieval-based approach (kNN-MT) for effective adaptation (+0.8 BLEU for speech translation). We also use adapters to easily integrate incremental training data from data augmentation, and show that it matches the performance of re-training. We observe that cascaded systems are more easily adaptable towards specific target domains, due to their separate modules. Our cascaded speech system outperforms its end-to-end counterpart on scientific talk translation, although their performance remains similar on TED talks.
In this paper, we describe our submission to the Simultaneous Track at IWSLT 2023. This year, we continue with the successful setup from the last year, however, we adopt the latest methods that further improve the translation quality. Additionally, we propose a novel online policy for attentional encoder-decoder models. The policy prevents the model to generate translation beyond the current speech input by using an auxiliary CTC output layer. We show that the proposed simultaneous policy can be applied to both streaming blockwise models and offline encoder-decoder models. We observe significant improvements in quality (up to 1.1 BLEU) and the computational footprint (up to 45% relative RTF).
The challenge of low-latency speech translation has recently draw significant interest in the research community as shown by several publications and shared tasks. Therefore, it is essential to evaluate these different approaches in realistic scenarios. However, currently only specific aspects of the systems are evaluated and often it is not possible to compare different approaches. In this work, we propose the first framework to perform and evaluate the various aspects of low-latency speech translation under realistic conditions. The evaluation is carried out in an end-to-end fashion. This includes the segmentation of the audio as well as the run-time of the different components. Secondly, we compare different approaches to low-latency speech translation using this framework. We evaluate models with the option to revise the output as well as methods with fixed output. Furthermore, we directly compare state-of-the-art cascaded as well as end-to-end systems. Finally, the framework allows to automatically evaluate the translation quality as well as latency and also provides a web interface to show the low-latency model outputs to the user.
Advancements in Neural Machine Translation (NMT) greatly benefit the software localization industry by decreasing the post-editing time of human annotators. Although the volume of the software being localized is growing significantly, techniques for improving NMT for user interface (UI) texts are lacking. These UI texts have different properties than other collections of texts, presenting unique challenges for NMT. For example, they are often very short, causing them to be ambiguous and needing additional context (button, title text, a table item, etc.) for disambiguation. However, no such UI data sets are readily available with contextual information for NMT models to exploit. This work aims to provide a first step in improving UI translations and highlight its challenges. To achieve this, we provide a novel multilingual UI corpus collection (∼ 1.3M for English ↔ German) with a targeted test set and analyze the limitations of state-of-the-art methods on this challenging task. Specifically, we present a targeted test set for disambiguation from English to German to evaluate reliably and emphasize UI translation challenges. Furthermore, we evaluate several state-of-the-art NMT techniques from domain adaptation and document-level NMT on this challenging task. All the scripts to replicate the experiments and data sets are available here.ˆ,
Neural machine translation (NMT) models often suffer from gender biases that harm users and society at large. In this work, we explore how bridging the gap between languages for which parallel data is not available affects gender bias in multilingual NMT, specifically for zero-shot directions. We evaluate translation between grammatical gender languages which requires preserving the inherent gender information from the source in the target language. We study the effect of encouraging language-agnostic hidden representations on models’ ability to preserve gender and compare pivot-based and zero-shot translation regarding the influence of the bridge language (participating in all language pairs during training) on gender preservation. We find that language-agnostic representations mitigate zero-shot models’ masculine bias, and with increased levels of gender inflection in the bridge language, pivoting surpasses zero-shot translation regarding fairer gender preservation for speaker-related gender agreement.
Quality Estimation (QE) is the task of predicting the quality of Machine Translation (MT) system output, without using any gold-standard translation references. State-of-the-art QE models are supervised: they require human-labeled quality of some MT system output on some datasets for training, making them domain-dependent and MT-system-dependent. There has been research on unsupervised QE, which requires glass-box access to the MT systems, or parallel MT data to generate synthetic errors for training QE models. In this paper, we present Perturbation-based QE - a word-level Quality Estimation approach that works simply by analyzing MT system output on perturbed input source sentences. Our approach is unsupervised, explainable, and can evaluate any type of blackbox MT systems, including the currently prominent large language models (LLMs) with opaque internal processes. For language directions with no labeled QE data, our approach has similar or better performance than the zero-shot supervised approach on the WMT21 shared task. Our approach is better at detecting gender bias and word-sense-disambiguation errors in translation than supervised QE, indicating its robustness to out-of-domain usage. The performance gap is larger when detecting errors on a nontraditional translation-prompting LLM, indicating that our approach is more generalizable to different MT systems. We give examples demonstrating our approach’s explainability power, where it shows which input source words have influence on a certain MT output word.
In many humanitarian scenarios, translation into severely low resource languages often does not require a universal translation engine, but a dedicated text-specific translation engine. For example, healthcare records, hygienic procedures, government communication, emergency procedures and religious texts are all limited texts. While generic translation engines for all languages do not exist, translation of multilingually known limited texts into new, endangered languages may be possible and reduce human translation effort. We attempt to leverage translation resources from rich resource languages to efficiently produce best possible translation quality for well known texts, which is available in multiple languages, in a new, severely low resource language. We examine two approaches: 1.) best selection of seed sentences to jump start translations in a new language in view of best generalization to the remainder of a larger targeted text(s), and 2.) we adapt large general multilingual translation engines from many other languages to focus on a specific text in a new, unknown language. We find that adapting large pretrained multilingual models to the domain/text first and then to the severely low resource language works best. If we also select a best set of seed sentences, we can improve average chrF performance on new test languages from a baseline of 21.9 to 50.7, while reducing the number of seed sentences to only ∼1,000 in the new, unknown language.
