PyTorchKR Papers

최신 논문

총 9,339편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2208.01252

A Novel Transformer Network with Shifted Window Cross-Attention for Spatiotemporal Weather Forecasting

Earth Observatory is a growing research area that can capitalize on the powers of AI for short time forecasting, a Now-casting scenario. In this work, we tackle the challenge of weather forecasting using a video transformer network. Vision transformer architectures have been explored in various applications, with major constraints being the computational complexity of Attention and the data hungry training. To address these issues, we propose the use of Video Swin-Transformer, coupled with a dedicated augmentation scheme. Moreover, we employ gradual spatial reduction on the encoder side and cross-attention on the decoder. The proposed approach is tested on the Weather4Cast2021 weather forecasting challenge data, which requires the prediction of 8 hours ahead future frames (4 per hour) from an hourly weather product sequence. The dataset was normalized to 0-1 to facilitate using the evaluation metrics across different datasets. The model results in an MSE score of 0.4750 when provided with training data, and 0.4420 during transfer learning without using training data, respectively.

arXiv:2208.01373

DAPDAG: Domain Adaptation via Perturbed DAG Reconstruction

Leveraging labelled data from multiple domains to enable prediction in another domain without labels is a significant, yet challenging problem. To address this problem, we introduce the framework DAPDAG (\textbf{D}omain \textbf{A}daptation via \textbf{P}erturbed \textbf{DAG} Reconstruction) and propose to learn an auto-encoder that undertakes inference on population statistics given features and reconstructing a directed acyclic graph (DAG) as an auxiliary task. The underlying DAG structure is assumed invariant among observed variables whose conditional distributions are allowed to vary across domains led by a latent environmental variable $E$. The encoder is designed to serve as an inference device on $E$ while the decoder reconstructs each observed variable conditioned on its graphical parents in the DAG and the inferred $E$. We train the encoder and decoder jointly in an end-to-end manner and conduct experiments on synthetic and real datasets with mixed variables. Empirical results demonstrate that reconstructing the DAG benefits the approximate inference. Furthermore, our approach can achieve competitive performance against other benchmarks in prediction tasks, with better adaptation ability, especially in the target domain significantly different from the source domains.

arXiv:2208.00748공식 코드 · Papers with Code

Efficient Long-Text Understanding with Short-Text Models

Transformer-based pretrained language models (LMs) are ubiquitous across natural language understanding, but cannot be applied to long sequences such as stories, scientific articles and long documents, due to their quadratic complexity. While a myriad of efficient transformer variants have been proposed, they are typically based on custom implementations that require expensive pretraining from scratch. In this work, we propose SLED: SLiding-Encoder and Decoder, a simple approach for processing long sequences that re-uses and leverages battle-tested short-text pretrained LMs. Specifically, we partition the input into overlapping chunks, encode each with a short-text LM encoder and use the pretrained decoder to fuse information across chunks (fusion-in-decoder). We illustrate through controlled experiments that SLED offers a viable strategy for long text understanding and evaluate our approach on SCROLLS, a benchmark with seven datasets across a wide range of language understanding tasks. We find that SLED is competitive with specialized models that are up to 50x larger and require a dedicated and expensive pretraining step.

arXiv:2208.00293

Global Attention-based Encoder-Decoder LSTM Model for Temperature Prediction of Permanent Magnet Synchronous Motors

Temperature monitoring is critical for electrical motors to determine if device protection measures should be executed. However, the complexity of the internal structure of Permanent Magnet Synchronous Motors (PMSM) makes the direct temperature measurement of the internal components difficult. This work pragmatically develops three deep learning models to estimate the PMSMs' internal temperature based on readily measurable external quantities. The proposed supervised learning models exploit Long Short-Term Memory (LSTM) modules, bidirectional LSTM, and attention mechanism to form encoder-decoder structures to predict simultaneously the temperatures of the stator winding, tooth, yoke, and permanent magnet. Experiments were conducted in an exhaustive manner on a benchmark dataset to verify the proposed models' performances. The comparative analysis shows that the proposed global attention-based encoder-decoder (EnDec) model provides a competitive overall performance of 1.72 Mean Squared Error (MSE) and 5.34 Mean Absolute Error (MAE).

arXiv:2208.00246

Geometric deep learning for computational mechanics Part II: Graph embedding for interpretable multiscale plasticity

