How Good is Your Tokenizer? On the Monolingual Performance of Multilingual Language Models
01 Aug 2021-pp 3118-3135
TL;DR: The authors provide a systematic and comprehensive empirical comparison of pretrained multilingual language models versus their monolingual counterparts with regard to their monolinguistic task performance, and find that while the pretraining data size is an important factor in the downstream performance, a designated mon-olingual tokenizer plays an equally important role in downstream performance.
Abstract: In this work, we provide a systematic and comprehensive empirical comparison of pretrained multilingual language models versus their monolingual counterparts with regard to their monolingual task performance. We study a set of nine typologically diverse languages with readily available pretrained monolingual models on a set of five diverse monolingual downstream tasks. We first aim to establish, via fair and controlled comparisons, if a gap between the multilingual and the corresponding monolingual representation of that language exists, and subsequently investigate the reason for any performance difference. To disentangle conflating factors, we train new monolingual models on the same data, with monolingually and multilingually trained tokenizers. We find that while the pretraining data size is an important factor, a designated monolingual tokenizer plays an equally important role in the downstream performance. Our results show that languages that are adequately represented in the multilingual model’s vocabulary exhibit negligible performance decreases over their monolingual counterparts. We further find that replacing the original multilingual tokenizer with the specialized monolingual tokenizer improves the downstream performance of the multilingual model for almost every task and language.
Citations
More filters
21 Jun 2021
TL;DR: In this article, the authors present a new dataset called Case Holdings On Legal Decisions (CaseHOLD), which consists of over 53,000+ multiple choice questions to identify the relevant holding of a cited case.
Abstract: While self-supervised learning has made rapid advances in natural language processing, it remains unclear when researchers should engage in resource-intensive domain-specific pretraining (domain pretraining). The law, puzzlingly, has yielded few documented instances of substantial gains to domain pretraining in spite of the fact that legal language is widely seen to be unique. We hypothesize that these existing results stem from the fact that existing legal NLP tasks are too easy and fail to meet conditions for when domain pretraining can help. To address this, we first present CaseHOLD (Case Holdings On Legal Decisions), a new dataset comprised of over 53,000+ multiple choice questions to identify the relevant holding of a cited case. This dataset presents a fundamental task to lawyers and is both legally meaningful and difficult from an NLP perspective (F1 of 0.4 with a BiLSTM baseline). Second, we assess performance gains on CaseHOLD and existing legal NLP datasets. While a Transformer architecture (BERT) pretrained on a general corpus (Google Books and Wikipedia) improves performance, domain pretraining (on a corpus of ≈3.5M decisions across all courts in the U.S. that is larger than BERT's) with a custom legal vocabulary exhibits the most substantial performance gains with CaseHOLD (gain of 7.2% on F1, representing a 12% improvement on BERT) and consistent performance gains across two other legal tasks. Third, we show that domain pretraining may be warranted when the task exhibits sufficient similarity to the pretraining corpus: the level of performance increase in three legal tasks was directly tied to the domain specificity of the task. Our findings inform when researchers should engage in resource-intensive pretraining and show that Transformer-based architectures, too, learn embeddings suggestive of distinct legal language.
83 citations
Proceedings Article•
26 Aug 2021TL;DR: This paper propose a series of data-efficient methods that enable quick and effective adaptation of pretrained multilingual models to such low-resource languages and unseen scripts. But their methods rely on matrix factorization, which is not suitable for low resource languages.
Abstract: Massively multilingual language models such as multilingual BERT offer state-of-the-art cross-lingual transfer performance on a range of NLP tasks. However, due to limited capacity and large differences in pretraining data sizes, there is a profound performance gap between resource-rich and resource-poor target languages. The ultimate challenge is dealing with under-resourced languages not covered at all by the models and written in scripts unseen during pretraining. In this work, we propose a series of novel data-efficient methods that enable quick and effective adaptation of pretrained multilingual models to such low-resource languages and unseen scripts. Relying on matrix factorization, our methods capitalize on the existing latent knowledge about multiple languages already available in the pretrained model’s embedding matrix. Furthermore, we show that learning of the new dedicated embedding matrix in the target language can be improved by leveraging a small number of vocabulary items (i.e., the so-called lexically overlapping tokens) shared between mBERT’s and target language vocabulary. Our adaptation techniques offer substantial performance gains for languages with unseen scripts. We also demonstrate that they can yield improvements for low-resource languages written in scripts covered by the pretrained model.
25 citations
Proceedings Article•
01 Nov 2021
TL;DR: The authors propose MAD-G (Multilingual ADapter Generation) which generates language adapters from language representations based on typological features. But this approach is not viable for the vast majority of languages, due to limitations in their corpus size or compute budgets.