Recently, we have seen an increasing interest in the area of speech-to-text translation. This has led to astonishing improvements in this area. In contrast, the activities in the area of speech-to-speech translation is still limited, although it is essential to overcome the language barrier. We believe that one of the limiting factors is the availability of appropriate training data. We address this issue by creating LibriS2S, to our knowledge the first publicly available speech-to-speech training corpus between German and English. For this corpus, we used independently created audio for German and English leading to an unbiased pronunciation of the text in both languages. This allows the creation of a new text-to-speech and speech-to-speech translation model that directly learns to generate the speech signal based on the pronunciation of the source language. Using this created corpus, we propose Text-to-Speech models based on the example of the recently proposed FastSpeech 2 model that integrates source language information. We do this by adapting the model to take information such as the pitch, energy or transcript from the source speech as additional input.
The cornerstone of multilingual neural translation is shared representations across languages. Given the theoretically infinite representation power of neural networks, semantically identical sentences are likely represented differently. While representing sentences in the continuous latent space ensures expressiveness, it introduces the risk of capturing of irrelevant features which hinders the learning of a common representation. In this work, we discretize the encoder output latent space of multilingual models by assigning encoder states to entries in a codebook,which in effect represents source sentences in a new artificial language. This discretization process not only offers a new way to interpret the otherwise black-box model representations,but, more importantly, gives potential for increasing robustness in unseen testing conditions. We validate our approach on large-scale experiments with realistic data volumes and domains. When tested in zero-shot conditions, our approach is competitive with two strong alternatives from the literature. We also use the learned artificial language to analyze model behavior, and discover that using a similar bridge language increases knowledge-sharing among the remaining languages.
The evaluation campaign of the 19th International Conference on Spoken Language Translation featured eight shared tasks: (i) Simultaneous speech translation, (ii) Offline speech translation, (iii) Speech to speech translation, (iv) Low-resource speech translation, (v) Multilingual speech translation, (vi) Dialect speech translation, (vii) Formality control for speech translation, (viii) Isometric speech translation. A total of 27 teams participated in at least one of the shared tasks. This paper details, for each shared task, the purpose of the task, the data that were released, the evaluation metrics that were applied, the submissions that were received and the results that were achieved.
Pretrained models in acoustic and textual modalities can potentially improve speech translation for both Cascade and End-to-end approaches. In this evaluation, we aim at empirically looking for the answer by using the wav2vec, mBART50 and DeltaLM models to improve text and speech translation models. The experiments showed that the presence of these models together with an advanced audio segmentation method results in an improvement over the previous end-to-end system by up to 7 BLEU points. More importantly, the experiments showed that given enough data and modeling capacity to overcome the training difficulty, we can outperform even very competitive Cascade systems. In our experiments, this gap can be as large as 2.0 BLEU points, the same gap that the Cascade often led over the years.
In this paper, we describe our submission to the Simultaneous Speech Translation at IWSLT 2022. We explore strategies to utilize an offline model in a simultaneous setting without the need to modify the original model. In our experiments, we show that our onlinization algorithm is almost on par with the offline setting while being 3x faster than offline in terms of latency on the test set. We also show that the onlinized offline model outperforms the best IWSLT2021 simultaneous system in medium and high latency regimes and is almost on par in the low latency regime. We make our system publicly available.
Multilingual neural machine translation has shown the capability of directly translating between language pairs unseen in training, i.e. zero-shot translation. Despite being conceptually attractive, it often suffers from low output quality. The difficulty of generalizing to new translation directions suggests the model representations are highly specific to those language pairs seen in training. We demonstrate that a main factor causing the language-specific representations is the positional correspondence to input tokens. We show that this can be easily alleviated by removing residual connections in an encoder layer. With this modification, we gain up to 18.5 BLEU points on zero-shot translation while retaining quality on supervised directions. The improvements are particularly prominent between related languages, where our proposed model outperforms pivot-based translation. Moreover, our approach allows easy integration of new languages, which substantially expands translation coverage. By thorough inspections of the hidden layer outputs, we show that our approach indeed leads to more language-independent representations.
The evaluation campaign of the International Conference on Spoken Language Translation (IWSLT 2021) featured this year four shared tasks: (i) Simultaneous speech translation, (ii) Offline speech translation, (iii) Multilingual speech translation, (iv) Low-resource speech translation. A total of 22 teams participated in at least one of the tasks. This paper describes each shared task, data and evaluation metrics, and reports results of the received submissions.
This paper describes Maastricht University’s participation in the IWSLT 2021 multilingual speech translation track. The task in this track is to build multilingual speech translation systems in supervised and zero-shot directions. Our primary system is an end-to-end model that performs both speech transcription and translation. We observe that the joint training for the two tasks is complementary especially when the speech translation data is scarce. On the source and target side, we use data augmentation and pseudo-labels respectively to improve the performance of our systems. We also introduce an ensembling technique that consistently improves the quality of transcriptions and translations. The experiments show that the end-to-end system is competitive with its cascaded counterpart especially in zero-shot conditions.
While recent advances in deep learning led to significant improvements in machine translation, neural machine translation is often still not able to continuously adapt to the environment. For humans, as well as for machine translation, bilingual dictionaries are a promising knowledge source to continuously integrate new knowledge. However, their exploitation poses several challenges: The system needs to be able to perform one-shot learning as well as model the morphology of source and target language. In this work, we proposed an evaluation framework to assess the ability of neural machine translation to continuously learn new phrases. We integrate one-shot learning methods for neural machine translation with different word representations and show that it is important to address both in order to successfully make use of bilingual dictionaries. By addressing both challenges we are able to improve the ability to translate new, rare words and phrases from 30% to up to 70%. The correct lemma is even generated by more than 90%.
