Additive interventions are a recently-proposed mechanism for controlling target-side attributes in neural machine translation. In contrast to tag-based approaches which manipulate the raw source sequence, interventions work by directly modulating the encoder representation of all tokens in the sequence. We examine the role of additive interventions in a large-scale multi-domain machine translation setting and compare its performance in various inference scenarios. We find that while the performance difference is small between intervention-based systems and tag-based systems when the domain label matches the test domain, intervention-based systems are robust to label error, making them an attractive choice under label uncertainty. Further, we find that the superiority of single-domain fine-tuning comes under question when training data size is scaled, contradicting previous findings.
Traditional machine translation (MT) metrics provide an average measure of translation quality that is insensitive to the long tail of behavioral problems in MT. Examples include translation of numbers, physical units, dropped content and hallucinations. These errors, which occur rarely and unpredictably in Neural Machine Translation (NMT), greatly undermine the reliability of state-of-the-art MT systems. Consequently, it is important to have visibility into these problems during model development. Towards this direction, we introduce SALTED, a specifications-based framework for behavioral testing of MT models that provides fine-grained views of salient long-tail errors, permitting trustworthy visibility into previously invisible problems. At the core of our approach is the development of high-precision detectors that flag errors (or alternatively, verify output correctness) between a source sentence and a system output. We demonstrate that such detectors could be used not just to identify salient long-tail errors in MT systems, but also for higher-recall filtering of the training data, fixing targeted errors with model fine-tuning in NMT and generating novel data for metamorphic testing to elicit further bugs in models.
Neural transducers have been widely used in automatic speech recognition (ASR). In this paper, we introduce it to streaming end-to-end speech translation (ST), which aims to convert audio signals to texts in other languages directly. Compared with cascaded ST that performs ASR followed by text-based machine translation (MT), the proposed Transformer transducer (TT)-based ST model drastically reduces inference latency, exploits speech information, and avoids error propagation from ASR to MT. To improve the modeling capacity, we propose attention pooling for the joint network in TT. In addition, we extend TT-based ST to multilingual ST, which generates texts of multiple languages at the same time. Experimental results on a large-scale 50 thousand (K) hours pseudo-labeled training set show that TT-based ST not only significantly reduces inference time but also outperforms non-streaming cascaded ST for English-German translation.
This paper presents the JHU-Microsoft joint submission for WMT 2021 quality estimation shared task. We only participate in Task 2 (post-editing effort estimation) of the shared task, focusing on the target-side word-level quality estimation. The techniques we experimented with include Levenshtein Transformer training and data augmentation with a combination of forward, backward, round-trip translation, and pseudo post-editing of the MT output. We demonstrate the competitiveness of our system compared to the widely adopted OpenKiwi-XLM baseline. Our system is also the top-ranking system on the MT MCC metric for the English-German language pair.
We propose a novel scheme to use the Levenshtein Transformer to perform the task of word-level quality estimation. A Levenshtein Transformer is a natural fit for this task: trained to perform decoding in an iterative manner, a Levenshtein Transformer can learn to post-edit without explicit supervision. To further minimize the mismatch between the translation task and the word-level QE task, we propose a two-stage transfer learning procedure on both augmented data and human post-editing data. We also propose heuristics to construct reference labels that are compatible with subword-level finetuning and inference. Results on WMT 2020 QE shared task dataset show that our proposed method has superior data efficiency under the data-constrained setting and competitive performance under the unconstrained setting.
Machine translation models have discrete vocabularies and commonly use subword segmentation techniques to achieve an 'open-vocabulary.' This approach relies on consistent and correct underlying unicode sequences, and makes models susceptible to degradation from common types of noise and variation. Motivated by the robustness of human language processing, we propose the use of visual text representations, which dispense with a finite set of text embeddings in favor of continuous vocabularies created by processing visually rendered text. We show that models using visual text representations approach or match performance of text baselines on clean TED datasets. More importantly, models with visual embeddings demonstrate significant robustness to varied types of noise, achieving e.g., 25.9 BLEU on a character permuted German--English task where subword models degrade to 1.9.
We present the Multilingual TEDx corpus, built to support speech recognition (ASR) and speech translation (ST) research across many non-English source languages. The corpus is a collection of audio recordings from TEDx talks in 8 source languages. We segment transcripts into sentences and align them to the source-language audio and target-language translations. The corpus is released along with open-sourced code enabling extension to new talks and languages as they become available. Our corpus creation methodology can be applied to more languages than previous work, and creates multi-way parallel evaluation sets. We provide baselines in multiple ASR and ST settings, including multilingual models to improve translation performance for low-resource language pairs.
Recent work has shown that a multilingual neural machine translation (NMT) model can be used to judge how well a sentence paraphrases another sentence in the same language; however, attempting to generate paraphrases from the model using beam search produces trivial copies or near copies. We introduce a simple paraphrase generation algorithm which discourages the production of n-grams that are present in the input. Our approach enables paraphrase generation in many languages from a single multilingual NMT model. Furthermore, the trade-off between semantic similarity and lexical/syntactic diversity between the input and output can be controlled at generation time. We conduct human evaluation to compare our method to a paraphraser trained on a large English synthetic paraphrase database and find that our model produces paraphrases that better preserve semantic meaning and grammatically, for the same level of lexical/syntactic diversity. Additional smaller human assessments demonstrate our approach also works in non-English languages.
Following previous work on automatic paraphrasing, we assess the feasibility of improving BLEU (Papineni et al., 2002) using state-of-the-art neural paraphrasing techniques to generate additional references. We explore the extent to which diverse paraphrases can adequately cover the space of valid translations and compare to an alternative approach of generating paraphrases constrained by MT outputs. We compare both approaches to human-produced references in terms of diversity and the improvement in BLEU's correlation with human judgements of MT quality. Our experiments on the WMT19 metrics tasks for all into-English language directions show that somewhat surprisingly, the addition of diverse paraphrases, even those produced by humans, leads to only small, inconsistent changes in BLEU's correlation with human judgments, suggesting that BLEU's ability to correctly exploit multiple references is limited
We propose the use of a sequence-to-sequence paraphraser for automatic machine translation evaluation. The paraphraser takes a human reference as input and then force-decodes and scores an MT system output. We propose training the aforementioned paraphraser as a multilingual NMT system, treating paraphrasing as a zero-shot "language pair" (e.g., Russian to Russian). We denote our paraphraser "unbiased" because the mode of our model's output probability is centered around a copy of the input sequence, which in our case represent the best case scenario where the MT system output matches a human reference. Our method is simple and intuitive, and our single model (trained in 39 languages) outperforms or statistically ties with all prior metrics on the WMT19 segment-level shared metrics task in all languages, excluding Gujarati where the model had no training data. We also explore using our model conditioned on the source instead of the reference, and find that it outperforms every quality estimation as a metric system from the WMT19 shared task on quality estimation by a statistically significant margin in every language pair.