Transformers are being used extensively across several sequence modeling tasks. Significant research effort has been devoted to experimentally probe the inner workings of Transformers. However, our conceptual and theoretical understanding of their power and inherent limitations is still nascent. In particular, the roles of various components in Transformers such as positional encodings, attention heads, residual connections, and feedforward networks, are not clear. In this paper, we take a step towards answering these questions. We analyze the computational power as captured by Turing-completeness. We first provide an alternate and simpler proof to show that vanilla Transformers are Turing-complete and then we prove that Transformers with only positional masking and without any positional encoding are also Turing-complete. We further analyze the necessity of each component for the Turing-completeness of the network; interestingly, we find that a particular type of residual connection is necessary. We demonstrate the practical implications of our results via experiments on machine translation and synthetic tasks.
College of Intelligence and Computing, Tianjin University（天津大学智能与计算学部）；Microsoft Research Asia（微软亚洲研究院）；School of Computer Science and Technology, Beijing Institute of Technology（北京理工大学计算机学院）
Latest development of neural models has connected the encoder and decoder through a self-attention mechanism. In particular, Transformer, which is solely based on self-attention, has led to breakthroughs in Natural Language Processing (NLP) tasks. However, the multi-head attention mechanism, as a key component of Transformer, limits the effective deployment of the model to a resource-limited setting. In this paper, based on the ideas of tensor decomposition and parameters sharing, we propose a novel self-attention model (namely Multi-linear attention) with Block-Term Tensor Decomposition (BTD)…
Large Transformers have achieved state-of-the-art performance across many tasks. Most open-source libraries on scaling Transformers focus on improving training or inference with better parallelization. In this work, we present TorchScale, an open-source toolkit that allows researchers and developers to scale up Transformers efficiently and effectively. TorchScale has the implementation of several modeling techniques, which can improve modeling generality and capability, as well as training stability and efficiency. Experimental results on language modeling and neural machine translation demons…
Shanghai Jiao Tong University（上海交通大学）；Massachusetts Institute of Technology（麻省理工学院）
Transformer has become ubiquitous in natural language processing (e.g., machine translation, question answering); however, it requires enormous amount of computations to achieve high performance, which makes it not suitable for mobile applications that are tightly constrained by the hardware resources and battery. In this paper, we present an efficient mobile NLP architecture, Lite Transformer to facilitate deploying mobile NLP applications on edge devices. The key primitive is the Long-Short Range Attention (LSRA), where one group of heads specializes in the local context modeling (by convolu…
Bidirectional Encoder Representations from Transformers (BERT) has recently achieved state-of-the-art performance on a broad range of NLP tasks including sentence classification, machine translation, and question answering. The BERT model architecture is derived primarily from the transformer. Prior to the transformer era, bidirectional Long Short-Term Memory (BLSTM) has been the dominant modeling architecture for neural machine translation and question answering. In this paper, we investigate how these two modeling techniques can be combined to create a more powerful model architecture. We pr…
University of Mannheim（曼海姆大学）；Educational Testing Service (ETS)（教育考试服务中心（ETS））
Breaking down the structure of long texts into semantically coherent segments makes the texts more readable and supports downstream applications like summarization and retrieval. Starting from an apparent link between text coherence and segmentation, we introduce a novel supervised model for text segmentation with simple but explicit coherence modeling. Our model -- a neural architecture consisting of two hierarchically connected Transformer networks -- is a multi-task learning model that couples the sentence-level segmentation objective with the coherence objective that differentiates correct…
Yandex（Yandex（亚德斯））；University of Amsterdam（阿姆斯特丹大学）；University of Edinburgh（爱丁堡大学）；University of Zurich（苏黎世大学）
We seek to understand how the representations of individual tokens and the structure of the learned feature space evolve between layers in deep neural networks under different learning objectives. We focus on the Transformers for our analysis as they have been shown effective on various tasks, including machine translation (MT), standard left-to-right language models (LM) and masked language modeling (MLM). Previous work used black-box probing tasks to show that the representations learned by the Transformer differ significantly depending on the objective. In this work, we use canonical correl…
Florida State University（佛罗里达州立大学）；Tencent AI Lab（腾讯人工智能实验室）；University of Macau（澳门大学）
