推理任务（例如答案句子选择（AS2）或事实验证）通常通过将基于变压器的模型作为单个句子对分类器来解决。最近的研究表明，这些任务受益于共同跨多个候选句子的建模依赖性。在本文中，我们首先表明，当用于多转化推理任务进行微调时，流行的预训练的变压器的性能很差。然后，我们提出了一个新的预训练目标，该目标对段落级的语义进行了对多个输入句子进行建模。我们对三个AS2和一个事实验证数据集的评估证明了我们的预训练技术优于传统技术的优势，用于变压器用作多用途推理任务的关节模型，以及用作句子对配方的跨编码器的优势这些任务。我们的代码和预培训模型将在https://github.com/amazon-research/wqa-multi-sentence-inference上发布。

We introduce a new language representation model called BERT, which stands for Bidirectional Encoder Representations from Transformers. Unlike recent language representation models (Peters et al., 2018a;Radford et al., 2018), BERT is designed to pretrain deep bidirectional representations from unlabeled text by jointly conditioning on both left and right context in all layers. As a result, the pre-trained BERT model can be finetuned with just one additional output layer to create state-of-the-art models for a wide range of tasks, such as question answering and language inference, without substantial taskspecific architecture modifications.BERT is conceptually simple and empirically powerful. It obtains new state-of-the-art results on eleven natural language processing tasks, including pushing the GLUE score to 80.5% (7.7% point absolute improvement), MultiNLI accuracy to 86.7% (4.6% absolute improvement), SQuAD v1.1 question answering Test F1 to 93.2 (1.5 point absolute improvement) and SQuAD v2.0 Test F1 to 83.1 (5.1 point absolute improvement).

我们研究了检查问题的事实，旨在识别给定索赔的真实性。具体而言，我们专注于事实提取和验证（发烧）及其伴随数据集的任务。该任务包括从维基百科检索相关文件（和句子）并验证文件中的信息是否支持或驳斥所索赔的索赔。此任务至关重要，可以是假新闻检测和医疗索赔验证等应用程序块。在本文中，我们以通过以结构化和全面的方式呈现文献来更好地了解任务的挑战。我们通过分析不同方法的技术视角并讨论发热数据集的性能结果，描述了所提出的方法，这是最熟悉的和正式结构化的数据集，就是事实提取和验证任务。我们还迄今为止迄今为止确定句子检索组件的有益损失函数的最大实验研究。我们的分析表明，采样负句对于提高性能并降低计算复杂性很重要。最后，我们描述了开放的问题和未来的挑战，我们激励了未来的任务研究。

远程变压器模型取得了令人鼓舞的令人鼓舞的结果，即长上下文问题应答（QA）任务。这些任务通常需要超过一个长文件的推理，并且他们受益于识别一组证据跨度（例如，句子），为解决问题提供支持证据。在这项工作中，我们提出了一种用于装备远程变压器的新方法，其具有额外的序列级目标，以便更好地识别支持证据跨度。我们通过提出FineTuning的额外对比监督信号来实现这一目标，鼓励模型通过最大化问题证据相似性来明确歧视来自消极的证据句。拟议的额外损失表现出三种不同强大的长情绪变压器模型的一致改进，跨两个具有挑战性的问题回答基准 - 热杆菌和Qasper。

Pre-trained Transformers currently dominate most NLP tasks. They impose, however, limits on the maximum input length (512 sub-words in BERT), which are too restrictive in the legal domain. Even sparse-attention models, such as Longformer and BigBird, which increase the maximum input length to 4,096 sub-words, severely truncate texts in three of the six datasets of LexGLUE. Simpler linear classifiers with TF-IDF features can handle texts of any length, require far less resources to train and deploy, but are usually outperformed by pre-trained Transformers. We explore two directions to cope with long legal texts: (i) modifying a Longformer warm-started from LegalBERT to handle even longer texts (up to 8,192 sub-words), and (ii) modifying LegalBERT to use TF-IDF representations. The first approach is the best in terms of performance, surpassing a hierarchical version of LegalBERT, which was the previous state of the art in LexGLUE. The second approach leads to computationally more efficient models at the expense of lower performance, but the resulting models still outperform overall a linear SVM with TF-IDF features in long legal document classification.

