在基于变压器的模型中通常观察到令牌均匀性，在经过变压器中经过堆叠的多个自我发场层后，不同的令牌共享大量相似信息。在本文中，我们建议使用每个变压器层的输出的奇异值的分布来表征令牌均匀性的现象，并从经验上说明，偏斜的奇异值分布可以减轻“令牌均匀性”问题。基于我们的观察结果，我们定义了奇异值分布的几种理想特性，并提出了一种新的转换函数，以更新奇异值。我们表明，除了减轻令牌均匀性外，转换功能还应保留原始嵌入空间中的当地邻域结构。我们提出的奇异价值变换函数应用于伯特，阿尔伯特，罗伯塔和德文尔特等一系列基于变压器的语言模型，并且在语义文本相似性评估和一系列胶水任务中观察到了改善的性能。我们的源代码可在https://github.com/hanqi-qi/tokenuni.git上找到。

This paper presents SimCSE, a simple contrastive learning framework that greatly advances state-of-the-art sentence embeddings. We first describe an unsupervised approach, which takes an input sentence and predicts itself in a contrastive objective, with only standard dropout used as noise. This simple method works surprisingly well, performing on par with previous supervised counterparts. We find that dropout acts as minimal data augmentation, and removing it leads to a representation collapse. Then, we propose a supervised approach, which incorporates annotated pairs from natural language inference datasets into our contrastive learning framework by using "entailment" pairs as positives and "contradiction" pairs as hard negatives. We evaluate SimCSE on standard semantic textual similarity (STS) tasks, and our unsupervised and supervised models using BERT base achieve an average of 76.3% and 81.6% Spearman's correlation respectively, a 4.2% and 2.2% improvement compared to the previous best results. We also show-both theoretically and empirically-that the contrastive learning objective regularizes pre-trained embeddings' anisotropic space to be more uniform, and it better aligns positive pairs when supervised signals are available. 1 2 We randomly sample 10 6 sentences from English Wikipedia and fine-tune BERTbase with learning rate = 3e-5, N = 64. In all our experiments, no STS training sets are used.

大规模预训练的语言模型的出现为自然语言处理的最新进展做出了巨大贡献。许多最先进的语言模型首先在大型文本语料库上进行培训，然后在下游任务上进行微调。尽管它最近获得了成功和广泛的采用，但对预训练的语言模型的微调通常会遭受过度拟合，这会导致由于模型的复杂性极高的复杂性和下游任务的有限培训样本而导致的普遍性差。为了解决这个问题，我们提出了一个新颖有效的微调框架，称为Layerwise噪声稳定性正则化（LNSR）。具体而言，我们建议注入标准的高斯噪声或势内噪声，并将微调模型的隐藏表示形式定向。我们首先提供理论分析以支持我们方法的功效。然后，我们证明了所提出的方法的优势，而不是其他最先进的算法，包括L2-SP，MixOut和Smart。尽管这些先前的作品仅验证其方法对相对简单的文本分类任务的有效性，但我们还验证了方法对问题答案任务的有效性，而目标问题更加困难，并且可以使用更多的培训示例。此外，广泛的实验结果表明，所提出的算法不仅可以提高语言模型的内域性能，而且还可以改善域外数据的域概括性能。

近年来，基于变压器的预训练模型已获得了很大的进步，成为自然语言处理中最重要的骨干之一。最近的工作表明，变压器内部的注意力机制可能不需要，卷积神经网络和基于多层感知器的模型也已被研究为变压器替代方案。在本文中，我们考虑了一个用于语言模型预训练的图形循环网络，该网络通过本地令牌级通信为每个序列构建一个图形结构，以及与其他代币解耦的句子级表示。原始模型在受监督培训下的特定领域特定文本分类中表现良好，但是，其通过自我监督的方式学习转移知识的潜力尚未得到充分利用。我们通过优化体系结构并验证其在更通用的语言理解任务（英语和中文）中的有效性来填补这一空白。至于模型效率，我们的模型在基于变压器的模型中而不是二次复杂性，而是具有线性复杂性，并且在推断过程中的性能更有效。此外，我们发现与现有基于注意力的模型相比，我们的模型可以生成更多样化的输出，而背景化的功能冗余性较小。

