Knowledge distillation describes a method for training a student network to perform better by learning from a stronger teacher network. Translating a sentence with an Neural Machine Translation (NMT) engine is time expensive and having a smaller model speeds up this process. We demonstrate how to transfer the translation quality of an ensemble and an oracle BLEU teacher network into a single NMT system. Further, we present translation improvements from a teacher network that has the same architecture and dimensions of the student network. As the training of the student model is still expensive, we introduce a data filtering method based on the knowledge of the teacher model that not only speeds up the training, but also leads to better translation quality. Our techniques need no code change and can be easily reproduced with any NMT architecture to speed up the decoding process.
University of Science and Technology of China（中国科学技术大学）；Microsoft Research（微软研究院）
Neural machine translation on low-resource language is challenging due to the lack of bilingual sentence pairs. Previous works usually solve the low-resource translation problem with knowledge transfer in a multilingual setting. In this paper, we propose the concept of Language Graph and further design a novel graph distillation algorithm that boosts the accuracy of low-resource translations in the graph with forward and backward knowledge distillation. Preliminary experiments on the TED talks multilingual dataset demonstrate the effectiveness of our proposed method. Specifically, we improve t…
Institute of Automation, CAS（中国科学院自动化研究所）；University of Chinese Academy of Sciences（中国科学院大学）；Fanyu AI Research, Beijing Fanyu Technology Ltd.（北京凡语科技有限公司凡语人工智能研究院）
A common scenario of Multilingual Neural Machine Translation (MNMT) is that each translation task arrives in a sequential manner, and the training data of previous tasks is unavailable. In this scenario, the current methods suffer heavily from catastrophic forgetting (CF). To alleviate the CF, we investigate knowledge distillation based life-long learning methods. Specifically, in one-tomany scenario, we propose a multilingual distillation method to make the new model (student) jointly learn multilingual output from old model (teacher) and new task. In many-to one scenario, we find that direct…
Microsoft Research Asia（微软亚洲研究院）；Zhejiang University（浙江大学）；Peking University（北京大学）
Multilingual machine translation, which translates multiple languages with a single model, has attracted much attention due to its efficiency of offline training and online serving. However, traditional multilingual translation usually yields inferior accuracy compared with the counterpart using individual models for each language pair, due to language diversity and model capacity limitations. In this paper, we propose a distillation-based approach to boost the accuracy of multilingual machine translation. Specifically, individual models are first trained and regarded as teachers, and then the…
Neural Machine Translation(NMT) models are usually trained via unidirectional decoder which corresponds to optimizing one-step-ahead prediction. However, this kind of unidirectional decoding framework may incline to focus on local structure rather than global coherence. To alleviate this problem, we propose a novel method, Self-Knowledge Distillation with Bidirectional Decoder for Neural Machine Translation(SBD-NMT). We deploy a backward decoder which can act as an effective regularization method to the forward decoder. By leveraging the backward decoder's information about the longer-term fut…
Multilingual Non-Autoregressive Machine Translation without Knowledge Distillation
University of Alberta（阿尔伯塔大学）；Alberta Machine Intelligence Institute (Amii)（阿尔伯塔机器智能研究所（Amii））；Institute for AI Industry Research (AIR), Tsinghua University（清华大学人工智能产业研究院（AIR））；ByteDance Research（字节跳动研究院）；Damo Academy, Alibaba（阿里巴巴达摩院）
Carnegie Mellon University（卡内基梅隆大学）；Facebook AI Research（脸书人工智能研究）
Non-autoregressive machine translation (NAT) systems predict a sequence of output tokens in parallel, achieving substantial improvements in generation speed compared to autoregressive models. Existing NAT models usually rely on the technique of knowledge distillation, which creates the training data from a pretrained autoregressive model for better performance. Knowledge distillation is empirically useful, leading to large gains in accuracy for NAT models, but the reason for this success has, as of yet, been unclear. In this paper, we first design systematic experiments to investigate why know…