The history-dependent behaviors of classical plasticity models are often driven by internal variables evolved according to phenomenological laws. The difficulty to interpret how these internal variables represent a history of deformation, the lack of direct measurement of these internal variables for calibration and validation, and the weak physical underpinning of those phenomenological laws have long been criticized as barriers to creating realistic models. In this work, geometric machine learning on graph data (e.g. finite element solutions) is used as a means to establish a connection between nonlinear dimensional reduction techniques and plasticity models. Geometric learning-based encoding on graphs allows the embedding of rich time-history data onto a low-dimensional Euclidean space such that the evolution of plastic deformation can be predicted in the embedded feature space. A corresponding decoder can then convert these low-dimensional internal variables back into a weighted graph such that the dominating topological features of plastic deformation can be observed and analyzed.

arXiv:2208.00657

SiamixFormer: a fully-transformer Siamese network with temporal Fusion for accurate building detection and change detection in bi-temporal remote sensing images

Building detection and change detection using remote sensing images can help urban and rescue planning. Moreover, they can be used for building damage assessment after natural disasters. Currently, most of the existing models for building detection use only one image (pre-disaster image) to detect buildings. This is based on the idea that post-disaster images reduce the model's performance because of presence of destroyed buildings. In this paper, we propose a siamese model, called SiamixFormer, which uses pre- and post-disaster images as input. Our model has two encoders and has a hierarchical transformer architecture. The output of each stage in both encoders is given to a temporal transformer for feature fusion in a way that query is generated from pre-disaster images and (key, value) is generated from post-disaster images. To this end, temporal features are also considered in feature fusion. Another advantage of using temporal transformers in feature fusion is that they can better maintain large receptive fields generated by transformer encoders compared with CNNs. Finally, the output of the temporal transformer is given to a simple MLP decoder at each stage. The SiamixFormer model is evaluated on xBD, and WHU datasets, for building detection and on LEVIR-CD and CDD datasets for change detection and could outperform the state-of-the-art.

arXiv:2208.00050

Generating Multiple 4D Expression Transitions by Learning Face Landmark Trajectories

In this work, we address the problem of 4D facial expressions generation. This is usually addressed by animating a neutral 3D face to reach an expression peak, and then get back to the neutral state. In the real world though, people show more complex expressions, and switch from one expression to another. We thus propose a new model that generates transitions between different expressions, and synthesizes long and composed 4D expressions. This involves three sub-problems: (i) modeling the temporal dynamics of expressions, (ii) learning transitions between them, and (iii) deforming a generic mesh. We propose to encode the temporal evolution of expressions using the motion of a set of 3D landmarks, that we learn to generate by training a manifold-valued GAN (Motion3DGAN). To allow the generation of composed expressions, this model accepts two labels encoding the starting and the ending expressions. The final sequence of meshes is generated by a Sparse2Dense mesh Decoder (S2D-Dec) that maps the landmark displacements to a dense, per-vertex displacement of a known mesh topology. By explicitly working with motion trajectories, the model is totally independent from the identity. Extensive experiments on five public datasets show that our proposed approach brings significant improvements with respect to previous solutions, while retaining good generalization to unseen data.

arXiv:2207.14467공식 코드 · Papers with Code

GTrans: Grouping and Fusing Transformer Layers for Neural Machine Translation

Transformer structure, stacked by a sequence of encoder and decoder network layers, achieves significant development in neural machine translation. However, vanilla Transformer mainly exploits the top-layer representation, assuming the lower layers provide trivial or redundant information and thus ignoring the bottom-layer feature that is potentially valuable. In this work, we propose the Group-Transformer model (GTrans) that flexibly divides multi-layer representations of both encoder and decoder into different groups and then fuses these group features to generate target words. To corroborate the effectiveness of the proposed method, extensive experiments and analytic experiments are conducted on three bilingual translation benchmarks and two multilingual translation tasks, including the IWLST-14, IWLST-17, LDC, WMT-14 and OPUS-100 benchmark. Experimental and analytical results demonstrate that our model outperforms its Transformer counterparts by a consistent gain. Furthermore, it can be successfully scaled up to 60 encoder layers and 36 decoder layers.

arXiv:2207.14812공식 코드 · Papers with Code

GLEAN: Generative Latent Bank for Image Super-Resolution and Beyond

We show that pre-trained Generative Adversarial Networks (GANs) such as StyleGAN and BigGAN can be used as a latent bank to improve the performance of image super-resolution. While most existing perceptual-oriented approaches attempt to generate realistic outputs through learning with adversarial loss, our method, Generative LatEnt bANk (GLEAN), goes beyond existing practices by directly leveraging rich and diverse priors encapsulated in a pre-trained GAN. But unlike prevalent GAN inversion methods that require expensive image-specific optimization at runtime, our approach only needs a single forward pass for restoration. GLEAN can be easily incorporated in a simple encoder-bank-decoder architecture with multi-resolution skip connections. Employing priors from different generative models allows GLEAN to be applied to diverse categories (\eg~human faces, cats, buildings, and cars). We further present a lightweight version of GLEAN, named LightGLEAN, which retains only the critical components in GLEAN. Notably, LightGLEAN consists of only 21% of parameters and 35% of FLOPs while achieving comparable image quality. We extend our method to different tasks including image colorization and blind image restoration, and extensive experiments show that our proposed models perform favorably in comparison to existing methods. Codes and models are available at https://github.com/open-mmlab/mmediting.

