mT5: A Massively Multilingual Pre-trained Text-to-Text Transformer
01 Jun 2021-pp 483-498
TL;DR: This paper proposed a multilingual variant of T5, mT5, which was pre-trained on a new Common Crawl-based dataset covering 101 languages and achieved state-of-the-art performance on many multilingual benchmarks.
Abstract: The recent “Text-to-Text Transfer Transformer” (T5) leveraged a unified text-to-text format and scale to attain state-of-the-art results on a wide variety of English-language NLP tasks. In this paper, we introduce mT5, a multilingual variant of T5 that was pre-trained on a new Common Crawl-based dataset covering 101 languages. We detail the design and modified training of mT5 and demonstrate its state-of-the-art performance on many multilingual benchmarks. We also describe a simple technique to prevent “accidental translation” in the zero-shot setting, where a generative model chooses to (partially) translate its prediction into the wrong language. All of the code and model checkpoints used in this work are publicly available.
Citations
More filters
01 Jan 2022
13 citations
TL;DR: The authors showed that fine-tuning a language model on a single task can outperform a multi-task-prompted language model trained on 300+ different tasks by a mean accuracy of 3.20% and 1.29%.
Abstract: Recently, Language Models (LMs) instruction-tuned on multiple tasks, also known as multitask-prompted fine-tuning (MT), have shown the capability to generalize to unseen tasks. Previous work has shown that scaling the number of training tasks is the key component in making stronger MT LMs. In this work, we report an unexpected finding that an expert LM fine-tuned on just a single task can outperform an MT LM trained with 300+ different tasks on 11 different unseen datasets and on 13 datasets of the BIG-bench benchmark by a mean accuracy of 3.20% and 1.29%, respectively. This finding casts doubt on the previously held belief that simply scaling the number of tasks makes stronger MT LMs. Leveraging this finding, we further show that this distributed approach of training a separate expert LM per training task instead of a single MT LM for zero-shot inference possesses many benefits including (1) avoiding negative task transfer that often occurs during instruction tuning, (2) being able to continually learn new tasks without having to re-train on previous tasks to avoid catastrophic forgetting, and (3) showing compositional capabilities when merging individual experts together. The code is available at https://github.com/joeljang/ELM.
12 citations
TL;DR: This article showed that exposure to pretraining data may break the distributional control of pre-trained models and lead to a lower generalization performance in the COGS benchmark. But the performance degradation is more extreme with novel embeddings, and the degradation increases with the amount of pretraining training data, highlighting an interesting case of inverse scaling.
Abstract: Human linguistic capacity is often characterized by compositionality and the generalization it enables -- human learners can produce and comprehend novel complex expressions by composing known parts. Several benchmarks exploit distributional control across training and test to gauge compositional generalization, where certain lexical items only occur in limited contexts during training. While recent work using these benchmarks suggests that pretrained models achieve impressive generalization performance, we argue that exposure to pretraining data may break the aforementioned distributional control. Using the COGS benchmark of Kim and Linzen (2020), we test two modified evaluation setups that control for this issue: (1) substituting context-controlled lexical items with novel character sequences, and (2) substituting them with special tokens represented by novel embeddings. We find that both of these setups lead to lower generalization performance in T5 (Raffel et al., 2020), suggesting that previously reported results have been overestimated due to uncontrolled lexical exposure during pretraining. The performance degradation is more extreme with novel embeddings, and the degradation increases with the amount of pretraining data, highlighting an interesting case of inverse scaling.
12 citations
01 Aug 2021
TL;DR: This paper proposed a multilingual contrastive pretraining (MCP) method to improve the performance of ML-LMs for cross-lingual commonsense reasoning, which significantly enhances sentence representations.
Abstract: Commonsense reasoning research has so far been limited to English. We aim to evaluate and improve popular multilingual language models (ML-LMs) to help advance commonsense reasoning (CSR) beyond English. We collect the Mickey corpus, consisting of 561k sentences in 11 different languages, which can be used for analyzing and improving ML-LMs. We propose Mickey Probe, a language-general probing task for fairly evaluating the common sense of popular ML-LMs across different languages. In addition, we also create two new datasets, X-CSQA and X-CODAH, by translating their English versions to 14 other languages, so that we can evaluate popular ML-LMs for cross-lingual commonsense reasoning. To improve the performance beyond English, we propose a simple yet effective method — multilingual contrastive pretraining (MCP). It significantly enhances sentence representations, yielding a large performance gain on both benchmarks (e.g., +2.7% accuracy for X-CSQA over XLM-R_L).