This paper explores augmenting monolingual data for knowledge distillation in neural machine translation. Source language monolingual text can be incorporated as a forward translation. Interestingly, we find the best way to incorporate target language monolingual text is to translate it to the source language and round-trip translate it back to the target language, resulting in a fully synthetic corpus. We find that combining monolingual data from both source and target languages yields better performance than a corpus twice as large only in one language. Moreover, experiments reveal that the …
Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）；University of Chinese Academy of Sciences（中国科学院大学）
Non-autoregressive neural machine translation (NAT) suffers from the multi-modality problem: the source sentence may have multiple correct translations, but the loss function is calculated only according to the reference sentence. Sequence-level knowledge distillation makes the target more deterministic by replacing the target with the output from an autoregressive model. However, the multi-modality problem in the distilled dataset is still nonnegligible. Furthermore, learning from a specific teacher limits the upper bound of the model capability, restricting the potential of NAT models. In th…
Continual Knowledge Distillation for Neural Machine Translation
Tsinghua University（清华大学）；Beijing National Research Center for Information Science and Technology（北京国家信息科学与技术研究中心）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；Institute for AI Industry Research (AIR), Tsinghua University（清华大学人工智能产业研究院（AIR））
Selective Knowledge Distillation for Non-Autoregressive Neural Machine Translation
National Key Laboratory for Novel Software Technology, Nanjing University（南京大学现代软件技术国家重点实验室）；ByteDance AI Lab（字节跳动人工智能实验室）；Collaborative Innovation Center of Novel Software Technology and Industrialization（现代软件技术与产业化协同创新中心）
Sequence-level knowledge distillation (SLKD) is a model compression technique that leverages large, accurate teacher models to train smaller, under-parameterized student models. Why does pre-processing MT data with SLKD help us train smaller models? We test the common hypothesis that SLKD addresses a capacity deficiency in students by "simplifying" noisy data points and find it unlikely in our case. Models trained on concatenations of original and "simplified" datasets generalize just as well as baseline SLKD. We then propose an alternative hypothesis under the lens of data augmentation and re…
Multi-Teacher Knowledge Distillation For Text Image Machine Translation
University of Chinese Academy of Sciences（中国科学院大学）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；Samsung Research China - Beijing（三星中国研究院（北京））；Zhongke Fanyu Technology Co., Ltd（中科凡语科技有限公司）
Wangxuan Institute of Computer Technology, Peking University（北京大学王选计算机研究所）；Center for Data Science, Peking University（北京大学数据科学中心）；The MOE Key Laboratory of Computational Linguistics, Peking University（北京大学计算语言学教育部重点实验室）
k-nearest-neighbor machine translation (NN-MT), proposed by Khandelwal et al. (2021), has achieved many state-of-the-art results in machine translation tasks. Although effective, NN-MT requires conducting NN searches through the large datastore for each decoding step during inference, prohibitively increasing the decoding cost and thus leading to the difficulty for the deployment in real-world applications. In this paper, we propose to move the time-consuming NN search forward to the preprocessing phase, and then introduce Nearest Neighbor Knowledge Distillation (NN-KD) that trains the base NM…
University of Maryland（马里兰大学）；Microsoft Research Asia（微软亚洲研究院）
While non-autoregressive (NAR) models are showing great promise for machine translation, their use is limited by their dependence on knowledge distillation from autoregressive models. To address this issue, we seek to understand why distillation is so effective. Prior work suggests that distilled training data is less complex than manual translations. Based on experiments with the Levenshtein Transformer and the Mask-Predict NAR models on the WMT14 German-English task, this paper shows that different types of complexity have different impacts: while reducing lexical diversity and decreasing re…
We explore best practices for training small, memory efficient machine translation models with sequence-level knowledge distillation in the domain adaptation setting. While both domain adaptation and knowledge distillation are widely-used, their interaction remains little understood. Our large-scale empirical results in machine translation (on three language pairs with three domains each) suggest distilling twice for best performance: once using general-domain data and again using in-domain data with an adapted teacher.