Speech translation is the translation of speech in one language typically to text in another, traditionally accomplished through a combination of automatic speech recognition and machine translation. Speech translation has attracted interest for many years, but the recent successful applications of deep learning to both individual tasks have enabled new opportunities through joint modeling, in what we today call ‘end-to-end speech translation.’ In this tutorial we introduce the techniques used in cutting-edge research on speech translation. Starting from the traditional cascaded approach, we give an overview on data sources and model architectures to achieve state-of-the art performance with end-to-end speech translation for both high- and low-resource languages. In addition, we discuss methods to evaluate analyze the proposed solutions, as well as the challenges faced when applying speech translation models for real-world applications.
We present our development of the multilingual machine translation system for the large-scale multilingual machine translation task at WMT 2021. Starting form the provided baseline system, we investigated several techniques to improve the translation quality on the target subset of languages. We were able to significantly improve the translation quality by adapting the system towards the target subset of languages and by generating synthetic data using the initial model. Techniques successfully applied in zero-shot multilingual machine translation (e.g. similarity regularizer) only had a minor effect on the final translation performance.
Unsupervised Neural Machine translation (UNMT) is beneficial especially for under-resourced languages such as from the Dravidian family. They learn to translate between the source and target, relying solely on only monolingual corpora. However, UNMT systems fail in scenarios that occur often when dealing with low resource languages. Recent works have achieved state-of-the-art results by adding auxiliary parallel data with similar languages. In this work, we focus on unsupervised translation between English and Kannada by using limited amounts of auxiliary data between English and other Dravidian languages. We show that transliteration is essential in unsupervised translation between Dravidian languages, as they do not share a common writing system. We explore several model architectures that use the auxiliary data in order to maximize knowledge sharing and enable UNMT for dissimilar language pairs. We show from our experiments it is crucial for Kannada and reference languages to be similar. Further, we propose a method to measure language similarity to choose the most beneficial reference languages.
The evaluation campaign of the International Conference on Spoken Language Translation (IWSLT 2020) featured this year six challenge tracks: (i) Simultaneous speech translation, (ii) Video speech translation, (iii) Offline speech translation, (iv) Conversational speech translation, (v) Open domain translation, and (vi) Non-native speech translation. A total of teams participated in at least one of the tracks. This paper introduces each track’s goal, data and evaluation metrics, and reports the results of the received submissions.
Automatic speech recognition (ASR) systems are primarily evaluated on transcription accuracy. However, in some use cases such as subtitling, verbatim transcription would reduce output readability given limited screen size and reading time. Therefore, this work focuses on ASR with output compression, a task challenging for supervised approaches due to the scarcity of training data. We first investigate a cascaded system, where an unsupervised compression model is used to post-edit the transcribed speech. We then compare several methods of end-to-end speech recognition under output length constraints. The experiments show that with limited data far less than needed for training a model from scratch, we can adapt a Transformer-based ASR model to incorporate both transcription and compression capabilities. Furthermore, the best performance in terms of WER and ROUGE scores is achieved by explicitly modeling the length constraints within the end-to-end ASR system.
Speech translation has traditionally been approached through cascaded models consisting of a speech recognizer trained on a corpus of transcribed speech, and a machine translation system trained on parallel texts. Several recent works have shown the feasibility of collapsing the cascade into a single, direct model that can be trained in an end-to-end fashion on a corpus of translated speech. However, experiments are inconclusive on whether the cascade or the direct model is stronger, and have only been conducted under the unrealistic assumption that both are trained on equal amounts of data, ignoring other available speech recognition and machine translation corpora. In this paper, we demonstrate that direct speech translation models require more data to perform well than cascaded models, and although they allow including auxiliary data through multi-task training, they are poor at exploiting such data, putting them at a severe disadvantage. As a remedy, we propose the use of end- to-end trainable models with two attention mechanisms, the first establishing source speech to source text alignments, the second modeling source to target text alignment. We show that such models naturally decompose into multi-task–trainable recognition and translation tasks and propose an attention-passing technique that alleviates error propagation issues in a previous formulation of a model with two attention stages. Our proposed model outperforms all examined baselines and is able to exploit auxiliary training data much more effectively than direct attentional models.
The state-of-the-art neural network architectures make it possible to create spoken language understanding systems with high quality and fast processing time. One major challenge for real-world applications is the high latency of these systems caused by triggered actions with high executions times. If an action can be separated into subactions, the reaction time of the systems can be improved through incremental processing of the user utterance and starting subactions while the utterance is still being uttered. In this work, we present a model-agnostic method to achieve high quality in processing incrementally produced partial utterances. Based on clean and noisy versions of the ATIS dataset, we show how to create datasets with our method to create low-latency natural language understanding components. We get improvements of up to 47.91 absolute percentage points in the metric F1-score.
The IWSLT 2019 evaluation campaign featured three tasks: speech translation of (i) TED talks and (ii) How2 instructional videos from English into German and Portuguese, and (iii) text translation of TED talks from English into Czech. For the first two tasks we encouraged submissions of end- to-end speech-to-text systems, and for the second task participants could also use the video as additional input. We received submissions by 12 research teams. This overview provides detailed descriptions of the data and evaluation conditions of each task and reports results of the participating systems.
This paper describes KIT’s submission to the IWSLT 2019 Speech Translation task on two sub-tasks corresponding to two different datasets. We investigate different end-to-end architectures for the speech recognition module, including our new transformer-based architectures. Overall, our modules in the pipe-line are based on the transformer architecture which has recently achieved great results in various fields. In our systems, using transformer is also advantageous compared to traditional hybrid systems in term of simplicity while still having competent results.