Abstract: Adapter modules have emerged as a general parameter-efficient means to specialize a pretrained encoder to new domains. Massively multilingual transformers (MMTs) have particularly benefited from additional training of language-specific adapters. However, this approach is not viable for the vast majority of languages, due to limitations in their corpus size or compute budgets. In this work, we propose MAD-G (Multilingual ADapter Generation), which contextually generates language adapters from language representations based on typological features. In contrast to prior work, our time- and space-efficient MAD-G approach enables (1) sharing of linguistic knowledge across languages and (2) zero-shot inference by generating language adapters for unseen languages. We thoroughly evaluate MAD-G in zero-shot cross-lingual transfer on part-of-speech tagging, dependency parsing, and named entity recognition. While offering (1) improved fine-tuning efficiency (by a factor of around 50 in our experiments), (2) a smaller parameter budget, and (3) increased language coverage, MAD-G remains competitive with more expensive methods for language-specific adapter training across the board. Moreover, it offers substantial benefits for low-resource languages, particularly on the NER task in low-resource African languages. Finally, we demonstrate that MAD-G’s transfer performance can be further improved via: (i) multi-source training, i.e., by generating and combining adapters of multiple languages with available task-specific training data; and (ii) by further fine-tuning generated MAD-G adapters for languages with monolingual data.
10 citations
Posted Content•
TL;DR: This article showed that efficient embedding based methods that rely solely on the respective datasets outperform computational expensive few-shot fine-tuning approaches, demonstrating that they are able to efficiently identify the best datasets for intermediate training.
Abstract: Intermediate task fine-tuning has been shown to culminate in large transfer gains across many NLP tasks. With an abundance of candidate datasets as well as pre-trained language models, it has become infeasible to run the cross-product of all combinations to find the best transfer setting. In this work we first establish that similar sequential fine-tuning gains can be achieved in adapter settings, and subsequently consolidate previously proposed methods that efficiently identify beneficial tasks for intermediate transfer learning. We experiment with a diverse set of 42 intermediate and 11 target English classification, multiple choice, question answering, and sequence tagging tasks. Our results show that efficient embedding based methods that rely solely on the respective datasets outperform computational expensive few-shot fine-tuning approaches. Our best methods achieve an average Regret@3 of less than 1% across all target tasks, demonstrating that we are able to efficiently identify the best datasets for intermediate training.
8 citations
Posted Content•
TL;DR: The Korean Language Understanding Evaluation (KLUE) benchmark as mentioned in this paper is a collection of 8 Korean NLP tasks, including Topic Classification, SemanticTextual Similarity, Natural Language Inference, Named Entity Recognition, Relation Extraction, Dependency Parsing, Machine Reading Comprehension, and Dialogue State Tracking.
Abstract: We introduce Korean Language Understanding Evaluation (KLUE) benchmark. KLUE is a collection of 8 Korean natural language understanding (NLU) tasks, including Topic Classification, SemanticTextual Similarity, Natural Language Inference, Named Entity Recognition, Relation Extraction, Dependency Parsing, Machine Reading Comprehension, and Dialogue State Tracking. We build all of the tasks from scratch from diverse source corpora while respecting copyrights, to ensure accessibility for anyone without any restrictions. With ethical considerations in mind, we carefully design annotation protocols. Along with the benchmark tasks and data, we provide suitable evaluation metrics and fine-tuning recipes for pretrained language models for each task. We furthermore release the pretrained language models (PLM), KLUE-BERT and KLUE-RoBERTa, to help reproducing baseline models on KLUE and thereby facilitate future research. We make a few interesting observations from the preliminary experiments using the proposed KLUE benchmark suite, already demonstrating the usefulness of this new benchmark suite. First, we find KLUE-RoBERTa-large outperforms other baselines, including multilingual PLMs and existing open-source Korean PLMs. Second, we see minimal degradation in performance even when we replace personally identifiable information from the pretraining corpus, suggesting that privacy and NLU capability are not at odds with each other. Lastly, we find that using BPE tokenization in combination with morpheme-level pre-tokenization is effective in tasks involving morpheme-level tagging, detection and generation. In addition to accelerating Korean NLP research, our comprehensive documentation on creating KLUE will facilitate creating similar resources for other languages in the future. KLUE is available at this https URL.
7 citations
References
More filters
15 Jul 2019
TL;DR: An approach for sentiment analysis of product reviews in Russian using convolutional neural networks using Word2Vec pre-trained vectors as inputs for neural networks, utilizing no hand-crafted features or sentiment lexicons.
