Transformers have become one of the most important architectural innovations in deep learning and have enabled many breakthroughs over the past few years. Here we propose a simple network architecture, gMLP, based on MLPs with gating, and show that it can perform as well as Transformers in key language and vision applications. Our comparisons show that self-attention is not critical for Vision Transformers, as gMLP can achieve the same accuracy. For BERT, our model achieves parity with Transformers on pretraining perplexity and is better on some downstream NLP tasks. On finetuning tasks where gMLP performs worse, making the gMLP model substantially larger can close the gap with Transformers. In general, our experiments show that gMLP can scale as well as Transformers over increased data and compute.
In the first week of May, 2021, researchers from four different institutions: Google, Tsinghua University, Oxford University and Facebook, shared their latest work [16, 7, 12, 17] on arXiv.org almost at the same time, each proposing new learning architectures, consisting mainly of linear layers, claiming them to be comparable, or even superior to convolutional-based models. This sparked immediate discussion and debate in both academic and industrial communities as to whether MLPs are sufficient, many thinking that learning architectures are returning to MLPs. Is this true? In this perspective,…
We initiate the first empirical study on the use of MLP architectures for vision-and-language (VL) fusion. Through extensive experiments on 5 VL tasks and 5 robust VQA benchmarks, we find that: (i) Without pre-training, using MLPs for multimodal fusion has a noticeable performance gap compared to transformers; (ii) However, VL pre-training can help close the performance gap; (iii) Instead of heavy multi-head attention, adding tiny one-head attention to MLPs is sufficient to achieve comparable performance to transformers. Moreover, we also find that the performance gap between MLPs and transfor…
DWAFM: Dynamic Weighted Graph Structure Embedding Integrated with Attention and Frequency-Domain MLPs for Traffic Forecasting
School of Mathematics and Statistics, Nanjing University of Information Science and Technology（南京信息工程大学数学与统计学院）；Hubei Key Laboratory of Applied Mathematics, Hubei University（湖北应用数学重点实验室）；Key Laboratory of System Control and Information Processing, Ministry of Education（教育部系统控制与信息处理重点实验室）；Nanjing Institute of Technology（南京理工大学）；Jiangsu Province Engineering Research Center of IntelliSense Technology and System（江苏省智能感知技术与系统工程研究中心）
MLPs Learn In-Context on Regression and Classification Tasks
School of Engineering and Applied Sciences（工程与应用科学学院）；Center for Brain Sciences（脑科学中心）；Kempner Institute for the Study of Artificial and Natural Intelligence（肯普纳人工智能与自然智能研究所）；Harvard University（哈佛大学）
Convolutional Neural Networks (CNNs) are the go-to model for computer vision. Recently, attention-based networks, such as the Vision Transformer, have also become popular. In this paper we show that while convolutions and attention are both sufficient for good performance, neither of them are necessary. We present MLP-Mixer, an architecture based exclusively on multi-layer perceptrons (MLPs). MLP-Mixer contains two types of layers: one with MLPs applied independently to image patches (i.e. "mixing" the per-location features), and one with MLPs applied across patches (i.e. "mixing" spatial info…
University of Queensland（昆士兰大学）；International Computer Science Institute（国际计算机科学研究所）
An interesting approach to analyzing neural networks that has received renewed attention is to examine the equivalent kernel of the neural network. This is based on the fact that a fully connected feedforward network with one hidden layer, a certain weight distribution, an activation function, and an infinite number of neurons can be viewed as a mapping into a Hilbert space. We derive the equivalent kernels of MLPs with ReLU or Leaky ReLU activations for all rotationally-invariant weight distributions, generalizing a previous result that required Gaussian weight distributions. Additionally, th…
Transformer-based models consist of interleaved feed-forward blocks - that capture content meaning, and relatively more expensive self-attention blocks - that capture context meaning. In this paper, we explored trade-offs and ordering of the blocks to improve upon the current Transformer architecture and proposed PAR Transformer. It needs 35% lower compute time than Transformer-XL achieved by replacing ~63% of the self-attention blocks with feed-forward blocks, and retains the perplexity on WikiText-103 language modelling benchmark. We further validated our results on text8 and enwiki8 dataset…
Nanjing University of Information Science and Technology（南京信息工程大学）；Nanjing Normal University（南京师范大学）
Sequential visual task usually requires to pay attention to its current interested object conditional on its previous observations. Different from popular soft attention mechanism, we propose a new attention framework by introducing a novel conditional global feature which represents the weak feature descriptor of the current focused object. Specifically, for a standard CNN (Convolutional Neural Network) pipeline, the convolutional layers with different receptive fields are used to produce the attention maps by measuring how the convolutional features align to the conditional global feature. T…
We have successfully implemented the "Learn to Pay Attention" model of attention mechanism in convolutional neural networks, and have replicated the results of the original paper in the categories of image classification and fine-grained recognition.