An important concern in training multilingual neural machine translation (NMT) is to translate between language pairs unseen during training, i.e zero-shot translation. Improving this ability kills two birds with one stone by providing an alternative to pivot translation which also allows us to better understand how the model captures information between languages. In this work, we carried out an investigation on this capability of the multilingual NMT models. First, we intentionally create an encoder architecture which is independent with respect to the source language. Such experiments shed light on the ability of NMT encoders to learn multilingual representations, in general. Based on such proof of concept, we were able to design regularization methods into the standard Transformer model, so that the whole architecture becomes more robust in zero-shot conditions. We investigated the behaviour of such models on the standard IWSLT 2017 multilingual dataset. We achieved an average improvement of 2.23 BLEU points across 12 language pairs compared to the zero-shot performance of a state-of-the-art multilingual system. Additionally, we carry out further experiments in which the effect is confirmed even for language pairs with multiple intermediate pivots.
Recently, significant improvements have been achieved in various natural language processing tasks using neural sequence-to-sequence models. While aiming for the best generation quality is important, ultimately it is also necessary to develop models that can assess the quality of their output. In this work, we propose to use the similarity between training and test conditions as a measure for models’ confidence. We investigate methods solely using the similarity as well as methods combining it with the posterior probability. While traditionally only target tokens are annotated with confidence measures, we also investigate methods to annotate source tokens with confidence. By learning an internal alignment model, we can significantly improve confidence projection over using state-of-the-art external alignment tools. We evaluate the proposed methods on downstream confidence estimation for machine translation (MT). We show improvements on segment-level confidence estimation as well as on confidence estimation for source tokens. In addition, we show that the same methods can also be applied to other tasks using sequence-to-sequence models. On the automatic speech recognition (ASR) task, we are able to find 60% of the errors by looking at 20% of the data.
The International Workshop of Spoken Language Translation (IWSLT) 2018 Evaluation Campaign featured two tasks: low-resource machine translation and speech translation. In the first task, manually transcribed speech had to be translated from Basque to English. Since this translation direction is a under-resourced language pair, participants were encouraged to use additional parallel data from related languages. In the second task, participants had to translate English audio into German text with a full speech-translation system. In the baseline condition, participants were free to use composite architectures, while in the end-to-end condition they were restricted to use a single model for the task. This year, eight research groups took part in the low-resource machine translation task and nine in the speech translation task.
This paper describes KIT’s submission to the IWSLT 2018 Translation task. We describe a system participating in the baseline condition and a system participating in the end-to-end condition. The baseline system is a cascade of an ASR system, a system to segment the ASR output and a neural machine translation system. We investigate the combination of different ASR systems. For the segmentation and machine translation components, we focused on transformer-based architectures.
In today’s globalized world we have the ability to communicate with people across the world. However, in many situations the language barrier still presents a major issue. For example, many foreign students coming to KIT to study are initially unable to follow a lecture in German. Therefore, we offer an automatic simultaneous interpretation service for students. To fulfill this task, we have developed a low-latency translation system that is adapted to lectures and covers several language pairs. While the switch from traditional Statistical Machine Translation to Neural Machine Translation (NMT) significantly improved performance, to integrate NMT into the speech translation framework required several adjustments. We have addressed the run-time constraints and different types of input. Furthermore, we utilized one-shot learning to easily add new topic-specific terms to the system. Besides better performance, NMT also enabled us increase our covered languages through multilingual NMT. % Combining these techniques, we are able to provide an adapted speech translation system for several European languages.
Deep learning models are often not easily adaptable to new tasks and require task-specific adjustments. The differentiable neural computer (DNC), a memory-augmented neural network, is designed as a general problem solver which can be used in a wide range of tasks. But in reality, it is hard to apply this model to new tasks. We analyze the DNC and identify possible improvements within the application of question answering. This motivates a more robust and scalable DNC (rsDNC). The objective precondition is to keep the general character of this model intact while making its application more reliable and speeding up its required training time. The rsDNC is distinguished by a more robust training, a slim memory unit and a bidirectional architecture. We not only achieve new state-of-the-art performance on the bAbI task, but also minimize the performance variance between different initializations. Furthermore, we demonstrate the simplified applicability of the rsDNC to new tasks with passable results on the CNN RC task without adaptions.
Neural machine translation (NMT) has significantly improved the quality of automatic translation models. One of the main challenges in current systems is the translation of rare words. We present a generic approach to address this weakness by having external models annotate the training data as Experts, and control the model-expert interaction with a pointer network and reinforcement learning. Our experiments using phrase-based models to simulate Experts to complement neural machine translation models show that the model can be trained to copy the annotations into the output consistently. We demonstrate the benefit of our proposed framework in outof domain translation scenarios with only lexical resources, improving more than 1.0 BLEU point in both translation directions English-Spanish and German-English.
We present our experiments in the scope of the news translation task in WMT 2018, in directions: English→German. The core of our systems is the encoder-decoder based neural machine translation models using the transformer architecture. We enhanced the model with a deeper architecture. By using techniques to limit the memory consumption, we were able to train models that are 4 times larger on one GPU and improve the performance by 1.2 BLEU points. Furthermore, we performed sentence selection for the newly available ParaCrawl corpus. Thereby, we could improve the effectiveness of the corpus by 0.5 BLEU points.
The input to a neural sequence-to-sequence model is often determined by an up-stream system, e.g. a word segmenter, part of speech tagger, or speech recognizer. These up-stream models are potentially error-prone. Representing inputs through word lattices allows making this uncertainty explicit by capturing alternative sequences and their posterior probabilities in a compact form. In this work, we extend the TreeLSTM (Tai et al., 2015) into a LatticeLSTM that is able to consume word lattices, and can be used as encoder in an attentional encoder-decoder model. We integrate lattice posterior scores into this architecture by extending the TreeLSTM’s child-sum and forget gates and introducing a bias term into the attention mechanism. We experiment with speech translation lattices and report consistent improvements over baselines that translate either the 1-best hypothesis or the lattice without posterior scores.
