This is an internal, incomplete preview of a proposed change to the ACL Anthology.
For efficiency reasons, we generate only three BibTeX files per volume, and the preview may be incomplete in other ways, or contain mistakes.
Do not treat this content as an official publication.
In this paper, we present a new online dictionary of Akuzipik, an Indigenous language of St. Lawrence Island (Alaska) and Chukotka (Russia).We discuss community desires for strengthening language use in the community and in educational settings, and present specific features of an online dictionary designed to serve these community goals.
In this paper, we challenge the ACL community to reckon with historical and ongoing colonialism by adopting a set of ethical obligations and best practices drawn from the Indigenous studies literature. While the vast majority of NLP research focuses on a very small number of very high resource languages (English, Chinese, etc), some work has begun to engage with Indigenous languages. No research involving Indigenous language data can be considered ethical without first acknowledging that Indigenous languages are not merely very low resource languages. The toxic legacy of colonialism permeates every aspect of interaction between Indigenous communities and outside researchers. To this end, we propose that the ACL draft and adopt an ethical framework for NLP researchers and computational linguists wishing to engage in research involving Indigenous languages.
In this paper, we present a straightforward technique for constructing interpretable word embeddings from morphologically analyzed examples (such as interlinear glosses) for all of the world’s languages. Currently, fewer than 300-400 languages out of approximately 7000 have have more than a trivial amount of digitized texts; of those, between 100-200 languages (most in the Indo-European language family) have enough text data for BERT embeddings of reasonable quality to be trained. The word embeddings in this paper are explicitly designed to be both linguistically interpretable and fully capable of handling the broad variety found in the world’s diverse set of 7000 languages, regardless of corpus size or morphological characteristics. We demonstrate the applicability of our representation through examples drawn from a typologically diverse set of languages whose morphology includes prefixes, suffixes, infixes, circumfixes, templatic morphemes, derivational morphemes, inflectional morphemes, and reduplication.
Prior studies in multilingual language modeling (e.g., Cotterell et al., 2018; Mielke et al., 2019) disagree on whether or not inflectional morphology makes languages harder to model. We attempt to resolve the disagreement and extend those studies. We compile a larger corpus of 145 Bible translations in 92 languages and a larger number of typological features.1 We fill in missing typological data for several languages and consider corpus-based measures of morphological complexity in addition to expert-produced typological features. We find that several morphological measures are significantly associated with higher surprisal when LSTM models are trained with BPE-segmented data. We also investigate linguistically motivated subword segmentation strategies like Morfessor and Finite-State Transducers (FSTs) and find that these segmentation strategies yield better performance and reduce the impact of a language’s morphology on language modeling.
This paper describes the development of the first Universal Dependencies (UD) treebank for St. Lawrence Island Yupik, an endangered language spoken in the Bering Strait region. While the UD guidelines provided a general framework for our annotations, language-specific decisions were made necessary by the rich morphology of the polysynthetic language. Most notably, we annotated a corpus at the morpheme level as well as the word level. The morpheme level annotation was conducted using an existing morphological analyzer and manual disambiguation. By comparing the two resulting annotation schemes, we argue that morpheme-level annotation is essential for polysynthetic languages like St. Lawrence Island Yupik. Word-level annotation results in degenerate trees for some Yupik sentences and often fails to capture syntactic relations that can be manifested at the morpheme level. Dependency parsing experiments provide further support for morpheme-level annotation. Implications for UD annotation of other polysynthetic languages are discussed.
This article describes a simple PCFG induction model with a fixed category domain that predicts a large majority of attested constituent boundaries, and predicts labels consistent with nearly half of attested constituent labels on a standard evaluation data set of child-directed speech. The article then explores the idea that the difference between simple grammars exhibited by child learners and fully recursive grammars exhibited by adult learners may be an effect of increasing working memory capacity, where the shallow grammars are constrained images of the recursive grammars. An implementation of these memory bounds as limits on center embedding in a depth-specific transform of a recursive grammar yields a significant improvement over an equivalent but unbounded baseline, suggesting that this arrangement may indeed confer a learning advantage.