Institute of Formal and Applied Linguistics, Charles University（查尔斯大学形式与应用语言学研究所）
In most of neural machine translation distillation or stealing scenarios, the goal is to preserve the performance of the target model (teacher). The highest-scoring hypothesis of the teacher model is commonly used to train a new model (student). If reference translations are also available, then better hypotheses (with respect to the references) can be upsampled and poor hypotheses either removed or undersampled. This paper explores the importance sampling method landscape (pruning, hypothesis upsampling and undersampling, deduplication and their combination) with English to Czech and English …
Neural Machine Translation (NMT) models achieve state-of-the-art performance on many translation benchmarks. As an active research field in NMT, knowledge distillation is widely applied to enhance the model's performance by transferring teacher model's knowledge on each training sample. However, previous work rarely discusses the different impacts and connections among these samples, which serve as the medium for transferring teacher knowledge. In this paper, we design a novel protocol that can effectively analyze the different impacts of samples by comparing various samples' partitions. Based…
This work introduces a simple regressive ensemble for evaluating machine translation quality based on a set of novel and established metrics. We evaluate the ensemble using a correlation to expert-based MQM scores of the WMT 2021 Metrics workshop. In both monolingual and zero-shot cross-lingual settings, we show a significant performance improvement over single metrics. In the cross-lingual settings, we also demonstrate that an ensemble approach is well-applicable to unseen languages. Furthermore, we identify a strong reference-free baseline that consistently outperforms the commonly-used BLEU…
NTT Communication Science Laboratories（NTT通信科学实验室）；Nara Institute of Science and Technology（奈良先端科学技术大学院大学）
Non-autoregressive neural machine translation (NAT) usually employs sequence-level knowledge distillation using autoregressive neural machine translation (AT) as its teacher model. However, a NAT model often outputs shorter sentences than an AT model. In this work, we propose sequence-level knowledge distillation (SKD) using perturbed length-aware positional encoding and apply it to a student model, the Levenshtein Transformer. Our method outperformed a standard Levenshtein Transformer by 2.5 points in bilingual evaluation understudy (BLEU) at maximum in a WMT14 German to English translation. …
Nanyang Technological University（南洋理工大学）；Salesforce Research Asia（赛富时亚洲研究院）
Recent unsupervised machine translation (UMT) systems usually employ three main principles: initialization, language modeling and iterative back-translation, though they may apply them differently. Crucially, iterative back-translation and denoising auto-encoding for language modeling provide data diversity to train the UMT systems. However, the gains from these diversification processes has seemed to plateau. We introduce a novel component to the standard UMT framework called Cross-model Back-translated Distillation (CBD), that is aimed to induce another level of data diversification that exi…
Oregon State University（俄勒冈州立大学）；Baidu Research（百度研究院）
This paper describes multimodal machine translation systems developed jointly by Oregon State University and Baidu Research for WMT 2018 Shared Task on multimodal translation. In this paper, we introduce a simple approach to incorporate image information by feeding image features to the decoder side. We also explore different sequence level training methods including scheduled sampling and reinforcement learning which lead to substantial improvements. Our systems ensemble several models using different architectures and training methods and achieve the best performance for three subtasks: En-D…
Ensembling is a well-known technique in neural machine translation (NMT) to improve system performance. Instead of a single neural net, multiple neural nets with the same topology are trained separately, and the decoder generates predictions by averaging over the individual models. Ensembling often improves the quality of the generated translations drastically. However, it is not suitable for production systems because it is cumbersome and slow. This work aims to reduce the runtime to be on par with a single system without compromising the translation quality. First, we show that the ensemble …
University of Illinois at Chicago（伊利诺伊大学芝加哥分校）；IQVIA
Quality Estimation (QE) is an important component of the machine translation workflow as it assesses the quality of the translated output without consulting reference translations. In this paper, we discuss our submission to the WMT 2021 QE Shared Task. We participate in Task 2 sentence-level sub-task that challenge participants to predict the HTER score for sentence-level post-editing effort. Our proposed system is an ensemble of multilingual BERT (mBERT)-based regression models, which are generated by fine-tuning on different input settings. It demonstrates comparable performance with respec…