Abstract: Nowadays, product reviews on e-commerce sites tend to be a valuable resource in terms of evaluation of customers' behavior, their preferences, and needs. This paper provides an approach for sentiment analysis of product reviews in Russian using convolutional neural networks. We use Word2Vec pre-trained vectors as inputs for neural networks. This approach utilizes no hand-crafted features or sentiment lexicons. The training dataset was collected from reviews on top-ranked goods from the major e-commerce site in Russia, where the user-ranked scores were used as class labels. The system demonstrated the F-measure score up to 75.45% in a three-class classification. The collected training dataset and word embeddings are available to the research community.
41 citations
01 Apr 2021
TL;DR: The authors showed that explicit formalized syntax, injected into transformers through intermediate parsing, has very limited and inconsistent effect on downstream semantic language understanding (LU) performance, and showed that the use of supervised parsing is not necessary for high-level semantic natural language understanding.
Abstract: Traditional NLP has long held (supervised) syntactic parsing necessary for successful higher-level semantic language understanding (LU). The recent advent of end-to-end neural models, self-supervised via language modeling (LM), and their success on a wide range of LU tasks, however, questions this belief. In this work, we empirically investigate the usefulness of supervised parsing for semantic LU in the context of LM-pretrained transformer networks. Relying on the established fine-tuning paradigm, we first couple a pretrained transformer with a biaffine parsing head, aiming to infuse explicit syntactic knowledge from Universal Dependencies treebanks into the transformer. We then fine-tune the model for LU tasks and measure the effect of the intermediate parsing training (IPT) on downstream LU task performance. Results from both monolingual English and zero-shot language transfer experiments (with intermediate target-language parsing) show that explicit formalized syntax, injected into transformers through IPT, has very limited and inconsistent effect on downstream LU performance. Our results, coupled with our analysis of transformers’ representation spaces before and after intermediate parsing, make a significant step towards providing answers to an essential question: how (un)availing is supervised parsing for high-level semantic natural language understanding in the era of large neural models?
36 citations
Posted Content•
TL;DR: CANINE as discussed by the authors is a neural encoder that operates directly on character sequences, without explicit tokenization or vocabulary, and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias.
Abstract: Pipelined NLP systems have largely been superseded by end-to-end neural modeling, yet nearly all commonly-used models still require an explicit tokenization step. While recent tokenization approaches based on data-derived subword lexicons are less brittle than manually engineered tokenizers, these techniques are not equally suited to all languages, and the use of any fixed vocabulary may limit a model's ability to adapt. In this paper, we present CANINE, a neural encoder that operates directly on character sequences, without explicit tokenization or vocabulary, and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias. To use its finer-grained input effectively and efficiently, CANINE combines downsampling, which reduces the input sequence length, with a deep transformer stack, which encodes context. CANINE outperforms a comparable mBERT model by 2.8 F1 on TyDi QA, a challenging multilingual benchmark, despite having 28% fewer model parameters.
36 citations
01 Sep 2019
TL;DR: It is found that the English and German models perform well at generation, whereas the multilingual model is lacking, in particular, for Nordic languages.
Abstract: The multilingual BERT model is trained on 104 languages and meant to serve as a universal language model and tool for encoding sentences. We explore how well the model performs on several languages across several tasks: a diagnostic classification probing the embeddings for a particular syntactic property, a cloze task testing the language modelling ability to fill in gaps in a sentence, and a natural language generation task testing for the ability to produce coherent text fitting a given context. We find that the currently available multilingual BERT model is clearly inferior to the monolingual counterparts, and cannot in many cases serve as a substitute for a well-trained monolingual model. We find that the English and German models perform well at generation, whereas the multilingual model is lacking, in particular, for Nordic languages. The code of the experiments in the paper is available at: https://github.com/TurkuNLP/bert-eval
34 citations
Posted Content•
TL;DR: This paper proposes two tagging methods to solve the problem of data inconsistency, including a heuristic tagging method and a machine auxiliary tagging method, and builds a discourse-level dataset from hundreds of Chinese literature articles for improving this task.
Abstract: Named Entity Recognition and Relation Extraction for Chinese literature text is regarded as the highly difficult problem, partially because of the lack of tagging sets. In this paper, we build a discourse-level dataset from hundreds of Chinese literature articles for improving this task. To build a high quality dataset, we propose two tagging methods to solve the problem of data inconsistency, including a heuristic tagging method and a machine auxiliary tagging method. Based on this corpus, we also introduce several widely used models to conduct experiments. Experimental results not only show the usefulness of the proposed dataset, but also provide baselines for further research. The dataset is available at this https URL
34 citations