12 citations
01 Jan 2022
TL;DR: The CLEF 2022 SimpleText track as discussed by the authors addresses the challenges of text simplification approaches in the context of promoting scientific information access, by providing appropriate data and benchmarks, and creating a community of NLP and IR researchers working together to resolve one of the greatest challenges of today.
Abstract: The Web and social media have become the main source of information for citizens, with the risk that users rely on shallow information in sources prioritizing commercial or political incentives rather than the correctness and informational value. Non-experts tend to avoid scientific literature due to its complex language or their lack of prior background knowledge. Text simplification promises to remove some of these barriers. The CLEF 2022 SimpleText track addresses the challenges of text simplification approaches in the context of promoting scientific information access, by providing appropriate data and benchmarks, and creating a community of NLP and IR researchers working together to resolve one of the greatest challenges of today. The track will use a corpus of scientific literature abstracts and popular science requests. It features three tasks. First, content selection (what is in, or out?) challenges systems to select passages to include in a simplified summary in response to a query. Second, complexity spotting (what is unclear?) given a passage and a query, aims to rank terms/concepts that are required to be explained for understanding this passage (definitions, context, applications). Third, text simplification (rewrite this!) given a query, asks to simplify passages from scientific abstracts while preserving the main content.
12 citations
References
More filters
Proceedings Article•
12 Jun 2017TL;DR: This paper proposed a simple network architecture based solely on an attention mechanism, dispensing with recurrence and convolutions entirely and achieved state-of-the-art performance on English-to-French translation.
Abstract: The dominant sequence transduction models are based on complex recurrent orconvolutional neural networks in an encoder and decoder configuration. The best performing such models also connect the encoder and decoder through an attentionm echanisms. We propose a novel, simple network architecture based solely onan attention mechanism, dispensing with recurrence and convolutions entirely.Experiments on two machine translation tasks show these models to be superiorin quality while being more parallelizable and requiring significantly less timeto train. Our single model with 165 million parameters, achieves 27.5 BLEU onEnglish-to-German translation, improving over the existing best ensemble result by over 1 BLEU. On English-to-French translation, we outperform the previoussingle state-of-the-art with model by 0.7 BLEU, achieving a BLEU score of 41.1.
52,856 citations
Posted Content•
TL;DR: It is found that BERT was significantly undertrained, and can match or exceed the performance of every model published after it, and the best model achieves state-of-the-art results on GLUE, RACE and SQuAD.
Abstract: Language model pretraining has led to significant performance gains but careful comparison between different approaches is challenging. Training is computationally expensive, often done on private datasets of different sizes, and, as we will show, hyperparameter choices have significant impact on the final results. We present a replication study of BERT pretraining (Devlin et al., 2019) that carefully measures the impact of many key hyperparameters and training data size. We find that BERT was significantly undertrained, and can match or exceed the performance of every model published after it. Our best model achieves state-of-the-art results on GLUE, RACE and SQuAD. These results highlight the importance of previously overlooked design choices, and raise questions about the source of recently reported improvements. We release our models and code.
13,994 citations
"mT5: A Massively Multilingual Pre-t..." refers methods in this paper
..., 2020b), and RoBERTa (Liu et al., 2019), respectively....
[...]
...It uses data in 26 languages from Wikipedia and CC-News (Liu et al., 2019)....
[...]
...XLM-R (Conneau et al., 2020) is an improved version of XLM based on the RoBERTa model (Liu et al., 2019)....
[...]
..., 2020) is an improved version of XLM based on the RoBERTa model (Liu et al., 2019)....
[...]
...Popular models of this type are mBERT (Devlin, 2018), mBART (Liu et al., 2020a), and XLM-R (Conneau et al., 2020), which are multilingual variants of BERT (Devlin et al., 2019), BART (Lewis et al., 2020b), and RoBERTa (Liu et al., 2019), respectively....