Subword regularizations use multiple subword segmentations during training to improve the robustness of neural machine translation models. In previous subword regularizations, we use multiple segmentations in the training process but use only one segmentation in the inference. In this study, we propose an inference strategy to address this discrepancy. The proposed strategy approximates the marginalized likelihood by using multiple segmentations including the most plausible segmentation and several sampled segmentations. Because the proposed strategy aggregates predictions from several segment…
Quality Estimation (QE) of Machine Translation (MT) is a task to estimate the quality scores for given translation outputs from an unknown MT system. However, QE scores for low-resource languages are usually intractable and hard to collect. In this paper, we focus on the Sentence-Level QE Shared Task of the Fifth Conference on Machine Translation (WMT20), but in a more challenging setting. We aim to predict QE scores of given translation outputs when barely none of QE scores of that paired languages are given during training. We propose an ensemble-based predictor-estimator QE model with trans…
CycleDistill: Bootstrapping Machine Translation using LLMs with Cyclical Distillation
Nilekani Centre at AI4Bharat（AI4Bharat的Nilekani中心）；Indian Institute of Technology（印度理工学院）；Indian Institute of Technology Bombay（印度理工学院孟买分校）；Indian Institute of Engineering Science and Technology Shibpur（Shibpur的印度工程科学与技术学院）
IBM Research AI（IBM研究院人工智能部门）；IBM Research - Ireland（IBM研究院爱尔兰分院）
AMR parsing has experienced an unprecendented increase in performance in the last three years, due to a mixture of effects including architecture improvements and transfer learning. Self-learning techniques have also played a role in pushing performance forward. However, for most recent high performant parsers, the effect of self-learning and silver data augmentation seems to be fading. In this paper we propose to overcome this diminishing returns of silver data by combining Smatch-based ensembling techniques with ensemble distillation. In an extensive experimental setup, we push single model …
Huawei Translation Services Center（华为翻译服务中心）；The University of Melbourne（墨尔本大学）
Recently, non-autoregressive (NAT) models predict outputs in parallel, achieving substantial improvements in generation speed compared to autoregressive (AT) models. While performing worse on raw data, most NAT models are trained as student models on distilled data generated by AT teacher models, which is known as sequence-level Knowledge Distillation. An effective training strategy to improve the performance of AT models is Self-Distillation Mixup (SDM) Training, which pre-trains a model on raw data, generates distilled data by the pre-trained model itself and finally re-trains a model on the…
Recent state-of-the-art approaches to summarization utilize large pre-trained Transformer models. Distilling these models to smaller student models has become critically important for practical use; however there are many different distillation methods proposed by the NLP literature. Recent work on distilling BERT for classification and regression tasks shows strong performance using direct knowledge distillation. Alternatively, machine translation practitioners distill using pseudo-labeling, where a small model is trained on the translations of a larger model. A third, simpler approach is to …
Multilingual Neural Machine Translation (MNMT) trains a single NMT model that supports translation between multiple languages, rather than training separate models for different languages. Learning a single model can enhance the low-resource translation by leveraging data from multiple languages. However, the performance of an MNMT model is highly dependent on the type of languages used in training, as transferring knowledge from a diverse set of languages degrades the translation performance due to negative transfer. In this paper, we propose a Hierarchical Knowledge Distillation (HKD) approa…
Fondazione Bruno Kessler（布鲁诺·凯斯勒基金会）；University of Trento（特伦托大学）；AppTek
Direct speech translation (ST) has shown to be a complex task requiring knowledge transfer from its sub-tasks: automatic speech recognition (ASR) and machine translation (MT). For MT, one of the most promising techniques to transfer knowledge is knowledge distillation. In this paper, we compare the different solutions to distill knowledge in a sequence-to-sequence task like ST. Moreover, we analyze eventual drawbacks of this approach and how to alleviate them maintaining the benefits in terms of translation quality.