This system paper describes an end-to-end NMT pipeline for the Japanese ↔ English news translation task as submitted to WMT 2021, where we explore the efficacy of techniques such as tokenizing with language-independent and language-dependent tokenizers, normalizing by orthographic conversion, creating a politeness-and-formality-aware model by implementing a tagger, back-translation, model ensembling, and n-best reranking. We use parallel corpora provided by WMT 2021 organizers for training, and development and test data from WMT 2020 for evaluation of different experiment models. The preprocessed corpora are trained with a Transformer neural network model. We found that combining various techniques described herein, such as language-independent BPE tokenization, incorporating politeness and formality tags, model ensembling, n-best reranking, and back-translation produced the best translation models relative to other experiment systems.
St. Lawrence Island Yupik is an endangered polysynthetic language of the Bering Strait region. While conducting linguistic fieldwork between 2016 and 2019, we observed substantial support within the Yupik community for language revitalization and for resource development to support Yupik education. To that end, Chen & Schwartz (2018) implemented a finite-state morphological analyzer as a critical enabling technology for use in Yupik language education and technology. Chen & Schwartz (2018) reported a morphological analysis coverage rate of approximately 75% on a dataset of 60K Yupik tokens, leaving considerable room for improvement. In this work, we present a re-implementation of the Chen & Schwartz (2018) finite-state morphological analyzer for St. Lawrence Island Yupik that incorporates new linguistic insights; in particular, in this implementation we make use of the Paradigm Function Morphology (PFM) theory of morphology. We evaluate this new PFM-based morphological analyzer, and demonstrate that it consistently outperforms the existing analyzer of Chen & Schwartz (2018) with respect to accuracy and coverage rate across multiple datasets.
Unsupervised PCFG inducers hypothesize sets of compact context-free rules as explanations for sentences. PCFG induction not only provides tools for low-resource languages, but also plays an important role in modeling language acquisition (Bannard et al., 2009; Abend et al. 2017). However, current PCFG induction models, using word tokens as input, are unable to incorporate semantics and morphology into induction, and may encounter issues of sparse vocabulary when facing morphologically rich languages. This paper describes a neural PCFG inducer which employs context embeddings (Peters et al., 2018) in a normalizing flow model (Dinh et al., 2015) to extend PCFG induction to use semantic and morphological information. Linguistically motivated sparsity and categorical distance constraints are imposed on the inducer as regularization. Experiments show that the PCFG induction model with normalizing flow produces grammars with state-of-the-art accuracy on a variety of different languages. Ablation further shows a positive effect of normalizing flow, context embeddings and proposed regularizers.
In this paper, we introduce a morphologically-aware electronic dictionary for St. Lawrence Island Yupik, an endangered language of the Bering Strait region. Implemented using HTML, Javascript, and CSS, the dictionary is set in an uncluttered interface and permits users to search in Yupik or in English for Yupik root words and Yupik derivational suffixes. For each matching result, our electronic dictionary presents the user with the corresponding entry from the Badten (2008) Yupik-English paper dictionary. Because Yupik is a polysynthetic language, handling of multimorphemic word forms is critical. If a user searches for an inflected Yupik word form, we perform a morphological analysis and return entries for the root word and for any derivational suffixes present in the word. This electronic dictionary should serve not only as a valuable resource for all students and speakers of Yupik, but also for field linguists working towards documentation and conservation of the language.
There has been recent interest in applying cognitively- or empirically-motivated bounds on recursion depth to limit the search space of grammar induction models (Ponvert et al., 2011; Noji and Johnson, 2016; Shain et al., 2016). This work extends this depth-bounding approach to probabilistic context-free grammar induction (DB-PCFG), which has a smaller parameter space than hierarchical sequence models, and therefore more fully exploits the space reductions of depth-bounding. Results for this model on grammar acquisition from transcribed child-directed speech and newswire text exceed or are competitive with those of other models when evaluated on parse accuracy. Moreover, grammars acquired from this model demonstrate a consistent use of category labels, something which has not been demonstrated by other acquisition models.
There have been several recent attempts to improve the accuracy of grammar induction systems by bounding the recursive complexity of the induction model. Modern depth-bounded grammar inducers have been shown to be more accurate than early unbounded PCFG inducers, but this technique has never been compared against unbounded induction within the same system, in part because most previous depth-bounding models are built around sequence models, the complexity of which grows exponentially with the maximum allowed depth. The present work instead applies depth bounds within a chart-based Bayesian PCFG inducer, where bounding can be switched on and off, and then samples trees with or without bounding. Results show that depth-bounding is indeed significantly effective in limiting the search space of the inducer and thereby increasing accuracy of resulting parsing model, independent of the contribution of modern Bayesian induction techniques. Moreover, parsing results on English, Chinese and German show that this bounded model is able to produce parse trees more accurately than or competitively with state-of-the-art constituency grammar induction models.