最近的工作表明，（1）增加输入长度或（2）增加模型大小可以提高基于变压器的神经模型的性能。在本文中，我们提出了一个名为Longt5的新模型，我们探讨了同时缩放输入长度和模型大小的效果。具体而言，我们综合了从长输入变压器（ETC）的关注思路，并采用了从摘要预训练（PEGASU）的预训练策略进入可扩展的T5架构。结果是我们称之为{\ EM瞬态全球}（TGLOBAL）的新关注机制，这些机制是模仿等本地/全球注意力机制，但不需要额外的侧面输入。我们能够实现最先进的结果，以若干摘要任务，优于问题应答任务的原始T5模型。

我们提供了从文本到文本变换器（T5）的第一次探索句子嵌入式。句子嵌入式广泛适用于语言处理任务。虽然T5在作为序列到序列映射问题的语言任务上实现令人印象深刻的性能，但目前尚不清楚如何从编码器解码器模型生成陈列嵌入的句子。我们调查三种方法提取T5句子嵌入方法：两个仅利用T5编码器，一个使用全T5编码器解码器模型。为了支持我们的调查，我们建立了一个新的句子代表转移基准，SentGlue，它将Senteval Toolkit扩展到粘合基准的九个任务。我们的编码器的型号优于Senteval和SentGlue传输任务的句子 - BERT和SIMCSE句子嵌入，包括语义文本相似性（STS）。发现从数百万到数十亿参数的缩放T5产生一致的进一步改进。最后，我们的编码器 - 解码器方法在使用句子嵌入时在STS上实现了新的最先进的。我们的模型在https://tfhub.dev/google/collections/sentence-t5/1发布。

Laws and their interpretations, legal arguments and agreements\ are typically expressed in writing, leading to the production of vast corpora of legal text. Their analysis, which is at the center of legal practice, becomes increasingly elaborate as these collections grow in size. Natural language understanding (NLU) technologies can be a valuable tool to support legal practitioners in these endeavors. Their usefulness, however, largely depends on whether current state-of-the-art models can generalize across various tasks in the legal domain. To answer this currently open question, we introduce the Legal General Language Understanding Evaluation (LexGLUE) benchmark, a collection of datasets for evaluating model performance across a diverse set of legal NLU tasks in a standardized way. We also provide an evaluation and analysis of several generic and legal-oriented models demonstrating that the latter consistently offer performance improvements across multiple tasks.

大型语言模型在各种任务上显示出令人印象深刻的几次结果。但是，当知识是此类结果的关键时，就像问题回答和事实检查之类的任务一样，似乎需要存储知识的大量参数计数。众所周知，检索增强模型可以在不需要多个参数的情况下在知识密集的任务上表现出色，但是目前尚不清楚它们是否在几个弹药设置中工作。在这项工作中，我们介绍了地图集，这是一个经过精心设计和预先训练的增强语言模型，能够通过很少的培训示例学习知识密集型任务。我们对包括MMLU，苏格兰短裙和归类等各种任务进行评估，并研究文档索引内容的影响，表明它可以很容易地进行更新。值得注意的是，在自然问题上仅使用64个示例在自然问题上达到超过42 \％的准确性，尽管参数少了50倍，但比540B参数模型的表现优于540b参数模型。

信息检索是自然语言处理中的重要组成部分，用于知识密集型任务，如问题应答和事实检查。最近，信息检索已经看到基于神经网络的密集检索器的出现，作为基于术语频率的典型稀疏方法的替代方案。这些模型在数据集和任务中获得了最先进的结果，其中提供了大型训练集。但是，它们不会很好地转移到没有培训数据的新域或应用程序，并且通常因未经监督的术语 - 频率方法（例如BM25）的术语频率方法而言。因此，自然问题是如果没有监督，是否有可能训练密集的索取。在这项工作中，我们探讨了对比学习的限制，作为培训无人监督的密集检索的一种方式，并表明它导致强烈的检索性能。更确切地说，我们在15个数据集中出现了我们的模型胜过BM25的Beir基准测试。此外，当有几千例的示例可用时，我们显示微调我们的模型，与BM25相比，这些模型导致强大的改进。最后，当在MS-Marco数据集上微调之前用作预训练时，我们的技术在Beir基准上获得最先进的结果。