具有许多预训练模型（PTM）的模型中心已经是深度学习的基石。尽管以高成本建造，但它们仍然保持\ emph {探索}：从业人员通常会通过普及从提供的模型中心中选择一个PTM，然后对PTM进行微调以解决目标任务。这种na \“我的但共同的实践构成了两个障碍，以充分利用预训练的模型中心：（1）通过受欢迎程度选择的PTM选择没有最佳保证；（2）仅使用一个PTM，而其余的PTM则被忽略。理想情况下。理想情况下。 ，为了最大程度地利用预训练的模型枢纽，需要尝试所有PTM的所有组合和广泛的微调每个PTM组合，这会产生指数组合和不可偿还的计算预算。在本文中，我们提出了一种新的范围排名和调整预训练的模型：（1）我们的会议论文〜\ citep {you_logme：_2021}提出的logMe，以估算预先训练模型提取的标签证据的最大值，该标签证据可以在模型中排名所有PTMS用于各种类型的PTM和任务的枢纽\ Emph {微调之前}。（2）如果我们不偏爱模型的体系结构，则可以对排名最佳的PTM进行微调和部署，或者可以通过TOPE调整目标PTM -k通过t排名PTM他提出了b-tuning算法。排名部分基于会议论文，我们在本文中完成了其理论分析，包括启发式证据最大化程序的收敛证明和特征维度的影响。调整零件引入了一种用于调整多个PTM的新型贝叶斯调整（B-Tuning）方法，该方法超过了专门的方法，该方法旨在调整均匀的PTMS，并为调整异质PTMS设置了一种新的技术。利用PTM枢纽的新范式对于整个机器学习社区的大量受众来说可能会很有趣。

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。

近年来，人们对开发自然语言处理（NLP）中可解释模型的利益越来越多。大多数现有模型旨在识别输入功能，例如对于模型预测而言重要的单词或短语。然而，在NLP中开发的神经模型通常以层次结构的方式构成单词语义，文本分类需要层次建模来汇总本地信息，以便处理主题和标签更有效地转移。因此，单词或短语的解释不能忠实地解释文本分类中的模型决策。本文提出了一种新型的层次解释性神经文本分类器，称为提示，该分类器可以自动以层次结构方式以标记相关主题的形式生成模型预测的解释。模型解释不再处于单词级别，而是基于主题作为基本语义单元。评论数据集和新闻数据集的实验结果表明，我们所提出的方法与现有最新的文本分类器相当地达到文本分类结果，并比其他可解释的神经文本更忠实于模型的预测和更好地理解人类的解释分类器。

Incorporating contrastive learning objectives in sentence representation learning (SRL) has yielded significant improvements on many sentence-level NLP tasks. However, It is not well understood why contrastive learning works for learning sentence-level semantics. In this paper, we take a closer look at contrastive sentence representation learning through the lens of isotropy and learning dynamics. We interpret its success stories through the geometry of the representation shifts. We show that contrastive learning brings isotropy, and surprisingly learns to converge tokens to similar positions in the semantic space if given the signal that they are in the same sentence. Also, what we formalize as "spurious contextualization" is mitigated for semantically meaningful tokens, while augmented for functional ones. The embedding space is pushed toward the origin during training, with more areas now better defined. We ablate these findings by observing the learning dynamic with different training temperatures, batch sizes and pooling methods. With these findings, we aim to shed light on future designs of sentence representation learning methods.