Recently, the Transformer model that is based solely on attention mechanisms, has advanced the state-of-the-art on various machine translation tasks. However, recent studies reveal that the lack of recurrence hinders its further improvement of translation capacity. In response to this problem, we propose to directly model recurrence for Transformer with an additional recurrence encoder. In addition to the standard recurrent neural network, we introduce a novel attentive recurrent network to leverage the strengths of both attention and recurrent networks. Experimental results on the widely-used…
Transformers have recently taken the center stage in language modeling after LSTM's were considered the dominant model architecture for a long time. In this project, we investigate the performance of the Transformer architectures-BERT and Transformer-XL for the language modeling task. We use a sub-word model setting with the Finnish language and compare it to the previous State of the art (SOTA) LSTM model. BERT achieves a pseudo-perplexity score of 14.5, which is the first such measure achieved as far as we know. Transformer-XL improves upon the perplexity score to 73.58 which is 27\% better …
University of Waterloo（滑铁卢大学）；Peking University（北京大学）
Transformers are powerful for sequence modeling. Nearly all state-of-the-art language models and pre-trained language models are based on the Transformer architecture. However, it distinguishes sequential tokens only with the token position index. We hypothesize that better contextual representations can be generated from the Transformer with richer positional information. To verify this, we propose a segment-aware Transformer (Segatron), by replacing the original token position encoding with a combined position encoding of paragraph, sentence, and token. We first introduce the segment-aware m…
We explore deep autoregressive Transformer models in language modeling for speech recognition. We focus on two aspects. First, we revisit Transformer model configurations specifically for language modeling. We show that well configured Transformer models outperform our baseline models based on the shallow stack of LSTM recurrent neural network layers. We carry out experiments on the open-source LibriSpeech 960hr task, for both 200K vocabulary word-level and 10K byte-pair encoding subword-level language modeling. We apply our word-level models to conventional hybrid speech recognition by lattic…
Columbia University in the City of New York（纽约哥伦比亚大学）
We propose a way to use a transformer-based language model in conversational speech recognition. Specifically, we focus on decoding efficiently in a weighted finite-state transducer framework. We showcase an approach to lattice re-scoring that allows for longer range history captured by a transfomer-based language model and takes advantage of a transformer's ability to avoid computing sequentially.
Google（谷歌）；University of Cambridge（剑桥大学）；Alan Turing Institute（艾伦·图灵研究所）；DeepMind（深度思维）
Transformer models have achieved state-of-the-art results across a diverse range of domains. However, concern over the cost of training the attention mechanism to learn complex dependencies between distant inputs continues to grow. In response, solutions that exploit the structure and sparsity of the learned attention matrix have blossomed. However, real-world applications that involve long sequences, such as biological sequence analysis, may fall short of meeting these assumptions, precluding exploration of these models. To address this challenge, we present a new Transformer architecture, Pe…
Recent studies have demonstrated that the performance of transformers on the task of language modeling obeys a power-law relationship with model size over six orders of magnitude. While transformers exhibit impressive scaling, their performance hinges on processing large amounts of data, and their computational and memory requirements grow quadratically with sequence length. Motivated by these considerations, we construct a Legendre Memory Unit based model that introduces a general prior for sequence processing and exhibits an $O(n)$ and $O(n \ln n)$ (or better) dependency for memory and compu…
Skolkovo Institute of Science and Technology（斯科尔科沃科学技术研究院）
The identification of nonlinear dynamics from observations is essential for the alignment of the theoretical ideas and experimental data. The last, in turn, is often corrupted by the side effects and noise of different natures, so probabilistic approaches could give a more general picture of the process. At the same time, high-dimensional probabilities modeling is a challenging and data-intensive task. In this paper, we establish a parallel between the dynamical systems modeling and language modeling tasks. We propose a transformer-based model that incorporates geometrical properties of the da…
While deep learning has received a surge of interest in a variety of fields in recent years, major deep learning models barely use complex numbers. However, speech, signal and audio data are naturally complex-valued after Fourier Transform, and studies have shown a potentially richer representation of complex nets. In this paper, we propose a Complex Transformer, which incorporates the transformer model as a backbone for sequence modeling; we also develop attention and encoder-decoder network operating for complex input. The model achieves state-of-the-art performance on the MusicNet dataset a…