查询聚焦的文本摘要（QFTS）任务旨在构建基于给定查询的文本文档摘要的构建系统。解决此任务的关键挑战是缺乏培训摘要模型的大量标记数据。在本文中，我们通过探索一系列域适应技术来解决这一挑战。鉴于最近在广泛的自然语言处理任务中进行预先接受的变压器模型的成功，我们利用此类模型为单文档和多文件方案的QFTS任务产生抽象摘要。对于域适应，我们使用预先训练的变压器的摘要模型应用了各种技术，包括转移学习，弱监督学习和远程监督。六个数据集的广泛实验表明，我们所提出的方法非常有效地为QFTS任务产生抽象摘要，同时在一组自动和人类评估指标上设置新的最先进的结果。

Recent work on open domain question answering (QA) assumes strong supervision of the supporting evidence and/or assumes a blackbox information retrieval (IR) system to retrieve evidence candidates. We argue that both are suboptimal, since gold evidence is not always available, and QA is fundamentally different from IR. We show for the first time that it is possible to jointly learn the retriever and reader from question-answer string pairs and without any IR system. In this setting, evidence retrieval from all of Wikipedia is treated as a latent variable. Since this is impractical to learn from scratch, we pre-train the retriever with an Inverse Cloze Task. We evaluate on open versions of five QA datasets. On datasets where the questioner already knows the answer, a traditional IR system such as BM25 is sufficient. On datasets where a user is genuinely seeking an answer, we show that learned retrieval is crucial, outperforming BM25 by up to 19 points in exact match.

Using a single model across various tasks is beneficial for training and applying deep neural sequence models. We address the problem of developing generalist representations of text that can be used to perform a range of different tasks rather than being specialised to a single application. We focus on processing short questions and developing an embedding for these questions that is useful on a diverse set of problems, such as question topic classification, equivalent question recognition, and question answering. This paper introduces QBERT, a generalist model for processing questions. With QBERT, we demonstrate how we can train a multi-task network that performs all question-related tasks and has achieved similar performance compared to its corresponding single-task models.

Transformers-based models, such as BERT, have been one of the most successful deep learning models for NLP. Unfortunately, one of their core limitations is the quadratic dependency (mainly in terms of memory) on the sequence length due to their full attention mechanism. To remedy this, we propose, BIGBIRD, a sparse attention mechanism that reduces this quadratic dependency to linear. We show that BIGBIRD is a universal approximator of sequence functions and is Turing complete, thereby preserving these properties of the quadratic, full attention model. Along the way, our theoretical analysis reveals some of the benefits of having O(1) global tokens (such as CLS), that attend to the entire sequence as part of the sparse attention mechanism. The proposed sparse attention can handle sequences of length up to 8x of what was previously possible using similar hardware. As a consequence of the capability to handle longer context, BIGBIRD drastically improves performance on various NLP tasks such as question answering and summarization. We also propose novel applications to genomics data.

在这项工作中，我们探索如何学习专用的语言模型，旨在学习从文本文件中学习关键词的丰富表示。我们在判别和生成设置中进行预训练变压器语言模型（LMS）的不同掩蔽策略。在歧视性设定中，我们引入了一种新的预训练目标 - 关键边界，用替换（kbir）infifiling，在使用Kbir预先训练的LM进行微调时显示出在Sota上的性能（F1中高达9.26点）的大量增益关键酶提取的任务。在生成设置中，我们为BART - 键盘介绍了一个新的预训练设置，可再现与CATSeq格式中的输入文本相关的关键字，而不是Denoised原始输入。这也导致在关键词中的性能（F1 @ M）中的性能（高达4.33点），用于关键正版生成。此外，我们还微调了在命名实体识别（ner），问题应答（qa），关系提取（重新），抽象摘要和达到与SOTA的可比性表现的预训练的语言模型，表明学习丰富的代表关键词确实有利于许多其他基本的NLP任务。

