We introduce Performers, Transformer architectures which can estimate regular (softmax) full-rank-attention Transformers with provable accuracy, but using only linear (as opposed to quadratic) space and time complexity, without relying on any priors such as sparsity or low-rankness. To approximate softmax attentionkernels, Performers use a novel Fast Attention Via positive Orthogonal Random features approach (FAVOR+), which may be of independent interest for scalable kernel methods. FAVOR+ can also be used to efficiently model kernelizable attention mechanisms beyond softmax. This representational power is crucial to accurately compare softmax with other kernels for the first time on large-scale tasks, beyond the reach of regular Transformers, and investigate optimal attention-kernels. Performers are linear architectures fully compatible with regular Transformers and with strong theoretical guarantees: unbiased or nearly-unbiased estimation of the attention matrix, uniform convergence and low estimation variance. We tested Performers on a rich set of tasks stretching from pixel-prediction through text models to protein sequence modeling. We demonstrate competitive results with other examined efficient sparse and dense attention methods, showcasing effectiveness of the novel attention-learning paradigm leveraged by Performers.

在这项工作中，我们介绍了内核化变压器，这是一个通用，可扩展的，数据驱动的框架，用于学习变压器中的内核功能。我们的框架将变压器内核作为光谱特征图之间的点产物近似，并通过学习光谱分布来学习内核。这不仅有助于学习通用的内核端到端，而且还可以减少变压器从二次到线性的时间和空间复杂性。我们表明，在准确性和计算效率方面，内核化的变压器实现了与现有的有效变压器体系结构相当的性能。我们的研究还表明，内核的选择对性能有重大影响，而内核学习变体是固定内核变压器的竞争替代方案，无论是长时间的序列任务。

在本文中，据我们所知，我们提供了将各种掩盖机制纳入变形金刚以可扩展方式融入变形金刚结构的第一种综合方法。我们表明，有关线性因果关注的最新结果（Choromanski等，2021）和对数线性RPE注意力（Luo等，2021）是这种一般机制的特殊情况。但是，通过将问题作为对未掩盖注意力的拓扑调制（基于图的）调制，我们以前获得了几个未知结果，包括有效的D维RPE掩盖和图形内掩蔽。我们利用许多数学技术，从光谱分析到动态编程和随机步行到新算法，以求解图形上的马尔可夫过程。我们提供相应的经验评估。

最近，提出了随机特征专注（RFA），以通过线性化指数核来近似线性时间和空间复杂性的软磁性注意力。在本文中，我们首先提出了一种新颖的观点，以通过将RFA重新铸造为自称的重要性采样器来理解这种近似值的偏见。这种观点进一步阐明了整个软磁注意的\ emph {nobaled}估计量，称为随机注意（RA）。RA通过特定的分布构建积极的随机特征，并享有极大的改善近似保真度，尽管表现出二次复杂性。通过结合RA中的表现力和RFA的效率，我们开发了一种新型的线性复杂性自我发项机制，称为线性随机注意（LARA）。跨各个领域的广泛实验表明，RA和LARA可显着提高RFA的性能。

变压器架构现在是序列建模任务的核心。注意机制是核心，它可以在序列中对长期依赖性进行有效的建模。最近，变压器已成功地应用于计算机视觉域，在该域中首先将2D图像分割成斑块，然后将其视为1D序列。然而，这种线性化会损害图像中空间位置的概念，该图像具有重要的视觉线索。为了弥合差距，我们提出了连锁反应，这是视觉变压器的次级注意机制。基于最近基于内核的有效注意机制，我们设计了一种新型的动态编程算法，该算法将不同令牌的贡献加重了与它们在线性观察到的2D空间中相对空间距离的查询的贡献。广泛的实验和分析证明了连锁反应对各种视觉任务的有效性。

多头注意力是最先进的变压器背后的推动力，它在各种自然语言处理（NLP）和计算机视觉任务中实现了出色的性能。已经观察到，对于许多应用，这些注意力头会学习冗余嵌入，并且大多数可以在不降低模型性能的情况下去除。受到这一观察的启发，我们提出了变压器的混合物（变压器-MGK）的混合物，这是一种新型的变压器架构，用每个头部的钥匙混合了变压器中的冗余头部。这些键的混合物遵循高斯混合模型，并使每个注意力头有效地集中在输入序列的不同部分上。与传统的变压器对应物相比，变压器-MGK会加速训练和推理，具有较少的参数，并且需要更少的拖船来计算，同时实现跨任务的可比性或更高的准确性。 Transformer-MGK也可以轻松扩展到线性注意力。我们从经验上证明了在一系列实用应用中变形金属MGK的优势，包括语言建模和涉及非常长序列的任务。在Wikitext-103和远程竞技场基准中，具有4个头部的变压器MGK具有与基线变压器具有8个头的可比性或更好的性能。