The IWSLT 2017 evaluation campaign has organised three tasks. The Multilingual task, which is about training machine translation systems handling many-to-many language directions, including so-called zero-shot directions. The Dialogue task, which calls for the integration of context information in machine translation, in order to resolve anaphoric references that typically occur in human-human dialogue turns. And, finally, the Lecture task, which offers the challenge of automatically transcribing and translating real-life university lectures. Following the tradition of these reports, we will described all tasks in detail and present the results of all runs submitted by their participants.
In this paper, we present KIT’s multilingual neural machine translation (NMT) systems for the IWSLT 2017 evaluation campaign machine translation (MT) and spoken language translation (SLT) tasks. For our MT task submissions, we used our multi-task system, modified from a standard attentional neural machine translation framework, instead of building 20 individual NMT systems. We investigated different architectures as well as different data corpora in training such a multilingual system. We also suggested an effective adaptation scheme for multilingual systems which brings great improvements compared to monolingual systems. For the SLT track, in addition to a monolingual neural translation system used to generate correct punctuations and true cases of the data prior to training our multilingual system, we introduced a noise model in order to make our system more robust. Results show that our novel modifications improved our systems considerably on all tasks.
Punctuation and segmentation is crucial in spoken language translation, as it has a strong impact to translation performance. However, the impact of rare or unknown words in the performance of punctuation and segmentation insertion has not been thoroughly studied. In this work, we simulate various degrees of domain-match in testing scenario and investigate their impact to the punctuation insertion task. We explore three rare word generalizing schemes using part-of-speech (POS) tokens. Experiments show that generalizing rare and unknown words greatly improves the punctuation insertion performance, reaching up to 8.8 points of improvement in F-score when applied to the out-of-domain test scenario. We show that this improvement in punctuation quality has a positive impact on a following machine translation (MT) performance, improving it by 2 BLEU points.
Translating noisy inputs, such as the output of a speech recognizer, is a difficult but important challenge for neural machine translation. One way to increase robustness of neural models is by introducing artificial noise to the training data. In this paper, we experiment with appropriate forms of such noise, exploring a middle ground between general-purpose regularizers and highly task-specific forms of noise induction. We show that with a simple generative noise model, moderate gains can be achieved in translating erroneous speech transcripts, provided that type and amount of noise are properly calibrated. The optimal amount of noise at training time is much smaller than the amount of noise in our test data, indicating limitations due to trainability issues. We note that unlike our baseline model, models trained on noisy data are able to generate outputs of proper length even for noisy inputs, while gradually reducing output length for higher amount of noise, as might also be expected from a human translator. We discuss these findings in details and give suggestions for future work.
In this paper, we proposed two strategies which can be applied to a multilingual neural machine translation system in order to better tackle zero-shot scenarios despite not having any parallel corpus. The experiments show that they are effective in terms of both performance and computing resources, especially in multilingual translation of unbalanced data in real zero-resourced condition when they alleviate the language bias problem.
In this paper, we offer an in-depth analysis about the modeling and search performance. We address the question if a more complex search algorithm is necessary. Furthermore, we investigate the question if more complex models which might only be applicable during rescoring are promising. By separating the search space and the modeling using n-best list reranking, we analyze the influence of both parts of an NMT system independently. By comparing differently performing NMT systems, we show that the better translation is already in the search space of the translation systems with less performance. This results indicate that the current search algorithms are sufficient for the NMT systems. Furthermore, we could show that even a relatively small n-best list of 50 hypotheses already contain notably better translations.
Recently, the development of neural machine translation (NMT) has significantly improved the translation quality of automatic machine translation. While most sentences are more accurate and fluent than translations by statistical machine translation (SMT)-based systems, in some cases, the NMT system produces translations that have a completely different meaning. This is especially the case when rare words occur. When using statistical machine translation, it has already been shown that significant gains can be achieved by simplifying the input in a preprocessing step. A commonly used example is the pre-reordering approach. In this work, we used phrase-based machine translation to pre-translate the input into the target language. Then a neural machine translation system generates the final hypothesis using the pre-translation. Thereby, we use either only the output of the phrase-based machine translation (PBMT) system or a combination of the PBMT output and the source sentence. We evaluate the technique on the English to German translation task. Using this approach we are able to outperform the PBMT system as well as the baseline neural MT system by up to 2 BLEU points. We analyzed the influence of the quality of the initial system on the final result.
Evaluating the quality of output from language processing systems such as machine translation or speech recognition is an essential step in ensuring that they are sufficient for practical use. However, depending on the practical requirements, evaluation approaches can differ strongly. Often, reference-based evaluation measures (such as BLEU or WER) are appealing because they are cheap and allow rapid quantitative comparison. On the other hand, practitioners often focus on manual evaluation because they must deal with frequently changing domains and quality standards requested by customers, for which reference-based evaluation is insufficient or not possible due to missing in-domain reference data (Harris et al., 2016). In this paper, we attempt to bridge this gap by proposing a framework for lightly supervised quality estimation. We collect manually annotated scores for a small number of segments in a test corpus or document, and combine them with automatically predicted quality scores for the remaining segments to predict an overall quality estimate. An evaluation shows that our framework estimates quality more reliably than using fully automatic quality estimation approaches, while keeping annotation effort low by not requiring full references to be available for the particular domain.
The IWSLT 2016 Evaluation Campaign featured two tasks: the translation of talks and the translation of video conference conversations. While the first task extends previously offered tasks with talks from a different source, the second task is completely new. For both tasks, three tracks were organised: automatic speech recognition (ASR), spoken language translation (SLT), and machine translation (MT). Main translation directions that were offered are English to/from German and English to French. Additionally, the MT track included English to/from Arabic and Czech, as well as French to English. We received this year run submissions from 11 research labs. All runs were evaluated with objective metrics, while submissions for two of the MT talk tasks were also evaluated with human post-editing. Results of the human evaluation show improvements over the best submissions of last year.