This work applies natural language modeling to generate plausible strategic moves in the ancient game of Go. We train the Generative Pretrained Transformer (GPT-2) to mimic the style of Go champions as archived in Smart Game Format (SGF), which offers a text description of move sequences. The trained model further generates valid but previously unseen strategies for Go. Because GPT-2 preserves punctuation and spacing, the raw output of the text generator provides inputs to game visualization and creative patterns, such as the Sabaki project's game engine using auto-replays. Results demonstrate…
UC Berkeley（加州大学伯克利分校）；Facebook AI Research（脸书人工智能研究院）；Google Brain（谷歌大脑）
We introduce a framework that abstracts Reinforcement Learning (RL) as a sequence modeling problem. This allows us to draw upon the simplicity and scalability of the Transformer architecture, and associated advances in language modeling such as GPT-x and BERT. In particular, we present Decision Transformer, an architecture that casts the problem of RL as conditional sequence modeling. Unlike prior approaches to RL that fit value functions or compute policy gradients, Decision Transformer simply outputs the optimal actions by leveraging a causally masked Transformer. By conditioning an autoregr…
Ant Financial Services Group（蚂蚁金融服务集团）；Hasso Plattner Institute（哈索·普拉特纳研究所）；University of Potsdam（波茨坦大学）
Human language understanding operates at multiple levels of granularity (e.g., words, phrases, and sentences) with increasing levels of abstraction that can be hierarchically combined. However, existing deep models with stacked layers do not explicitly model any sort of hierarchical process. This paper proposes a recursive Transformer model based on differentiable CKY style binary trees to emulate the composition process. We extend the bidirectional language model pre-training objective to this architecture, attempting to predict each word given its left and right abstraction nodes. To scale u…
CSAIL, MIT（麻省理工学院计算机科学与人工智能实验室）；University of Texas at Austin（得克萨斯大学奥斯汀分校）；Amazon（亚马逊）
The architecture of Transformer is based entirely on self-attention, and has been shown to outperform models that employ recurrence on sequence transduction tasks such as machine translation. The superior performance of Transformer has been attributed to propagating signals over shorter distances, between positions in the input and the output, compared to the recurrent architectures. We establish connections between the dynamics in Transformer and recurrent networks to argue that several factors including gradient flow along an ensemble of multiple weakly dependent paths play a paramount role …
In modern NLP applications, word embeddings are a crucial backbone that can be readily shared across a number of tasks. However as the text distributions change and word semantics evolve over time, the downstream applications using the embeddings can suffer if the word representations do not conform to the data drift. Thus, maintaining word embeddings to be consistent with the underlying data distribution is a key problem. In this work, we tackle this problem and propose TransDrift, a transformer-based prediction model for word embeddings. Leveraging the flexibility of transformer, our model a…
Georgia Institute of Technology（佐治亚理工学院）；Microsoft（微软）
Transformer models have achieved superior performance in various natural language processing tasks. However, the quadratic computational cost of the attention mechanism limits its practicality for long sequences. There are existing attention variants that improve the computational efficiency, but they have limited ability to effectively compute global information. In parallel to Transformer models, state space models (SSMs) are tailored for long sequences, but they are not flexible enough to capture complicated local information. We propose SPADE, short for $\underline{\textbf{S}}$tate s$\unde…
Modeling unit and model architecture are two key factors of Recurrent Neural Network Transducer (RNN-T) in end-to-end speech recognition. To improve the performance of RNN-T for Mandarin speech recognition task, a novel transformer transducer with the combination architecture of self-attention transformer and RNN is proposed. And then the choice of different modeling units for transformer transducer is explored. In addition, we present a new mix-bandwidth training method to obtain a general model that is able to accurately recognize Mandarin speech with different sampling rates simultaneously.…
Efficient Transformers have been developed for long sequence modeling, due to their subquadratic memory and time complexity. Sparse Transformer is a popular approach to improving the efficiency of Transformers by restricting self-attention to locations specified by the predefined sparse patterns. However, leveraging sparsity may sacrifice expressiveness compared to full-attention, when important token correlations are multiple hops away. To combine advantages of both the efficiency of sparse transformer and the expressiveness of full-attention Transformer, we propose \textit{Diffuser}, a new s…