[...]
16 Jun 2016
TL;DR: The Stanford Question Answering Dataset (SQuAD) as mentioned in this paper is a reading comprehension dataset consisting of 100,000+ questions posed by crowdworkers on a set of Wikipedia articles, where the answer to each question is a segment of text from the corresponding reading passage.
Abstract: We present the Stanford Question Answering Dataset (SQuAD), a new reading comprehension dataset consisting of 100,000+ questions posed by crowdworkers on a set of Wikipedia articles, where the answer to each question is a segment of text from the corresponding reading passage. We analyze the dataset to understand the types of reasoning required to answer the questions, leaning heavily on dependency and constituency trees. We build a strong logistic regression model, which achieves an F1 score of 51.0%, a significant improvement over a simple baseline (20%). However, human performance (86.8%) is much higher, indicating that the dataset presents a good challenge problem for future research.
The dataset is freely available at this https URL
3,667 citations
01 Jul 2020
TL;DR: It is shown that pretraining multilingual language models at scale leads to significant performance gains for a wide range of cross-lingual transfer tasks, and the possibility of multilingual modeling without sacrificing per-language performance is shown for the first time.
Abstract: This paper shows that pretraining multilingual language models at scale leads to significant performance gains for a wide range of cross-lingual transfer tasks. We train a Transformer-based masked language model on one hundred languages, using more than two terabytes of filtered CommonCrawl data. Our model, dubbed XLM-R, significantly outperforms multilingual BERT (mBERT) on a variety of cross-lingual benchmarks, including +14.6% average accuracy on XNLI, +13% average F1 score on MLQA, and +2.4% F1 score on NER. XLM-R performs particularly well on low-resource languages, improving 15.7% in XNLI accuracy for Swahili and 11.4% for Urdu over previous XLM models. We also present a detailed empirical analysis of the key factors that are required to achieve these gains, including the trade-offs between (1) positive transfer and capacity dilution and (2) the performance of high and low resource languages at scale. Finally, we show, for the first time, the possibility of multilingual modeling without sacrificing per-language performance; XLM-R is very competitive with strong monolingual models on the GLUE and XNLI benchmarks. We will make our code and models publicly available.
3,248 citations
"mT5: A Massively Multilingual Pre-t..." refers background or methods in this paper
...XLM-R (Conneau et al., 2020) is an improved version of XLM based on the RoBERTa model (Liu et al., 2019)....
[...]
...Values used by prior work include α = 0.7 for mBERT (Devlin, 2018), α = 0.3 for XLM-R (Conneau et al., 2020), and α = 0.2 for MMNMT (Arivazhagan et al., 2019)....
[...]
...We therefore take the approach used in (Devlin, 2018; Conneau et al., 2020; Arivazhagan et al., 2019) and boost lower-resource languages by sampling examples according to the probability p(L) ∝ |L|α, where p(L) is the probability of sampling text from a given language during pre-training and |L| is the number of examples in the language....
[...]
...We therefore take the approach used in (Devlin, 2018; Conneau et al., 2020; Arivazhagan et al., 2019) and boost lower-resource languages by sampling examples according to the probability p(L) ∝ |L|α, where p(L) is the probability of sampling text from a given language during pre-training and |L| is…...
[...]
..., 2020a), and XLM-R (Conneau et al., 2020), which are multilingual variants of BERT (Devlin...
[...]
18 Jan 2018
TL;DR: Universal Language Model Fine-tuning (ULMFiT) as mentioned in this paper is an effective transfer learning method that can be applied to any task in NLP, and introduces techniques that are key for finetuning a language model.
Abstract: Inductive transfer learning has greatly impacted computer vision, but existing approaches in NLP still require task-specific modifications and training from scratch. We propose Universal Language Model Fine-tuning (ULMFiT), an effective transfer learning method that can be applied to any task in NLP, and introduce techniques that are key for fine-tuning a language model. Our method significantly outperforms the state-of-the-art on six text classification tasks, reducing the error by 18-24% on the majority of datasets. Furthermore, with only 100 labeled examples, it matches the performance of training from scratch on 100 times more data. We open-source our pretrained models and code.
2,128 citations