基于变压器的模型广泛用于自然语言处理（NLP）。变压器模型的核心是自我关注机制，它捕获了输入序列中的令牌对的相互作用，并在序列长度上逐步取决于逐行。在更长的序列上培训此类模型是昂贵的。在本文中，我们表明，基于局部敏感散列（LSH）的伯努利采样注意机制降低了这种模型到线性的二次复杂性。我们通过考虑自我关注作为与Bernoulli随机变量相关的单独令牌的总和来绕过二次成本，原则上可以通过单个哈希进行一次（尽管在实践中，这个数字可能是一个小常数）。这导致了有效的采样方案来估算依赖于LSH的特定修改的自我关注（以便在GPU架构上进行部署）。我们在标准512序列长度上评估了胶水基准的算法，在那里我们看到了相对于标准预磨削变压器的良好性能。在远程竞技场（LRA）基准中，为了评估长序列的性能，我们的方法实现了与Softmax自我关注的结果一致，但具有相当大的加速和内存节省，并且通常优于其他有效的自我关注方法。我们的代码可以在https://github.com/mlpen/yoso获得

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.

由于自我关注模块的二次空间和时间复杂性，基于变压器的模型在处理长序列中是不高的。为了解决此限制，建议通过分别通过低维投影和行选择来降低线性（模数对数因子）的二次复杂度。这两种型号本质上连接，并了解他们的连接，我们介绍了矩阵素描的理论框架。基于理论分析，我们提出了Skeinformer加速自我关注，进一步提高了三个精心设计的组件的自我关注的准确性：列采样，自适应行标准化和飞行员采样重新利用。关于长距离竞技场（LRA）基准的实验表明，我们的方法以始终如一的较小时间/空间占地面积优于替代方案。

过度分辨的神经网络概括井，但训练昂贵。理想情况下，人们希望减少其计算成本，同时保留其概括的益处。稀疏的模型培训是实现这一目标的简单和有希望的方法，但随着现有方法与准确性损失，慢速训练运行时的困难或困难，仍然存在挑战，仍然存在困难的挑战。核心问题是，在离散的一组稀疏矩阵上搜索稀疏性掩模是困难和昂贵的。为了解决此问题，我们的主要见解是通过具有称为蝴蝶矩阵产品的固定结构的固定结构来优化优化稀疏矩阵的连续超集。随着蝴蝶矩阵不是硬件效率，我们提出了简单的蝴蝶（块和平坦）的变体来利用现代硬件。我们的方法（像素化蝴蝶）使用基于扁平块蝴蝶和低秩矩阵的简单固定稀疏模式，以缩小大多数网络层（例如，注意，MLP）。我们经验验证了像素化蝴蝶比蝴蝶快3倍，加快培训，以实现有利的准确性效率权衡。在ImageNet分类和Wikitext-103语言建模任务中，我们的稀疏模型训练比致密的MLP - 混频器，视觉变压器和GPT-2媒体更快地训练高达2.5倍，没有精确下降。

序列建模的一个中心目标是设计一个单个原则模型，该模型可以解决各种方式和任务，尤其是在远程依赖方面的序列数据。尽管包括RNN，CNN和Transformers在内的传统模型具有用于捕获长期依赖性的专业变体，但它们仍然很难扩展到长时间的10000美元或更多步骤。通过模拟基本状态空间模型（SSM）\（x'（t）= ax（t）= ax（t） + bu（t），y（t）= cx（t） + du（t） + du（t）\ ), and showed that for appropriate choices of the state matrix \( A \), this system could handle long-range dependencies mathematically and empirically.但是，该方法具有过度的计算和内存需求，使其无法作为一般序列建模解决方案。我们根据SSM的新参数化提出了结构化状态空间序列模型（S4），并表明它可以比以前的方法更有效地计算出其理论强度。我们的技术涉及对\（a \）进行低级校正的调节，从而使其对角度稳定，并将SSM降低到库奇内核的精心研究的计算中。 S4在各种既定的基准测试范围内取得了强劲的经验结果，包括（i）在顺序CIFAR-10上的91 \％精度，没有数据增强或辅助损失，与较大的2-D Resnet相当，（ii）实质上关闭。在图像和语言建模任务上与变形金刚的差距，同时在远程竞技场基准的每个任务上执行每一代$ 60 \ times $ $（iii）sota，包括求解所有先前工作的挑战性path-x任务，而所有先前工作的长度为16K，同时与所有竞争对手一样高效。