We present SpanBERT, a pre-training method that is designed to better represent and predict spans of text. Our approach extends BERT by (1) masking contiguous random spans, rather than random tokens, and (2) training the span boundary representations to predict the entire content of the masked span, without relying on the individual token representations within it. Span-BERT consistently outperforms BERT and our better-tuned baselines, with substantial gains on span selection tasks such as question answering and coreference resolution. In particular, with the same training data and model size as BERT large , our single model obtains 94.6% and 88.7% F1 on SQuAD 1.1 and 2.0 respectively. We also achieve a new state of the art on the OntoNotes coreference resolution task (79.6% F1), strong performance on the TACRED relation extraction benchmark, and even gains on GLUE. 1 * Equal contribution. 1 Our code and pre-trained models are available at https://github.com/facebookresearch/ SpanBERT.

Bidirectional Encoder Representations from Transformers (BERT; Devlin et al. 2019) represents the latest incarnation of pretrained language models which have recently advanced a wide range of natural language processing tasks. In this paper, we showcase how BERT can be usefully applied in text summarization and propose a general framework for both extractive and abstractive models. We introduce a novel document-level encoder based on BERT which is able to express the semantics of a document and obtain representations for its sentences. Our extractive model is built on top of this encoder by stacking several intersentence Transformer layers. For abstractive summarization, we propose a new fine-tuning schedule which adopts different optimizers for the encoder and the decoder as a means of alleviating the mismatch between the two (the former is pretrained while the latter is not). We also demonstrate that a two-staged fine-tuning approach can further boost the quality of the generated summaries. Experiments on three datasets show that our model achieves stateof-the-art results across the board in both extractive and abstractive settings. 1

Transformer-based models have pushed state of the art in many areas of NLP, but our understanding of what is behind their success is still limited. This paper is the first survey of over 150 studies of the popular BERT model. We review the current state of knowledge about how BERT works, what kind of information it learns and how it is represented, common modifications to its training objectives and architecture, the overparameterization issue and approaches to compression. We then outline directions for future research.

Entities, as important carriers of real-world knowledge, play a key role in many NLP tasks. We focus on incorporating entity knowledge into an encoder-decoder framework for informative text generation. Existing approaches tried to index, retrieve, and read external documents as evidence, but they suffered from a large computational overhead. In this work, we propose an encoder-decoder framework with an entity memory, namely EDMem. The entity knowledge is stored in the memory as latent representations, and the memory is pre-trained on Wikipedia along with encoder-decoder parameters. To precisely generate entity names, we design three decoding methods to constrain entity generation by linking entities in the memory. EDMem is a unified framework that can be used on various entity-intensive question answering and generation tasks. Extensive experimental results show that EDMem outperforms both memory-based auto-encoder models and non-memory encoder-decoder models.

Large transformer models can highly improve Answer Sentence Selection (AS2) tasks, but their high computational costs prevent their use in many real-world applications. In this paper, we explore the following research question: How can we make the AS2 models more accurate without significantly increasing their model complexity? To address the question, we propose a Multiple Heads Student architecture (named CERBERUS), an efficient neural network designed to distill an ensemble of large transformers into a single smaller model. CERBERUS consists of two components: a stack of transformer layers that is used to encode inputs, and a set of ranking heads; unlike traditional distillation technique, each of them is trained by distilling a different large transformer architecture in a way that preserves the diversity of the ensemble members. The resulting model captures the knowledge of heterogeneous transformer models by using just a few extra parameters. We show the effectiveness of CERBERUS on three English datasets for AS2; our proposed approach outperforms all single-model distillations we consider, rivaling the state-of-the-art large AS2 models that have 2.7x more parameters and run 2.5x slower. Code for our model is available at https://github.com/amazon-research/wqa-cerberus