Soft robotic manipulators provide numerous advantages over conventional rigid manipulators in fragile environments such as the marine environment. However, developing analytic inverse models necessary for shape, motion, and force control of such robots remains a challenging problem. As an alternative to analytic models, numerical models can be learned using powerful machine learned methods. In this paper, the Kinematics Transformer is proposed for developing accurate and precise inverse kinematic models of soft robotic limbs. The proposed method re-casts the inverse kinematics problem as a seq…
Graduate University of Advanced Technology（高等技术研究生院）；Shahid Bahonar University of Kerman（克尔曼沙希德·巴霍纳尔大学）；University of Isfahan（伊斯法罕大学）
In vision-based action recognition, spatio-temporal features from different modalities are used for recognizing activities. Temporal modeling is a long challenge of action recognition. However, there are limited methods such as pre-computed motion features, three-dimensional (3D) filters, and recurrent neural networks (RNN) for modeling motion information in deep-based approaches. Recently, transformers success in modeling long-range dependencies in natural language processing (NLP) tasks has gotten great attention from other domains; including speech, image, and video, to rely entirely on sel…
The longitudinal modeling of neuroanatomical changes related to Alzheimer's disease (AD) is crucial for studying the progression of the disease. To this end, we introduce TransforMesh, a spatio-temporal network based on transformers that models longitudinal shape changes on 3D anatomical meshes. While transformer and mesh networks have recently shown impressive performances in natural language processing and computer vision, their application to medical image analysis has been very limited. To the best of our knowledge, this is the first work that combines transformer and mesh networks. Our re…
Large Transformer models have been central to recent advances in natural language processing. The training and inference costs of these models, however, have grown rapidly and become prohibitively expensive. Here we aim to reduce the costs of Transformers by searching for a more efficient variant. Compared to previous approaches, our search is performed at a lower level, over the primitives that define a Transformer TensorFlow program. We identify an architecture, named Primer, that has a smaller training cost than the original Transformer and other variants for auto-regressive language modeli…
Transformers are widely used in natural language processing due to their ability to model longer-term dependencies in text. Although these models achieve state-of-the-art performance for many language related tasks, their applicability outside of the natural language processing field has been minimal. In this work, we propose the use of transformer models for the prediction of dynamical systems representative of physical phenomena. The use of Koopman based embeddings provide a unique and powerful method for projecting any dynamical system into a vector representation which can then be predicte…
Transformers have demonstrated impressive strength in long-term series forecasting. Existing prediction research mostly focused on mapping past short sub-series (lookback window) to future series (forecast window). The longer training dataset time series will be discarded, once training is completed. Models can merely rely on lookback window information for inference, which impedes models from analyzing time series from a global perspective. And these windows used by Transformers are quite narrow because they must model each time-step therein. Under this point-wise processing style, broadening…
Transformer models have been developed in molecular science with excellent performance in applications including quantitative structure-activity relationship (QSAR) and virtual screening (VS). Compared with other types of models, however, they are large, which results in a high hardware requirement to abridge time for both training and inference processes. In this work, cross-layer parameter sharing (CLPS), and knowledge distillation (KD) are used to reduce the sizes of transformers in molecular science. Both methods not only have competitive QSAR predictive performance as compared to the orig…
KAIST（韩国科学技术院）；University of Michigan（密歇根大学）；LG AI Research（LG人工智能研究院）；UC Berkeley（加州大学伯克利分校）；Google Research（谷歌研究院）
Preference-based reinforcement learning (RL) provides a framework to train agents using human preferences between two behaviors. However, preference-based RL has been challenging to scale since it requires a large amount of human feedback to learn a reward function aligned with human intent. In this paper, we present Preference Transformer, a neural architecture that models human preferences using transformers. Unlike prior approaches assuming human judgment is based on the Markovian rewards which contribute to the decision equally, we introduce a new preference model based on the weighted sum…
Modeling neural population dynamics underlying noisy single-trial spiking activities is essential for relating neural observation and behavior. A recent non-recurrent method - Neural Data Transformers (NDT) - has shown great success in capturing neural dynamics with low inference latency without an explicit dynamical model. However, NDT focuses on modeling the temporal evolution of the population activity while neglecting the rich covariation between individual neurons. In this paper we introduce SpatioTemporal Neural Data Transformer (STNDT), an NDT-based architecture that explicitly models r…