变压器模型不仅在自然语言处理（NLP）中成功，而且还在计算机视觉（CV）中表现出高潜力。尽管提前很大，但大多数作品只关注建筑的改进，但很少关注分类头。多年来，变压器模型专门用于分类令牌来构建最终分类器，而无明确地利用高级字标记。在本文中，我们提出了一种名为二阶变压器（SOT）的新型变压器模型，同时利用分类器的分类令牌和单词令牌。具体地，我们经验披露了高级词令牌包含丰富的信息，其本身是对分类器非常竞争的，而且与分类令牌互补。为了有效地利用这种丰富的信息，我们提出了具有奇异值功率标准化的多头全球交叉协方差汇集，其符合相似的哲学，因此与变压器块兼容，比常用的汇集方法更好。然后，我们全面地研究了如何将单词令牌与分类令牌进行了解，以构建最终分类头。对于CV任务，我们的SOT显着提高了最先进的视觉变压器，以挑战基准，包括想象成和想象力-A。对于NLP任务，通过基于预磨料语言变压器的微调，我们的SOT大大提高了广泛使用的任务等性能，如可乐和RTE。代码将在https://peihuali.org/sot提供

Recent progress in pre-trained neural language models has significantly improved the performance of many natural language processing (NLP) tasks. In this paper we propose a new model architecture DeBERTa (Decoding-enhanced BERT with disentangled attention) that improves the BERT and RoBERTa models using two novel techniques. The first is the disentangled attention mechanism, where each word is represented using two vectors that encode its content and position, respectively, and the attention weights among words are computed using disentangled matrices on their contents and relative positions, respectively. Second, an enhanced mask decoder is used to incorporate absolute positions in the decoding layer to predict the masked tokens in model pre-training. In addition, a new virtual adversarial training method is used for fine-tuning to improve models' generalization. We show that these techniques significantly improve the efficiency of model pre-training and the performance of both natural language understand (NLU) and natural langauge generation (NLG) downstream tasks. Compared to RoBERTa-Large, a DeBERTa model trained on half of the training data performs consistently better on a wide range of NLP tasks, achieving improvements on MNLI by +0.9% (90.2% vs. 91.1%), on SQuAD v2.0 by +2.3% (88.4% vs. 90.7%) and RACE by +3.6% (83.2% vs. 86.8%). Notably, we scale up DeBERTa by training a larger version that consists of 48 Transform layers with 1.5 billion parameters. The significant performance boost makes the single DeBERTa model surpass the human performance on the SuperGLUE benchmark (Wang et al., 2019a) for the first time in terms of macro-average score (89.9 versus 89.8), and the ensemble DeBERTa model sits atop the SuperGLUE leaderboard as of January 6, 2021, outperforming the human baseline by a decent margin (90.3 versus 89.8). The pre-trained DeBERTa models and the source code were released at: https://github.com/microsoft/DeBERTa 1 .

Contrastive learning has become a new paradigm for unsupervised sentence embeddings. Previous studies focus on instance-wise contrastive learning, attempting to construct positive pairs with textual data augmentation. In this paper, we propose a novel Contrastive learning method with Prompt-derived Virtual semantic Prototypes (ConPVP). Specifically, with the help of prompts, we construct virtual semantic prototypes to each instance, and derive negative prototypes by using the negative form of the prompts. Using a prototypical contrastive loss, we enforce the anchor sentence embedding to be close to its corresponding semantic prototypes, and far apart from the negative prototypes as well as the prototypes of other sentences. Extensive experimental results on semantic textual similarity, transfer, and clustering tasks demonstrate the effectiveness of our proposed model compared to strong baselines. Code is available at https://github.com/lemon0830/promptCSE.

对比学习一直吸引着学习无监督的句子嵌入。当前的最新无监督方法是无监督的SIMCSE（UNSUP-SIMCSE）。 Unsup-Simcse将辍学作为最小数据增强方法，并将相同的输入句子传递给预训练的变压器编码器（带有掉落的掉落）两次，以获取两个相应的嵌入式以构建正对。由于句子的长度信息通常会由于使用嵌入变压器中的位置嵌入而编码到句子嵌入中，因此Unsup-Simcse中的每个正对实际上包含相同的长度信息。因此，接受这些正面对训练的Unsup-Simcse可能是有偏见的，这往往会考虑到语义上相同长度或相似长度的句子更相似。通过统计观察，我们发现Unsup-Simcse确实存在这样的问题。为了减轻它，我们应用了一个简单的重复操作来修改输入句子，然后分别将输入句子及其修改后的对应物传递给预训练的变压器编码器，以获取阳性对。此外，我们从计算机视觉社区中汲取灵感，并引入动量对比度，从而扩大了负面对的数量，而没有其他计算。提出的两种修改分别应用于正和负对，并构建一种新的句子嵌入方法，称为增强的Unsup-Simcse（ESIMCSE）。我们在几个基准数据集W.R.T上评估了所提出的ESIMCSE，语义文本相似性（STS）任务。实验结果表明，ESIMCSE的表现优于最先进的undup-Simcse，而Bert基碱的平均长矛相关性为2.02％。