由于其二次复杂性，是变压器中的关注模块，其是变压器中的重要组件不能高效地扩展到长序列。许多工作侧重于近似于尺寸的圆点 - 指数的软MAX功能，导致分二次甚至线性复杂性变压器架构。但是，我们表明这些方法不能应用于超出点的指数样式的更强大的注意模块，例如，具有相对位置编码（RPE）的变压器。由于在许多最先进的模型中，相对位置编码被用作默认，设计可以包含RPE的高效变压器是吸引人的。在本文中，我们提出了一种新颖的方法来加速对RPE的转化仪的关注计算在核心化的关注之上。基于观察到相对位置编码形成Toeplitz矩阵，我们数在数学上表明，可以使用快速傅里叶变换（FFT）有效地计算具有RPE的核化注意。使用FFT，我们的方法实现$ \ mathcal {o}（n \ log n）$时间复杂性。有趣的是，我们进一步证明使用相对位置编码适当地可以减轻香草群关注的培训不稳定问题。在广泛的任务上，我们经验证明我们的模型可以从头开始培训，没有任何优化问题。学习模型比许多高效的变压器变体更好地执行，并且在长序列制度中比标准变压器更快。

Large Transformer models routinely achieve state-of-the-art results on a number of tasks but training these models can be prohibitively costly, especially on long sequences. We introduce two techniques to improve the efficiency of Transformers. For one, we replace dot-product attention by one that uses locality-sensitive hashing, changing its complexity from O(L 2 ) to O(L log L), where L is the length of the sequence. Furthermore, we use reversible residual layers instead of the standard residuals, which allows storing activations only once in the training process instead of N times, where N is the number of layers. The resulting model, the Reformer, performs on par with Transformer models while being much more memory-efficient and much faster on long sequences.

Low-rank matrix approximations, such as the truncated singular value decomposition and the rank-revealing QR decomposition, play a central role in data analysis and scientific computing. This work surveys and extends recent research which demonstrates that randomization offers a powerful tool for performing low-rank matrix approximation. These techniques exploit modern computational architectures more fully than classical methods and open the possibility of dealing with truly massive data sets.This paper presents a modular framework for constructing randomized algorithms that compute partial matrix decompositions. These methods use random sampling to identify a subspace that captures most of the action of a matrix. The input matrix is then compressed-either explicitly or implicitly-to this subspace, and the reduced matrix is manipulated deterministically to obtain the desired low-rank factorization. In many cases, this approach beats its classical competitors in terms of accuracy, speed, and robustness. These claims are supported by extensive numerical experiments and a detailed error analysis.The specific benefits of randomized techniques depend on the computational environment. Consider the model problem of finding the k dominant components of the singular value decomposition of an m × n matrix. (i) For a dense input matrix, randomized algorithms require O(mn log(k)) floating-point operations (flops) in contrast with O(mnk) for classical algorithms. (ii) For a sparse input matrix, the flop count matches classical Krylov subspace methods, but the randomized approach is more robust and can easily be reorganized to exploit multi-processor architectures. (iii) For a matrix that is too large to fit in fast memory, the randomized techniques require only a constant number of passes over the data, as opposed to O(k) passes for classical algorithms. In fact, it is sometimes possible to perform matrix approximation with a single pass over the data.

变压器在长序列上是缓慢的，渴望记忆力，因为自我注意的时间和记忆复杂性在序列上是二次的。近似关注方法试图通过交易模型质量以降低计算复杂性来解决此问题，但通常无法实现墙壁锁定的加速。我们认为，缺失的原则是提出注意力算法，以考虑读取和在GPU记忆层次之间写入。我们提出了FlashAttention，这是一种IO意识的精确注意算法，该算法使用平铺来减少GPU高带宽内存（HBM）和GPU芯片SRAM之间的内存读数/写入/写入。我们分析了闪存的IO复杂性，表明它所需的HBM访问少于标准注意力，并且对于一系列SRAM尺寸而言是最佳的。我们还扩展了闪光词，以引起障碍物的注意，从而产生了比任何现有的近似关注方法更快的近似关注算法。闪存火车的变压器​​比现有基准快：与MLPERF 1.1训练速度记录相比，Bert-Large（第512秒）的端到端壁式锁定加速度为15％，GPT-2上的3 $ \ times $ speedup（seq） 。闪存表现和块状闪光词可在变压器中实现更长的上下文，从而产生更高质量的模型（GPT-2上的0.7更好的困惑和长期分类的6.4点升力）和全新的功能：第一个实现优于更好的Chance的变压器PATH-X挑战（Seq。Length16K，61.4％精度）和PATH-256（Seq。Length64K，63.1％精度）上的性能。