Neural models have recently shown big improvements in the performance of phrase-based machine translation. Recurrent language models, in particular, have been a great success due to their ability to model arbitrary long context. In this work, we integrate global semantic information extracted from large encyclopedic sources into neural network language models. We integrate semantic word classes extracted from Wikipedia and sentence level topic information into a recurrent neural network-based language model. The new resulting models exhibit great potential in alleviating data sparsity problems with the additional knowledge provided. This approach of integrating global information is not restricted to language modeling but can also be easily applied to any model that profits from context or further data resources, e.g. neural machine translation. Using this model has improved rescoring quality of a state-of-the-art phrase-based translation system by 0.84 BLEU points. We performed experiments on two language pairs.
In this paper, we present our first attempts in building a multilingual Neural Machine Translation framework under a unified approach in which the information shared among languages can be helpful in the translation of individual language pairs. We are then able to employ attention-based Neural Machine Translation for many-to-many multilingual translation tasks. Our approach does not require any special treatment on the network architecture and it allows us to learn minimal number of free parameters in a standard way of training. Our approach has shown its effectiveness in an under-resourced translation scenario with considerable improvements up to 2.6 BLEU points. In addition, we point out a novel way to make use of monolingual data with Neural Machine Translation using the same approach with a 3.15-BLEU-score gain in IWSLT’16 English→German translation task.
In this paper, we investigate a multilingual approach for speech disfluency removal. A major challenge of this task comes from the costly nature of disfluency annotation. Motivated by the fact that speech disfluencies are commonly observed throughout different languages, we investigate the potential of multilingual disfluency modeling. We suggest that learning a joint representation of the disfluencies in multiple languages can be a promising solution to the data sparsity issue. In this work, we utilize a multilingual neural machine translation system, where a disfluent speech transcript is directly transformed into a cleaned up text. Disfluency removal experiments on English and German speech transcripts show that multilingual disfluency modeling outperforms the single language systems. In a following experiment, we show that the improvements are also observed in a downstream application using the disfluency-removed transcripts as input.
In this paper, we present the KIT systems of the IWSLT 2016 machine translation evaluation. We participated in the machine translation (MT) task as well as the spoken language language translation (SLT) track for English→German and German→English translation. We use attentional neural machine translation (NMT) for all our submissions. We investigated different methods to adapt the system using small in-domain data as well as methods to train the system on these small corpora. In addition, we investigated methods to combine NMT systems that encode the input as well as the output differently. We combine systems using different vocabularies, reverse translation systems, multi-source translation system. In addition, we used pre-translation systems that facilitate phrase-based machine translation systems. Results show that applying domain adaptation and ensemble technique brings a crucial improvement of 3-4 BLEU points over the baseline system. In addition, system combination using n-best lists yields further 1-2 BLEU points.
The paper overviews the 11th evaluation campaign organized by the IWSLT workshop. The 2014 evaluation offered multiple tracks on lecture transcription and translation based on the TED Talks corpus. In particular, this year IWSLT included three automatic speech recognition tracks, on English, German and Italian, five speech translation tracks, from English to French, English to German, German to English, English to Italian, and Italian to English, and five text translation track, also from English to French, English to German, German to English, English to Italian, and Italian to English. In addition to the official tracks, speech and text translation optional tracks were offered, globally involving 12 other languages: Arabic, Spanish, Portuguese (B), Hebrew, Chinese, Polish, Persian, Slovenian, Turkish, Dutch, Romanian, Russian. Overall, 21 teams participated in the evaluation, for a total of 76 primary runs submitted. Participants were also asked to submit runs on the 2013 test set (progress test set), in order to measure the progress of systems with respect to the previous year. All runs were evaluated with objective metrics, and submissions for two of the official text translation tracks were also evaluated with human post-editing.
EU-BRIDGE is a European research project which is aimed at developing innovative speech translation technology. One of the collaborative efforts within EU-BRIDGE is to produce joint submissions of up to four different partners to the evaluation campaign at the 2014 International Workshop on Spoken Language Translation (IWSLT). We submitted combined translations to the German→English spoken language translation (SLT) track as well as to the German→English, English→German and English→French machine translation (MT) tracks. In this paper, we present the techniques which were applied by the different individual translation systems of RWTH Aachen University, the University of Edinburgh, Karlsruhe Institute of Technology, and Fondazione Bruno Kessler. We then show the combination approach developed at RWTH Aachen University which combined the individual systems. The consensus translations yield empirical gains of up to 2.3 points in BLEU and 1.2 points in TER compared to the best individual system.
In this paper, we present the KIT systems participating in the TED translation tasks of the IWSLT 2014 machine translation evaluation. We submitted phrase-based translation systems for all three official directions, namely English→German, German→English, and English→French, as well as for the optional directions English→Chinese and English→Arabic. For the official directions we built systems both for the machine translation as well as the spoken language translation track. This year we improved our systems’ performance over last year through n-best list rescoring using neural network-based translation and language models and novel preordering rules based on tree information of multiple syntactic levels. Furthermore, we could successfully apply a novel phrase extraction algorithm and transliteration of unknown words for Arabic. We also submitted a contrastive system for German→English built with stemmed German adjectives. For the SLT tracks, we used a monolingual translation system to translate the lowercased ASR hypotheses with all punctuation stripped to truecased, punctuated output as a preprocessing step to our usual translation system.