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；University of Chinese Academy of Sciences（中国科学院大学）；CAS Center for Excellence in Brain Science and Intelligence Technology（中国科学院脑科学与智能技术卓越创新中心）；Baidu Inc.（百度公司）
End-to-end speech translation (ST), which directly translates from source language speech into target language text, has attracted intensive attentions in recent years. Compared to conventional pipeline systems, end-to-end ST models have advantages of lower latency, smaller model size and less error propagation. However, the combination of speech recognition and text translation in one model is more difficult than each of these two tasks. In this paper, we propose a knowledge distillation approach to improve ST model by transferring the knowledge from text translation model. Specifically, we f…
Harbin Institute of Technology（哈尔滨工业大学）；National Institute of Information and Communications Technology (NICT)（信息通信研究机构（NICT））
Unsupervised neural machine translation (UNMT) has recently achieved remarkable results for several language pairs. However, it can only translate between a single language pair and cannot produce translation results for multiple language pairs at the same time. That is, research on multilingual UNMT has been limited. In this paper, we empirically introduce a simple method to translate between thirteen languages using a single encoder and a single decoder, making use of multilingual data to improve UNMT for all language pairs. On the basis of the empirical findings, we propose two knowledge di…
Lack of specialized data makes building a multi-domain neural machine translation tool challenging. Although emerging literature dealing with low resource languages starts to show promising results, most state-of-the-art models used millions of sentences. Today, the majority of multi-domain adaptation techniques are based on complex and sophisticated architectures that are not adapted for real-world applications. So far, no scalable method is performing better than the simple yet effective mixed-finetuning, i.e finetuning a generic model with a mix of all specialized data and generic data. In …
Scarcity of parallel sentence-pairs poses a significant hurdle for training high-quality Neural Machine Translation (NMT) models in bilingually low-resource scenarios. A standard approach is transfer learning, which involves taking a model trained on a high-resource language-pair and fine-tuning it on the data of the low-resource MT condition of interest. However, it is not clear generally which high-resource language-pair offers the best transfer learning for the target MT setting. Furthermore, different transferred models may have complementary semantic and/or syntactic strengths, hence usin…
Boosting LLM Translation Skills without General Ability Loss via Rationale Distillation
School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；Department of Computer Science, University of Illinois at Chicago（芝加哥大学计算机科学系）
Carnegie Mellon University（卡内基梅隆大学）；Pompeu Fabra University（庞培法布拉大学）；University of Washington（华盛顿大学）；Google DeepMind（谷歌DeepMind）
We introduce two first-order graph-based dependency parsers achieving a new state of the art. The first is a consensus parser built from an ensemble of independently trained greedy LSTM transition-based parsers with different random initializations. We cast this approach as minimum Bayes risk decoding (under the Hamming cost) and argue that weaker consensus within the ensemble is a useful signal of difficulty or ambiguity. The second parser is a "distillation" of the ensemble into a single model. We train the distillation parser using a structured hinge loss objective with a novel cost that in…
This paper contributes a new State Of The Art (SOTA) for Semantic Textual Similarity (STS). We compare and combine a number of recently proposed sentence embedding methods for STS, and propose a novel and simple ensemble knowledge distillation scheme that improves on previous approaches. Our experiments demonstrate that a model trained to learn the average embedding space from multiple ensemble students outperforms all the other individual models with high robustness. Utilizing our distillation method in combination with previous methods, we significantly improve on the SOTA unsupervised STS, …
National University of Defense Technology（中国人民解放军国防科技大学）；Microsoft Research Asia（微软亚洲研究院）
Despite that current reading comprehension systems have achieved significant advancements, their promising performances are often obtained at the cost of making an ensemble of numerous models. Besides, existing approaches are also vulnerable to adversarial attacks. This paper tackles these problems by leveraging knowledge distillation, which aims to transfer knowledge from an ensemble model to a single model. We first demonstrate that vanilla knowledge distillation applied to answer span prediction is effective for reading comprehension systems. We then propose two novel approaches that not on…
Baby Llama: knowledge distillation from an ensemble of teachers trained on a small dataset with no performance penalty
Niels Bohr Institute（尼尔斯·玻尔研究所）；University of Copenhagen（哥本哈根大学）；Departamento de Física Teórica（理论物理系）；Instituto de Física Teórica UAM/CSIC（马德里自治大学/西班牙国家研究委员会理论物理研究所）；Universidad Autónoma de Madrid（马德里自治大学）
Tsinghua University（清华大学）；Microsoft Research Asia（微软亚洲研究院）
Ensemble knowledge distillation can extract knowledge from multiple teacher models and encode it into a single student model. Many existing methods learn and distill the student model on labeled data only. However, the teacher models are usually learned on the same labeled data, and their predictions have high correlations with groudtruth labels. Thus, they cannot provide sufficient knowledge complementary to task labels for student teaching. Distilling on unseen unlabeled data has the potential to enhance the knowledge transfer from the teachers to the student. In this paper, we propose a uni…