arXiv:2207.14789

Encoder-Decoder Architecture for 3D Seismic Inversion

Inverting seismic data to build 3D geological structures is a challenging task due to the overwhelming amount of acquired seismic data, and the very-high computational load due to iterative numerical solutions of the wave equation, as required by industry-standard tools such as Full Waveform Inversion (FWI). For example, in an area with surface dimensions of 4.5km $\times$ 4.5km, hundreds of seismic shot-gather cubes are required for 3D model reconstruction, leading to Terabytes of recorded data. This paper presents a deep learning solution for the reconstruction of realistic 3D models in the presence of field noise recorded in seismic surveys. We implement and analyze a convolutional encoder-decoder architecture that efficiently processes the entire collection of hundreds of seismic shot-gather cubes. The proposed solution demonstrates that realistic 3D models can be reconstructed with a structural similarity index measure (SSIM) of 0.8554 (out of 1.0) in the presence of field noise at 10dB signal-to-noise ratio.

arXiv:2207.14742공식 코드 · Papers with Code

Graph Neural Networks for Channel Decoding

In this work, we propose a fully differentiable graph neural network (GNN)-based architecture for channel decoding and showcase a competitive decoding performance for various coding schemes, such as low-density parity-check (LDPC) and BCH codes. The idea is to let a neural network (NN) learn a generalized message passing algorithm over a given graph that represents the forward error correction (FEC) code structure by replacing node and edge message updates with trainable functions. Contrary to many other deep learning-based decoding approaches, the proposed solution enjoys scalability to arbitrary block lengths and the training is not limited by the curse of dimensionality. We benchmark our proposed decoder against state-of-the-art in conventional channel decoding as well as against recent deep learning-based results. For the (63,45) BCH code, our solution outperforms weighted belief propagation (BP) decoding by approximately 0.4 dB with significantly less decoding iterations and even for 5G NR LDPC codes, we observe a competitive performance when compared to conventional BP decoding. For the BCH codes, the resulting GNN decoder can be fully parametrized with only 9640 weights.

arXiv:2207.14678

AlphaVC: High-Performance and Efficient Learned Video Compression

Recently, learned video compression has drawn lots of attention and show a rapid development trend with promising results. However, the previous works still suffer from some criticial issues and have a performance gap with traditional compression standards in terms of widely used PSNR metric. In this paper, we propose several techniques to effectively improve the performance. First, to address the problem of accumulative error, we introduce a conditional-I-frame as the first frame in the GoP, which stabilizes the reconstructed quality and saves the bit-rate. Second, to efficiently improve the accuracy of inter prediction without increasing the complexity of decoder, we propose a pixel-to-feature motion prediction method at encoder side that helps us to obtain high-quality motion information. Third, we propose a probability-based entropy skipping method, which not only brings performance gain, but also greatly reduces the runtime of entropy coding. With these powerful techniques, this paper proposes AlphaVC, a high-performance and efficient learned video compression scheme. To the best of our knowledge, AlphaVC is the first E2E AI codec that exceeds the latest compression standard VVC on all common test datasets for both PSNR (-28.2% BD-rate saving) and MSSSIM (-52.2% BD-rate saving), and has very fast encoding (0.001x VVC) and decoding (1.69x VVC) speeds.

arXiv:2207.14508

Transfer Learning for Segmentation Problems: Choose the Right Encoder and Skip the Decoder

It is common practice to reuse models initially trained on different data to increase downstream task performance. Especially in the computer vision domain, ImageNet-pretrained weights have been successfully used for various tasks. In this work, we investigate the impact of transfer learning for segmentation problems, being pixel-wise classification problems that can be tackled with encoder-decoder architectures. We find that transfer learning the decoder does not help downstream segmentation tasks, while transfer learning the encoder is truly beneficial. We demonstrate that pretrained weights for a decoder may yield faster convergence, but they do not improve the overall model performance as one can obtain equivalent results with randomly initialized decoders. However, we show that it is more effective to reuse encoder weights trained on a segmentation or reconstruction task than reusing encoder weights trained on classification tasks. This finding implicates that using ImageNet-pretrained encoders for downstream segmentation problems is suboptimal. We also propose a contrastive self-supervised approach with multiple self-reconstruction tasks, which provides encoders that are suitable for transfer learning in segmentation problems in the absence of segmentation labels.