Translating meetings presents a challenge since multi-speaker speech shows a variety of disfluencies. In this paper we investigate the importance of transforming speech into well-written input prior to translating multi-party meetings. We first analyze the characteristics of this data and establish oracle scores. Sentence segmentation and punctuation are performed using a language model, turn information, or a monolingual translation system. Disfluencies are removed by a CRF model trained on in-domain and out-of-domain data. For comparison, we build a combined CRF model for punctuation insertion and disfluency removal. By applying these models, multi-party meetings are transformed into fluent input for machine translation. We evaluate the models with regard to translation performance and are able to achieve an improvement of 2.1 to 4.9 BLEU points depending on the availability of turn information.
In this paper we combine the advantages of a model using global source sentence contexts, the Discriminative Word Lexicon, and neural networks. By using deep neural networks instead of the linear maximum entropy model in the Discriminative Word Lexicon models, we are able to leverage dependencies between different source words due to the non-linearity. Furthermore, the models for different target words can share parameters and therefore data sparsity problems are effectively reduced. By using this approach in a state-of-the-art translation system, we can improve the performance by up to 0.5 BLEU points for three different language pairs on the TED translation task.
Word reordering is a difficult task for translation. Common automatic metrics such as BLEU have problems reflecting improvements in target language word order. However, it is a crucial aspect for humans when deciding on translation quality. This paper presents a detailed analysis of a structure-aware reordering approach applied in a German-to-English phrase-based machine translation system. We compare the translation outputs of two translation systems applying reordering rules based on parts-of-speech and syntax trees on a sentence-by-sentence basis. For each sentence-pair we examine the global translation performance and classify local changes in the translated sentences. This analysis is applied to three data sets representing different genres. While the improvement in BLEU differed substantially between the data sets, the manual evaluation showed that both global translation performance as well as individual types of improvements and degradations exhibit a similar behavior throughout the three data sets. We have observed that for 55-64% of the sentences with different translations, the translation produced using the tree-based reordering was considered to be the better translation. As intended by the investigated reordering model, most improvements are achieved by improving the position of the verb or being able to translate a verb that could not be translated before.
This paper presents two improvements of language models based on Restricted Boltzmann Machine (RBM) for large machine translation tasks. In contrast to other continuous space approach, RBM based models can easily be integrated into the decoder and are able to directly learn a hidden representation of the n-gram. Previous work on RBM-based language models do not use a shared word representation and therefore, they might suffer of a lack of generalization for larger contexts. Moreover, since the training step is very time consuming, they are only used for quite small copora. In this work we add a shared word representation for the RBM-based language model by factorizing the weight matrix. In addition, we propose an efficient and tailored sampling algorithm that allows us to drastically speed up the training process. Experiments are carried out on two German to English translation tasks and the results show that the training time could be reduced by a factor of 10 without any drop in performance. Furthermore, the RBM-based model can also be trained on large size corpora.
The paper overviews the tenth evaluation campaign organized by the IWSLT workshop. The 2013 evaluation offered multiple tracks on lecture transcription and translation based on the TED Talks corpus. In particular, this year IWSLT included two automatic speech recognition tracks, on English and German, three speech translation tracks, from English to French, English to German, and German to English, and three text translation track, also from English to French, English to German, and German to English. In addition to the official tracks, speech and text translation optional tracks were offered involving 12 other languages: Arabic, Spanish, Portuguese (B), Italian, Chinese, Polish, Persian, Slovenian, Turkish, Dutch, Romanian, Russian. Overall, 18 teams participated in the evaluation for a total of 217 primary runs submitted. All runs were evaluated with objective metrics on a current test set and two progress test sets, in order to compare the progresses against systems of the previous years. In addition, submissions of one of the official machine translation tracks were also evaluated with human post-editing.
EU-BRIDGE1 is a European research project which is aimed at developing innovative speech translation technology. This paper describes one of the collaborative efforts within EUBRIDGE to further advance the state of the art in machine translation between two European language pairs, English→French and German→English. Four research institutions involved in the EU-BRIDGE project combined their individual machine translation systems and participated with a joint setup in the machine translation track of the evaluation campaign at the 2013 International Workshop on Spoken Language Translation (IWSLT). We present the methods and techniques to achieve high translation quality for text translation of talks which are applied at RWTH Aachen University, the University of Edinburgh, Karlsruhe Institute of Technology, and Fondazione Bruno Kessler. We then show how we have been able to considerably boost translation performance (as measured in terms of the metrics BLEU and TER) by means of system combination. The joint setups yield empirical gains of up to 1.4 points in BLEU and 2.8 points in TER on the IWSLT test sets compared to the best single systems.
In this paper, we present the KIT systems participating in all three official directions, namely English→German, German→English, and English→French, in translation tasks of the IWSLT 2013 machine translation evaluation. Additionally, we present the results for our submissions to the optional directions English→Chinese and English→Arabic. We used phrase-based translation systems to generate the translations. This year, we focused on adapting the systems towards ASR input. Furthermore, we investigated different reordering models as well as an extended discriminative word lexicon. Finally, we added a data selection approach for domain adaptation.
We analyze the performance of source sentence reordering, a common reordering approach, using oracle experiments on German-English and English-German translation. First, we show that the potential of this approach is very promising. Compared to a monotone translation, the optimally reordered source sentence leads to improvements of up to 4.6 and 6.2 BLEU points, depending on the language. Furthermore, we perform a detailed evaluation of the different aspects of the approach. We analyze the impact of the restriction of the search space by reordering lattices and we can show that using more complex rule types for reordering results in better approximation of the optimally reordered source. However, a gap of about 3 to 3.8 BLEU points remains, presenting a promising perspective for research on extending the search space through better reordering rules. When evaluating the ranking of different reordering variants, the results reveal that the search for the best path in the lattice performs very well for German-English translation. For English-German translation there is potential for an improvement of up to 1.4 BLEU points through a better ranking of the different reordering possibilities in the reordering lattice.