在以前的作品中广泛讨论了句子语义相似性的原始伯特的表现不佳。我们发现不满意的性能主要是由于静态令牌嵌入偏差和无效的伯特层，而不是姓氏的高余弦相似性。为此，我们提出了一个迅速的句子嵌入方法，可以减少令牌嵌入偏差，使原始伯特层更有效。通过将句子嵌入式任务重新塑造为填充空白问题，我们的方法显着提高了原始伯特的性能。我们讨论了两个提示符，表示基于及时的句子嵌入的三个提示搜索方法。此外，我们提出了一种通过模板去噪技术的新型无监督培训目标，这大大缩短了监督和无人监督的环境之间的性能差距。对于实验，我们评估我们在非微调和微调的设置上的方法。即使是非微调方法也可以优于STS任务上的无监督服务器等微调的方法。我们的微调方法在无监督和监督设置中优于最先进的方法SIMCSE。与SIMCSE相比，我们分别在无监督环境下实现了2.29和2.58点的伯特和罗伯塔的改进。

最近编码的位置已显示在变压器体系结构中有效。它为序列不同位置的元素之间的依赖性建模提供了宝贵的监督。在本文中，我们首先研究了各种方法，以将位置信息整合到基于变压器的语言模型的学习过程中。然后，我们提出了一种名为旋转位置嵌入（绳索）的新颖方法，以有效利用位置信息。具体而言，提议的绳索用旋转矩阵编码绝对位置，同时将显式相对位置依赖性在自我发项公式中。值得注意的是，绳索具有宝贵的特性，包括序列长度的灵活性，衰减的相互依赖性随着相对距离的增加以及将线性自我注意力配备相对位置编码的能力。最后，我们在各种长文本分类基准数据集上使用旋转位置嵌入（也称为Roformer）评估增强的变压器。我们的实验表明，它始终如一地克服了其替代方案。此外，我们提供了理论分析来解释一些实验结果。 Roformer已经集成到HuggingFace：\ url {https://huggingface.co/docs/transformers/model_doc/roformer}。

Attention-based neural networks, such as Transformers, have become ubiquitous in numerous applications, including computer vision, natural language processing, and time-series analysis. In all kinds of attention networks, the attention maps are crucial as they encode semantic dependencies between input tokens. However, most existing attention networks perform modeling or reasoning based on representations, wherein the attention maps of different layers are learned separately without explicit interactions. In this paper, we propose a novel and generic evolving attention mechanism, which directly models the evolution of inter-token relationships through a chain of residual convolutional modules. The major motivations are twofold. On the one hand, the attention maps in different layers share transferable knowledge, thus adding a residual connection can facilitate the information flow of inter-token relationships across layers. On the other hand, there is naturally an evolutionary trend among attention maps at different abstraction levels, so it is beneficial to exploit a dedicated convolution-based module to capture this process. Equipped with the proposed mechanism, the convolution-enhanced evolving attention networks achieve superior performance in various applications, including time-series representation, natural language understanding, machine translation, and image classification. Especially on time-series representation tasks, Evolving Attention-enhanced Dilated Convolutional (EA-DC-) Transformer outperforms state-of-the-art models significantly, achieving an average of 17% improvement compared to the best SOTA. To the best of our knowledge, this is the first work that explicitly models the layer-wise evolution of attention maps. Our implementation is available at https://github.com/pkuyym/EvolvingAttention

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).

