Byte Pair Encoding is Suboptimal for Language Model Pretraining

• URL: http://arxiv.org/abs/2004.03720v2
• Date: Mon, 5 Oct 2020 17:35:44 GMT
• Title: Byte Pair Encoding is Suboptimal for Language Model Pretraining
• Authors: Kaj Bostrom and Greg Durrett
• Abstract summary: We analyze differences between unigram LM tokenization and byte-pair encoding (BPE) We find that the unigram LM tokenization method matches or outperforms BPE across downstream tasks and two languages. We hope that developers of future pretrained LMs will consider adopting the unigram LM method over the more prevalent BPE.
• Score: 49.30780227162387
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The success of pretrained transformer language models (LMs) in natural language processing has led to a wide range of pretraining setups. In particular, these models employ a variety of subword tokenization methods, most notably byte-pair encoding (BPE) (Sennrich et al., 2016; Gage, 1994), the WordPiece method (Schuster and Nakajima, 2012), and unigram language modeling (Kudo, 2018), to segment text. However, to the best of our knowledge, the literature does not contain a direct evaluation of the impact of tokenization on language model pretraining. We analyze differences between BPE and unigram LM tokenization, finding that the latter method recovers subword units that align more closely with morphology and avoids problems stemming from BPE's greedy construction procedure. We then compare the fine-tuned task performance of identical transformer masked language models pretrained with these tokenizations. Across downstream tasks and two languages (English and Japanese), we find that the unigram LM tokenization method matches or outperforms BPE. We hope that developers of future pretrained LMs will consider adopting the unigram LM method over the more prevalent BPE.

Related papers

• Assessing Phrase Break of ESL Speech with Pre-trained Language Models and Large Language Models [7.782346535009883]
  This work introduces approaches to assessing phrase breaks in ESL learners' speech using pre-trained language models (PLMs) and large language models (LLMs)
  arXiv  Detail & Related papers  (2023-06-08T07:10:39Z)
• PEACH: Pre-Training Sequence-to-Sequence Multilingual Models for Translation with Semi-Supervised Pseudo-Parallel Document Generation [5.004814662623874]
  This paper introduces a novel semi-supervised method, SPDG, that generates high-quality pseudo-parallel data for multilingual pre-training. Our experiments show that PEACH outperforms existing approaches used in training mT5 and mBART on various translation tasks.
  arXiv  Detail & Related papers  (2023-04-03T18:19:26Z)
• Multilingual Sentence Transformer as A Multilingual Word Aligner [15.689680887384847]
  We investigate whether multilingual sentence Transformer LaBSE is a strong multilingual word aligner. Experiment results on seven language pairs show that our best aligner outperforms previous state-of-the-art models of all varieties. Our aligner supports different language pairs in a single model, and even achieves new state-of-the-art on zero-shot language pairs that does not appear in the finetuning process.
  arXiv  Detail & Related papers  (2023-01-28T09:28:55Z)
• Language Model Pre-Training with Sparse Latent Typing [66.75786739499604]
  We propose a new pre-training objective, Sparse Latent Typing, which enables the model to sparsely extract sentence-level keywords with diverse latent types. Experimental results show that our model is able to learn interpretable latent type categories in a self-supervised manner without using any external knowledge.
  arXiv  Detail & Related papers  (2022-10-23T00:37:08Z)
• PERT: Pre-training BERT with Permuted Language Model [24.92527883997854]
  PERT is an auto-encoding model (like BERT) trained with Permuted Language Model (PerLM) We permute a proportion of the input text, and the training objective is to predict the position of the original token. We carried out extensive experiments on both Chinese and English NLU benchmarks.
  arXiv  Detail & Related papers  (2022-03-14T07:58:34Z)
• Training Multilingual Pre-trained Language Model with Byte-level Subwords [41.52056437015399]
  We present our practices on training multilingual pre-trained language models with BBPE: Byte-Level BPE (i.e., Byte Pair. In the experiment, we adopted the architecture of NEZHA as the underlying pre-trained language model and the results show that NEZHA trained with byte-level subwords consistently. We release the source code of our byte-level vocabulary building tools and the multilingual pre-trained language models.
  arXiv  Detail & Related papers  (2021-01-23T10:01:28Z)
• Unsupervised Paraphrasing with Pretrained Language Models [85.03373221588707]
  We propose a training pipeline that enables pre-trained language models to generate high-quality paraphrases in an unsupervised setting. Our recipe consists of task-adaptation, self-supervision, and a novel decoding algorithm named Dynamic Blocking. We show with automatic and human evaluations that our approach achieves state-of-the-art performance on both the Quora Question Pair and the ParaNMT datasets.
  arXiv  Detail & Related papers  (2020-10-24T11:55:28Z)
• Pre-training Multilingual Neural Machine Translation by Leveraging Alignment Information [72.2412707779571]
  mRASP is an approach to pre-train a universal multilingual neural machine translation model. We carry out experiments on 42 translation directions across a diverse setting, including low, medium, rich resource, and as well as transferring to exotic language pairs.
  arXiv  Detail & Related papers  (2020-10-07T03:57:54Z)
• Reusing a Pretrained Language Model on Languages with Limited Corpora for Unsupervised NMT [129.99918589405675]
  We present an effective approach that reuses an LM that is pretrained only on the high-resource language. The monolingual LM is fine-tuned on both languages and is then used to initialize a UNMT model. Our approach, RE-LM, outperforms a competitive cross-lingual pretraining model (XLM) in English-Macedonian (En-Mk) and English-Albanian (En-Sq)
  arXiv  Detail & Related papers  (2020-09-16T11:37:10Z)
• Multilingual Denoising Pre-training for Neural Machine Translation [132.66750663226287]
  mBART is a sequence-to-sequence denoising auto-encoder pre-trained on large-scale monolingual corpora. mBART is one of the first methods for pre-training a complete sequence-to-sequence model.
  arXiv  Detail & Related papers  (2020-01-22T18:59:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.