In this paper, we present the KIT systems participating in the English-French TED Translation tasks in the framework of the IWSLT 2012 machine translation evaluation. We also present several additional experiments on the English-German, English-Chinese and English-Arabic translation pairs. Our system is a phrase-based statistical machine translation system, extended with many additional models which were proven to enhance the translation quality. For instance, it uses the part-of-speech (POS)-based reordering, translation and language model adaptation, bilingual language model, word-cluster language model, discriminative word lexica (DWL), and continuous space language model. In addition to this, the system incorporates special steps in the preprocessing and in the post-processing step. In the preprocessing the noisy corpora are filtered by removing the noisy sentence pairs, whereas in the postprocessing the agreement between a noun and its surrounding words in the French translation is corrected based on POS tags with morphological information. Our system deals with speech transcription input by removing case information and punctuation except periods from the text translation model.
We present a novel approach for continuous space language models in statistical machine translation by using Restricted Boltzmann Machines (RBMs). The probability of an n-gram is calculated by the free energy of the RBM instead of a feedforward neural net. Therefore, the calculation is much faster and can be integrated into the translation process instead of using the language model only in a re-ranking step. Furthermore, it is straightforward to introduce additional word factors into the language model. We observed a faster convergence in training if we include automatically generated word classes as an additional word factor. We evaluated the RBM-based language model on the German to English and English to French translation task of TED lectures. Instead of replacing the conventional n-gram-based language model, we trained the RBM-based language model on the more important but smaller in-domain data and combined them in a log-linear way. With this approach we could show improvements of about half a BLEU point on the translation task.
In spoken language translation (SLT), finding proper segmentation and reconstructing punctuation marks are not only significant but also challenging tasks. In this paper we present our recent work on speech translation quality analysis for German-English by improving sentence segmentation and punctuation. From oracle experiments, we show an upper bound of translation quality if we had human-generated segmentation and punctuation on the output stream of speech recognition systems. In our oracle experiments we gain 1.78 BLEU points of improvements on the lecture test set. We build a monolingual translation system from German to German implementing segmentation and punctuation prediction as a machine translation task. Using the monolingual translation system we get an improvement of 1.53 BLEU points on the lecture test set, which is a comparable performance against the upper bound drawn by the oracle experiments.
We report here on the eighth evaluation campaign organized in 2011 by the IWSLT workshop series. That IWSLT 2011 evaluation focused on the automatic translation of public talks and included tracks for speech recognition, speech translation, text translation, and system combination. Unlike in previous years, all data supplied for the evaluation has been publicly released on the workshop website, and is at the disposal of researchers interested in working on our benchmarks and in comparing their results with those published at the workshop. This paper provides an overview of the IWSLT 2011 evaluation campaign, and describes the data supplied, the evaluation infrastructure made available to participants, and the subjective evaluation carried out.
This paper gives a detailed analysis of different approaches to adapt a statistical machine translation system towards a target domain using small amounts of parallel in-domain data. Therefore, we investigate the differences between the approaches addressing adaptation on the two main steps of building a translation model: The candidate selection and the phrase scoring. For the latter step we characterized the differences by four key aspects. We performed experiments on two different tasks of speech translation and analyzed the influence of the different aspects on the overall translation quality. On both tasks we could show significant improvements by using the presented adaptation techniques.
This paper presents the KIT system participating in the English→French TALK Translation tasks in the framework of the IWSLT 2011 machine translation evaluation. Our system is a phrase-based translation system using POS-based reordering extended with many additional features. First of all, a special preprocessing is devoted to the Giga corpus in order to minimize the effect of the great amount of noise it contains. In addition, the system gives more importance to the in-domain data by adapting the translation and the language models as well as by using a wordcluster language model. Furthermore, the system is extended by a bilingual language model and a discriminative word lexicon. The automatic speech transcription input usually has no or wrong punctuation marks, therefore these marks were especially removed from the source training data for the SLT system training.
The Quaero program is an international project promoting research and industrial innovation on technologies for automatic analysis and classification of multimedia and multilingual documents. Within the program framework, research organizations and industrial partners collaborate to develop prototypes of innovating applications and services for access and usage of multimedia data. One of the topics addressed is the translation of spoken language. Each year, a project-internal evaluation is conducted by DGA to monitor the technological advances. This work describes the design and results of the 2011 evaluation campaign. The participating partners were RWTH, KIT, LIMSI and SYSTRAN. Their approaches are compared on both ASR output and reference transcripts of speech data for the translation between French and German. The results show that the developed techniques further the state of the art and improve translation quality.
When building a university lecture translation system, one important step is to adapt it to the target domain. One problem in this adaptation task is to acquire translations for domain specific terms. In this approach we tried to get these translations from Wikipedia, which provides articles on very specific topics in many different languages. To extract translations for the domain specific terms, we used the interlanguage links of Wikipedia . We analyzed different methods to integrate this corpus into our system and explored methods to disambiguate between different translations by using the text of the articles. In addition, we developed methods to handle different morphological forms of the specific terms in morphologically rich input languages like German. The results show that the number of out-of-vocabulary (OOV) words could be reduced by 50% on computer science lectures and the translation quality could be improved by more than 1 BLEU point.
In an increasingly globalized world, situations in which people of different native tongues have to communicate with each other become more and more frequent. In many such situations, human interpreters are prohibitively expensive or simply not available. Automatic spoken language translation (SLT), as a cost-effective solution to this dilemma, has received increased attention in recent years. For a broad number of applications, including live SLT of lectures and oral presentations, these automatic systems should ideally operate in real time and with low latency. Large and highly specialized vocabularies as well as strong variations in speaking style – ranging from read speech to free presentations suffering from spontaneous events – make simultaneous SLT of lectures a challenging task. This paper presents our progress in building a simultaneous German-English lecture translation system. We emphasize some of the challenges which are particular to this language pair and propose solutions to tackle some of the problems encountered.