University of New Mexico（新墨西哥大学）；Verint Intelligent Self-Service（Verint智能自助服务）
In (Yang et al. 2016), a hierarchical attention network (HAN) is created for document classification. The attention layer can be used to visualize text influential in classifying the document, thereby explaining the model's prediction. We successfully applied HAN to a sequential analysis task in the form of real-time monitoring of turn taking in conversations. However, we discovered instances where the attention weights were uniform at the stopping point (indicating all turns were equivalently influential to the classifier), preventing meaningful visualization for real-time human review or cla…
Till now, attention-based models have been used with great success in the keyword spotting problem domain. However, in light of recent advances in deep learning, the question arises whether self-attention is truly irreplaceable for recognizing speech keywords. We thus explore the usage of gated MLPs --previously shown to be alternatives to transformers in vision tasks-- for the keyword spotting task. We provide a family of highly efficient MLP-based models for keyword spotting, with less than 0.5 million parameters. We show that our approach achieves competitive performance on Google Speech Co…
Canon Medical Research Europe（佳能医疗欧洲研究中心）；University of St Andrews（圣安德鲁斯大学）；Institute of Neuroscience & Psychology, University of Glasgow（格拉斯哥大学神经科学与心理学研究所）；University of Edinburgh（爱丁堡大学）
Training medical image analysis models requires large amounts of expertly annotated data which is time-consuming and expensive to obtain. Images are often accompanied by free-text radiology reports which are a rich source of information. In this paper, we tackle the automated extraction of structured labels from head CT reports for imaging of suspected stroke patients, using deep learning. Firstly, we propose a set of 31 labels which correspond to radiographic findings (e.g. hyperdensity) and clinical impressions (e.g. haemorrhage) related to neurological abnormalities. Secondly, inspired by p…
Neural processes (NPs) aim to stochastically complete unseen data points based on a given context dataset. NPs essentially leverage a given dataset as a context representation to derive a suitable identifier for a novel task. To improve the prediction accuracy, many variants of NPs have investigated context embedding approaches that generally design novel network architectures and aggregation functions satisfying permutation invariant. In this work, we propose a stochastic attention mechanism for NPs to capture appropriate context information. From the perspective of information theory, we dem…
Element AI（元素人工智能公司）；Computer Vision Center（计算机视觉中心）；Visual Tagging Services（视觉标记服务公司）；Universitat Autonoma de Barcelona（巴塞罗那自治大学）
Fine-grained image recognition is central to many multimedia tasks such as search, retrieval and captioning. Unfortunately, these tasks are still challenging since the appearance of samples of the same class can be more different than those from different classes. Attention has been typically implemented in neural networks by selecting the most informative regions of the image that improve classification. In contrast, in this paper, attention is not applied at the image level but to the convolutional feature activations. In essence, with our approach, the neural model learns to attend to lower…
Looped ReLU MLPs May Be All You Need as Practical Programmable Computers
The University of Hong Kong（香港大学）；University of Wisconsin-Madison（威斯康星大学麦迪逊分校）；Tsinghua University（清华大学）；The Simons Institute for the Theory of Computing（西蒙斯理论计算研究所）；University of Pennsylvania（宾夕法尼亚大学）
Rethinking Nonlinearity: Trainable Gaussian Mixture Modules for Modern Neural Architectures
Guangdong Institute of Intelligence Science and Technology（广东智能科学与技术研究院）；Great Bay University（大湾大学）；University of Massachusetts at Amherst（马萨诸塞大学阿姆赫斯特分校）
Tencent（腾讯）；Shenzhen University（深圳大学）；The Chinese University of Hong Kong（香港中文大学）
With the explosive growth of transaction activities in online payment systems, effective and realtime regulation becomes a critical problem for payment service providers. Thanks to the rapid development of artificial intelligence (AI), AI-enable regulation emerges as a promising solution. One main challenge of the AI-enabled regulation is how to utilize multimedia information, i.e., multimodal signals, in Financial Technology (FinTech). Inspired by the attention mechanism in nature language processing, we propose a novel cross-modal and intra-modal attention network (CIAN) to investigate the r…
We use data on 124 batteries released by Stanford University to first try to solve the binary classification problem of determining if a battery is "good" or "bad" given only the first 5 cycles of data (i.e., will it last longer than a certain threshold of cycles), as well as the prediction problem of determining the exact number of cycles a battery will last given the first 100 cycles of data. We approach the problem from a purely data-driven standpoint, hoping to use deep learning to learn the patterns in the sequences of data that the Stanford team engineered by hand. For both problems, we …
Recurrent Neural Network (RNN) has been successfully applied in many sequence learning problems. Such as handwriting recognition, image description, natural language processing and video motion analysis. After years of development, researchers have improved the internal structure of the RNN and introduced many variants. Among others, Gated Recurrent Unit (GRU) is one of the most widely used RNN model. However, GRU lacks the capability of adaptively paying attention to certain regions or locations, so that it may cause information redundancy or loss during leaning. In this paper, we propose a R…