arXiv:2207.14498공식 코드 · Papers with Code

Reference-Guided Texture and Structure Inference for Image Inpainting

Existing learning-based image inpainting methods are still in challenge when facing complex semantic environments and diverse hole patterns. The prior information learned from the large scale training data is still insufficient for these situations. Reference images captured covering the same scenes share similar texture and structure priors with the corrupted images, which offers new prospects for the image inpainting tasks. Inspired by this, we first build a benchmark dataset containing 10K pairs of input and reference images for reference-guided inpainting. Then we adopt an encoder-decoder structure to separately infer the texture and structure features of the input image considering their pattern discrepancy of texture and structure during inpainting. A feature alignment module is further designed to refine these features of the input image with the guidance of a reference image. Both quantitative and qualitative evaluations demonstrate the superiority of our method over the state-of-the-art methods in terms of completing complex holes.

arXiv:2207.14477

FCSN: Global Context Aware Segmentation by Learning the Fourier Coefficients of Objects in Medical Images

The encoder-decoder model is a commonly used Deep Neural Network (DNN) model for medical image segmentation. Conventional encoder-decoder models make pixel-wise predictions focusing heavily on local patterns around the pixel. This makes it challenging to give segmentation that preserves the object's shape and topology, which often requires an understanding of the global context of the object. In this work, we propose a Fourier Coefficient Segmentation Network~(FCSN) -- a novel DNN-based model that segments an object by learning the complex Fourier coefficients of the object's masks. The Fourier coefficients are calculated by integrating over the whole contour. Therefore, for our model to make a precise estimation of the coefficients, the model is motivated to incorporate the global context of the object, leading to a more accurate segmentation of the object's shape. This global context awareness also makes our model robust to unseen local perturbations during inference, such as additive noise or motion blur that are prevalent in medical images. When FCSN is compared with other state-of-the-art models (UNet+, DeepLabV3+, UNETR) on 3 medical image segmentation tasks (ISIC\_2018, RIM\_CUP, RIM\_DISC), FCSN attains significantly lower Hausdorff scores of 19.14 (6\%), 17.42 (6\%), and 9.16 (14\%) on the 3 tasks, respectively. Moreover, FCSN is lightweight by discarding the decoder module, which incurs significant computational overhead. FCSN only requires 22.2M parameters, 82M and 10M fewer parameters than UNETR and DeepLabV3+. FCSN attains inference and training speeds of 1.6ms/img and 6.3ms/img, that is 8$\times$ and 3$\times$ faster than UNet and UNETR.

arXiv:2207.14166

RHA-Net: An Encoder-Decoder Network with Residual Blocks and Hybrid Attention Mechanisms for Pavement Crack Segmentation

The acquisition and evaluation of pavement surface data play an essential role in pavement condition evaluation. In this paper, an efficient and effective end-to-end network for automatic pavement crack segmentation, called RHA-Net, is proposed to improve the pavement crack segmentation accuracy. The RHA-Net is built by integrating residual blocks (ResBlocks) and hybrid attention blocks into the encoder-decoder architecture. The ResBlocks are used to improve the ability of RHA-Net to extract high-level abstract features. The hybrid attention blocks are designed to fuse both low-level features and high-level features to help the model focus on correct channels and areas of cracks, thereby improving the feature presentation ability of RHA-Net. An image data set containing 789 pavement crack images collected by a self-designed mobile robot is constructed and used for training and evaluating the proposed model. Compared with other state-of-the-art networks, the proposed model achieves better performance and the functionalities of adding residual blocks and hybrid attention mechanisms are validated in a comprehensive ablation study. Additionally, a light-weighted version of the model generated by introducing depthwise separable convolution achieves better a performance and a much faster processing speed with 1/30 of the number of U-Net parameters. The developed system can segment pavement crack in real-time on an embedded device Jetson TX2 (25 FPS). The video taken in real-time experiments is released at https://youtu.be/3XIogk0fiG4.