The Transformer is widely used in natural language processing tasks. To train a Transformer however, one usually needs a carefully designed learning rate warm-up stage, which is shown to be crucial to the final performance but will slow down the optimization and bring more hyperparameter tunings. In this paper, we first study theoretically why the learning rate warm-up stage is essential and show that the location of layer normalization matters. Specifically, we prove with mean field theory that at initialization, for the original-designed Post-LN Transformer, which places the layer normalization between the residual blocks, the expected gradients of the parameters near the output layer are large. Therefore, using a large learning rate on those gradients makes the training unstable. The warm-up stage is practically helpful for avoiding this problem. On the other hand, our theory also shows that if the layer normalization is put inside the residual blocks (recently proposed as Pre-LN Transformer), the gradients are well-behaved at initialization. This motivates us to remove the warm-up stage for the training of Pre-LN Transformers. We show in our experiments that Pre-LN Transformers without the warm-up stage can reach comparable results with baselines while requiring significantly less training time and hyper-parameter tuning on a wide range of applications.

数据增强是通过转换为机器学习的人工创建数据的人工创建，是一个跨机器学习学科的研究领域。尽管它对于增加模型的概括功能很有用，但它还可以解决许多其他挑战和问题，从克服有限的培训数据到正规化目标到限制用于保护隐私的数据的数量。基于对数据扩展的目标和应用的精确描述以及现有作品的分类法，该调查涉及用于文本分类的数据增强方法，并旨在为研究人员和从业者提供简洁而全面的概述。我们将100多种方法划分为12种不同的分组，并提供最先进的参考文献来阐述哪种方法可以通过将它们相互关联，从而阐述了哪种方法。最后，提供可能构成未来工作的基础的研究观点。

Metric-based meta-learning is one of the de facto standards in few-shot learning. It composes of representation learning and metrics calculation designs. Previous works construct class representations in different ways, varying from mean output embedding to covariance and distributions. However, using embeddings in space lacks expressivity and cannot capture class information robustly, while statistical complex modeling poses difficulty to metric designs. In this work, we use tensor fields (``areas'') to model classes from the geometrical perspective for few-shot learning. We present a simple and effective method, dubbed hypersphere prototypes (HyperProto), where class information is represented by hyperspheres with dynamic sizes with two sets of learnable parameters: the hypersphere's center and the radius. Extending from points to areas, hyperspheres are much more expressive than embeddings. Moreover, it is more convenient to perform metric-based classification with hypersphere prototypes than statistical modeling, as we only need to calculate the distance from a data point to the surface of the hypersphere. Following this idea, we also develop two variants of prototypes under other measurements. Extensive experiments and analysis on few-shot learning tasks across NLP and CV and comparison with 20+ competitive baselines demonstrate the effectiveness of our approach.

Text clustering and topic extraction are two important tasks in text mining. Usually, these two tasks are performed separately. For topic extraction to facilitate clustering, we can first project texts into a topic space and then perform a clustering algorithm to obtain clusters. To promote topic extraction by clustering, we can first obtain clusters with a clustering algorithm and then extract cluster-specific topics. However, this naive strategy ignores the fact that text clustering and topic extraction are strongly correlated and follow a chicken-and-egg relationship. Performing them separately fails to make them mutually benefit each other to achieve the best overall performance. In this paper, we propose an unsupervised text clustering and topic extraction framework (ClusTop) which integrates text clustering and topic extraction into a unified framework and can achieve high-quality clustering result and extract topics from each cluster simultaneously. Our framework includes four components: enhanced language model training, dimensionality reduction, clustering and topic extraction, where the enhanced language model can be viewed as a bridge between clustering and topic extraction. On one hand, it provides text embeddings with a strong cluster structure which facilitates effective text clustering; on the other hand, it pays high attention on the topic related words for topic extraction because of its self-attention architecture. Moreover, the training of enhanced language model is unsupervised. Experiments on two datasets demonstrate the effectiveness of our framework and provide benchmarks for different model combinations in this framework.