Exploring the Limitations of Kolmogorov-Arnold Networks in Classification: Insights to Software Training and Hardware Implementation
Nara Institute of Science and Technology（奈良科学技术研究所）；University of Information Technology - VNUHCM（胡志明市国家大学信息学院）；Le Quy Don technical University（黎贵敦工业大学）
We present a generalization of Transformers to any-order permutation invariant data (sets, graphs, and hypergraphs). We begin by observing that Transformers generalize DeepSets, or first-order (set-input) permutation invariant MLPs. Then, based on recently characterized higher-order invariant MLPs, we extend the concept of self-attention to higher orders and propose higher-order Transformers for order-$k$ data ($k=2$ for graphs and $k>2$ for hypergraphs). Unfortunately, higher-order Transformers turn out to have prohibitive complexity $\mathcal{O}(n^{2k})$ to the number of input nodes $n$. To …
Efficient Graph Knowledge Distillation from GNNs to Kolmogorov--Arnold Networks via Self-Attention Dynamic Sampling
School of Railway Intelligent Engineering, Dalian Jiaotong University, Dalian 116028, Liaoning, China（大连交通大学铁路智能工程学院）；College of Electronic Information and Automation, Civil Aviation University of China, Tianjin 300300, Tianjin, China（中国民航大学电子信息与自动化学院）
Eindhoven University of Technology（埃因霍温理工大学）；University of Twente（特温特大学）
A new line of research for feature selection based on neural networks has recently emerged. Despite its superiority to classical methods, it requires many training iterations to converge and detect informative features. The computational time becomes prohibitively long for datasets with a large number of samples or a very high dimensional feature space. In this paper, we present a new efficient unsupervised method for feature selection based on sparse autoencoders. In particular, we propose a new sparse training algorithm that optimizes a model's sparse topology during training to pay attentio…
Carnegie Mellon University（卡内基梅隆大学）；Instituto Superior Técnico（高等技术学院）；Instituto de Telecomunicações（电信研究所）；Unbabel
Context-aware machine translation models are designed to leverage contextual information, but often fail to do so. As a result, they inaccurately disambiguate pronouns and polysemous words that require context for resolution. In this paper, we ask several questions: What contexts do human translators use to resolve ambiguous words? Are models paying large amounts of attention to the same context? What if we explicitly train them to do so? To answer these questions, we introduce SCAT (Supporting Context for Ambiguous Translations), a new English-French dataset comprising supporting context word…
ASUS Intelligent Cloud Services（华硕智能云服务）；National Taiwan University（国立台湾大学）
The importance of parameter selection in supervised learning is well known. However, due to the many parameter combinations, an incomplete or an insufficient procedure is often applied. This situation may cause misleading or confusing conclusions. In this opinion paper, through an intriguing example we point out that the seriousness goes beyond what is generally recognized. In the topic of multi-label classification for medical code prediction, one influential paper conducted a proper parameter selection on a set, but when moving to a subset of frequently occurring labels, the authors used the…
Deep learning has emerged as a compelling solution to many NLP tasks with remarkable performances. However, due to their opacity, such models are hard to interpret and trust. Recent work on explaining deep models has introduced approaches to provide insights toward the model's behaviour and predictions, which are helpful for assessing the reliability of the model's predictions. However, such methods do not improve the model's reliability. In this paper, we aim to teach the model to make the right prediction for the right reason by providing explanation training and ensuring the alignment of th…
Pay Attention to the Atlas: Atlas-Guided Test-Time Adaptation Method for Robust 3D Medical Image Segmentation
School of Computation, Information and Technology, Technical University of Munich（慕尼黑工业大学计算、信息与技术学院）；Imperial College London（伦敦帝国学院）；HeartFlow, Inc.（哈特弗洛公司）；School of Medicine, Klinikum rechts der Isar, Technical University of Munich（慕尼黑工业大学右岸医院医学院）；Department of Engineering Science, University of Oxford（牛津大学工程科学系）；Department of Computer Science, University of Sheffield（谢菲尔德大学计算机科学系）
University of North Carolina at Charlotte（北卡罗来纳大学夏洛特分校）
Graph Convolutional Neural Networks (GCNs) have become effective machine learning algorithms for many downstream network mining tasks such as node classification, link prediction, and community detection. However, most GCN methods have been developed for homogenous networks and are limited to a single embedding for each node. Complex systems, often represented by heterogeneous, multiplex networks present a more difficult challenge for GCN models and require that such techniques capture the diverse contexts and assorted interactions that occur between nodes. In this work, we propose RAHMeN, a n…
Lipschitz constrained networks have gathered considerable attention in the deep learning community, with usages ranging from Wasserstein distance estimation to the training of certifiably robust classifiers. However they remain commonly considered as less accurate, and their properties in learning are still not fully understood. In this paper we clarify the matter: when it comes to classification 1-Lipschitz neural networks enjoy several advantages over their unconstrained counterpart. First, we show that these networks are as accurate as classical ones, and can fit arbitrarily difficult bound…