arXiv:2207.14134

A Transformer-based Generative Adversarial Network for Brain Tumor Segmentation

Brain tumor segmentation remains a challenge in medical image segmentation tasks. With the application of transformer in various computer vision tasks, transformer blocks show the capability of learning long-distance dependency in global space, which is complementary with CNNs. In this paper, we proposed a novel transformer-based generative adversarial network to automatically segment brain tumors with multi-modalities MRI. Our architecture consists of a generator and a discriminator, which are trained in min-max game progress. The generator is based on a typical "U-shaped" encoder-decoder architecture, whose bottom layer is composed of transformer blocks with resnet. Besides, the generator is trained with deep supervision technology. The discriminator we designed is a CNN-based network with multi-scale $L_{1}$ loss, which is proved to be effective for medical semantic image segmentation. To validate the effectiveness of our method, we conducted experiments on BRATS2015 dataset, achieving comparable or better performance than previous state-of-the-art methods.

arXiv:2207.13861

DnSwin: Toward Real-World Denoising via Continuous Wavelet Sliding-Transformer

Real-world image denoising is a practical image restoration problem that aims to obtain clean images from in-the-wild noisy inputs. Recently, the Vision Transformer (ViT) has exhibited a strong ability to capture long-range dependencies, and many researchers have attempted to apply the ViT to image denoising tasks. However, a real-world image is an isolated frame that makes the ViT build long-range dependencies based on the internal patches, which divides images into patches, disarranges noise patterns and damages gradient continuity. In this article, we propose to resolve this issue by using a continuous Wavelet Sliding-Transformer that builds frequency correspondences under real-world scenes, called DnSwin. Specifically, we first extract the bottom features from noisy input images by using a convolutional neural network (CNN) encoder. The key to DnSwin is to extract high-frequency and low-frequency information from the observed features and build frequency dependencies. To this end, we propose a Wavelet Sliding-Window Transformer (WSWT) that utilizes the discrete wavelet transform (DWT), self-attention and the inverse DWT (IDWT) to extract deep features. Finally, we reconstruct the deep features into denoised images using a CNN decoder. Both quantitative and qualitative evaluations conducted on real-world denoising benchmarks demonstrate that the proposed DnSwin performs favorably against the state-of-the-art methods.

arXiv:2207.13820공식 코드 · Papers with Code

Cross-Attention of Disentangled Modalities for 3D Human Mesh Recovery with Transformers

Transformer encoder architectures have recently achieved state-of-the-art results on monocular 3D human mesh reconstruction, but they require a substantial number of parameters and expensive computations. Due to the large memory overhead and slow inference speed, it is difficult to deploy such models for practical use. In this paper, we propose a novel transformer encoder-decoder architecture for 3D human mesh reconstruction from a single image, called FastMETRO. We identify the performance bottleneck in the encoder-based transformers is caused by the token design which introduces high complexity interactions among input tokens. We disentangle the interactions via an encoder-decoder architecture, which allows our model to demand much fewer parameters and shorter inference time. In addition, we impose the prior knowledge of human body's morphological relationship via attention masking and mesh upsampling operations, which leads to faster convergence with higher accuracy. Our FastMETRO improves the Pareto-front of accuracy and efficiency, and clearly outperforms image-based methods on Human3.6M and 3DPW. Furthermore, we validate its generalizability on FreiHAND.

arXiv:2207.13532공식 코드 · Papers with Code

Contrastive Masked Autoencoders are Stronger Vision Learners

Masked image modeling (MIM) has achieved promising results on various vision tasks. However, the limited discriminability of learned representation manifests there is still plenty to go for making a stronger vision learner. Towards this goal, we propose Contrastive Masked Autoencoders (CMAE), a new self-supervised pre-training method for learning more comprehensive and capable vision representations. By elaboratively unifying contrastive learning (CL) and masked image model (MIM) through novel designs, CMAE leverages their respective advantages and learns representations with both strong instance discriminability and local perceptibility. Specifically, CMAE consists of two branches where the online branch is an asymmetric encoder-decoder and the momentum branch is a momentum updated encoder. During training, the online encoder reconstructs original images from latent representations of masked images to learn holistic features. The momentum encoder, fed with the full images, enhances the feature discriminability via contrastive learning with its online counterpart. To make CL compatible with MIM, CMAE introduces two new components, i.e. pixel shifting for generating plausible positive views and feature decoder for complementing features of contrastive pairs. Thanks to these novel designs, CMAE effectively improves the representation quality and transfer performance over its MIM counterpart. CMAE achieves the state-of-the-art performance on highly competitive benchmarks of image classification, semantic segmentation and object detection. Notably, CMAE-Base achieves $85.3\%$ top-1 accuracy on ImageNet and $52.5\%$ mIoU on ADE20k, surpassing previous best results by $0.7\%$ and $1.8\%$ respectively. The source code is publicly accessible at \url{https://github.com/ZhichengHuang/CMAE}.