变压器已成为自然兰格格处理和视觉中许多任务的首选模型。在更有效地进行培训和部署变压器的最新努力已经确定了许多策略，以近似自我发挥作用矩阵，这是变压器体系结构中的关键模块。有效的想法包括各种预先指定的稀疏模式，低级基础扩展及其组合。在本文中，我们重新访问了小波等经典多分辨率分析（MRA）概念，在这种情况下，在这种情况下的潜在价值迄今仍未被逐渐解散。我们表明，基于现代硬件和实施挑战所告知的经验反馈和设计选择的简单近似值，最终在大多数感兴趣的标准中产生了基于MRA的自我注意力方法，具有出色的性能。我们进行了一系列广泛的实验，并证明该多分辨率方案的表现优于最有效的自我注意力建议，并且对短序列和长序列都有利。代码可在\ url {https://github.com/mlpen/mra-witchention}中获得。

Accurate uncertainty quantification is a major challenge in deep learning, as neural networks can make overconfident errors and assign high confidence predictions to out-of-distribution (OOD) inputs. The most popular approaches to estimate predictive uncertainty in deep learning are methods that combine predictions from multiple neural networks, such as Bayesian neural networks (BNNs) and deep ensembles. However their practicality in real-time, industrial-scale applications are limited due to the high memory and computational cost. Furthermore, ensembles and BNNs do not necessarily fix all the issues with the underlying member networks. In this work, we study principled approaches to improve uncertainty property of a single network, based on a single, deterministic representation. By formalizing the uncertainty quantification as a minimax learning problem, we first identify distance awareness, i.e., the model's ability to quantify the distance of a testing example from the training data, as a necessary condition for a DNN to achieve high-quality (i.e., minimax optimal) uncertainty estimation. We then propose Spectral-normalized Neural Gaussian Process (SNGP), a simple method that improves the distance-awareness ability of modern DNNs with two simple changes: (1) applying spectral normalization to hidden weights to enforce bi-Lipschitz smoothness in representations and (2) replacing the last output layer with a Gaussian process layer. On a suite of vision and language understanding benchmarks, SNGP outperforms other single-model approaches in prediction, calibration and out-of-domain detection. Furthermore, SNGP provides complementary benefits to popular techniques such as deep ensembles and data augmentation, making it a simple and scalable building block for probabilistic deep learning. Code is open-sourced at https://github.com/google/uncertainty-baselines

神经网络的经典发展主要集中在有限维欧基德空间或有限组之间的学习映射。我们提出了神经网络的概括，以学习映射无限尺寸函数空间之间的运算符。我们通过一类线性积分运算符和非线性激活函数的组成制定运营商的近似，使得组合的操作员可以近似复杂的非线性运算符。我们证明了我们建筑的普遍近似定理。此外，我们介绍了四类运算符参数化：基于图形的运算符，低秩运算符，基于多极图形的运算符和傅里叶运算符，并描述了每个用于用每个计算的高效算法。所提出的神经运营商是决议不变的：它们在底层函数空间的不同离散化之间共享相同的网络参数，并且可以用于零击超分辨率。在数值上，与现有的基于机器学习的方法，达西流程和Navier-Stokes方程相比，所提出的模型显示出卓越的性能，而与传统的PDE求解器相比，与现有的基于机器学习的方法有关的基于机器学习的方法。

Many real-world applications require the prediction of long sequence time-series, such as electricity consumption planning. Long sequence time-series forecasting (LSTF) demands a high prediction capacity of the model, which is the ability to capture precise long-range dependency coupling between output and input efficiently. Recent studies have shown the potential of Transformer to increase the prediction capacity. However, there are several severe issues with Transformer that prevent it from being directly applicable to LSTF, including quadratic time complexity, high memory usage, and inherent limitation of the encoder-decoder architecture. To address these issues, we design an efficient transformer-based model for LSTF, named Informer, with three distinctive characteristics: (i) a ProbSparse self-attention mechanism, which achieves O(L log L) in time complexity and memory usage, and has comparable performance on sequences' dependency alignment. (ii) the self-attention distilling highlights dominating attention by halving cascading layer input, and efficiently handles extreme long input sequences. (iii) the generative style decoder, while conceptually simple, predicts the long time-series sequences at one forward operation rather than a step-by-step way, which drastically improves the inference speed of long-sequence predictions. Extensive experiments on four large-scale datasets demonstrate that Informer significantly outperforms existing methods and provides a new solution to the LSTF problem.

Polynomial kernels are among the most popular kernels in machine learning, since their feature maps model the interactions between the dimensions of the input data. However, these features correspond to tensor products of the input with itself, which makes their dimension grow exponentially with the polynomial degree. We address this issue by proposing Complexto-Real (CtR) sketches for tensor products that can be used as random feature approximations of polynomial kernels. These sketches leverage intermediate complex random projections, leading to better theoretical guarantees and potentially much lower variances than analogs using real projections. Our sketches are simple to construct and their final output is real-valued, which makes their downstream use straightforward. Finally, we show that they achieve state-of-the-art performance in terms of accuracy and speed.