The utilization of statistical machine translation (SMT) has grown enormously over the last decade, many using open-source software developed by the NLP community. As commercial use has increased, there is need for software that is optimized for commercial requirements, in particular, fast phrase-based decoding and more efficient utilization of modern multicore servers. In this paper we re-examine the major components of phrase-based decoding and decoder implementation with particular emphasis on speed and scalability on multicore machines. The result is a drop-in replacement for the Moses decoder which is up to fifteen times faster and scales monotonically with the number of cores.
This paper presents a new memory-bounded left-corner parsing model for unsupervised raw-text syntax induction, using unsupervised hierarchical hidden Markov models (UHHMM). We deploy this algorithm to shed light on the extent to which human language learners can discover hierarchical syntax through distributional statistics alone, by modeling two widely-accepted features of human language acquisition and sentence processing that have not been simultaneously modeled by any existing grammar induction algorithm: (1) a left-corner parsing strategy and (2) limited working memory capacity. To model realistic input to human language learners, we evaluate our system on a corpus of child-directed speech rather than typical newswire corpora. Results beat or closely match those of three competing systems.
Various small-scale pilot studies have found that for at least some documents, monolingual target language speakers may be able to successfully post-edit machine translations. We begin by analyzing previously published post-editing data to ascertain the effect, if any, of original source language on post-editing quality. Schwartz et al. (2014) hypothesized that post-editing success may be more pronounced when the monolingual post-editors are experts in the domain of the translated documents. This work tests that hypothesis by asking a domain expert to post-edit machine translations of a French scientific article (Besacier, 2014) into English. We find that the monolingual domain expert post-editor was able to successfully post-edit 86.7% of the sentences without requesting assistance from a bilingual post-editor. We evaluate the post-edited sentences according to a bilingual adequacy metric, and find that 96.5% of those sentences post-edited by only a monolingual post-editor are judged to be completely correct. These results confirm that a monolingual domain expert can successfully triage the post-editing effort, substantially reducing the workload on the bilingual post-editor by only sending the most challenging sentences to the bilingual post-editor.
We present a simple user interface for post-editing that presents the user with the source sentence, machine translation, and word alignments for each sentence in a test document (Figure 1). This software is open source, written in Java, and has no external dependencies; it can be run on Linux, Mac OS X, and Windows. This software was originally designed for monolingual post-editors, but should be equally usable by bilingual post-editors. While it may seem counter-intuitive to present monolingual post-editors with the source sentence, we found that the presence of alignment links between source words and target words can in fact aid a monolingual post-editor, especially with regard to correcting word order. For example, in our experiments using this interface (Schwartz et al., 2014), post-editors encountered some sentences where a word or phrase was enclosed within bracketing punctuation marks (such as quotation marks, commas, or parentheses) in the source sentence, and the machine translation system incorrectly reordered the word or phrase outside the enclosing punctuation; by examining the alignment links the post-editors were able to correct such reordering mistakes.
This paper describes the MIT-LL/AFRL statistical MT system and the improvements that were developed during the IWSLT 2013 evaluation campaign [1]. As part of these efforts, we experimented with a number of extensions to the standard phrase-based model that improve performance on the Russian to English, Chinese to English, Arabic to English, and English to French TED-talk translation task. We also applied our existing ASR system to the TED-talk lecture ASR task. We discuss the architecture of the MIT-LL/AFRL MT system, improvements over our 2012 system, and experiments we ran during the IWSLT-2013 evaluation. Specifically, we focus on 1) cross-entropy filtering of MT training data, and 2) improved optimization techniques, 3) language modeling, and 4) approximation of out-of-vocabulary words.
Multi-parallel corpora provide a potentially rich resource for machine translation. This paper surveys existing methods for utilizing such resources, including hypothesis ranking and system combination techniques. We find that despite significant research into system combination, relatively little is know about how best to translate when multiple parallel source languages are available. We provide results to show that the MAX multilingual multi-source hypothesis ranking method presented by Och and Ney (2001) does not reliably improve translation quality when a broad range of language pairs are considered. We also show that the PROD multilingual multi-source hypothesis ranking method of Och and Ney (2001) cannot be used with standard phrase-based translation engines, due to a high number of unreachable hypotheses. Finally, we present an oracle experiment which shows that current hypothesis ranking methods fall far short of the best results reachable via sentence-level ranking.