PyTorchKR Papers

최신 논문

총 9,642편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2108.09605

Self-Supervised Delineation of Geological Structures using Orthogonal Latent Space Projection

We developed two machine learning frameworks that could assist in automated litho-stratigraphic interpretation of seismic volumes without any manual hand labeling from an experienced seismic interpreter. The first framework is an unsupervised hierarchical clustering model to divide seismic images from a volume into certain number of clusters determined by the algorithm. The clustering framework uses a combination of density and hierarchical techniques to determine the size and homogeneity of the clusters. The second framework consists of a self-supervised deep learning framework to label regions of geological interest in seismic images. It projects the latent-space of an encoder-decoder architecture unto two orthogonal subspaces, from which it learns to delineate regions of interest in the seismic images. To demonstrate an application of both frameworks, a seismic volume was clustered into various contiguous clusters, from which four clusters were selected based on distinct seismic patterns: horizons, faults, salt domes and chaotic structures. Images from the selected clusters are used to train the encoder-decoder network. The output of the encoder-decoder network is a probability map of the possibility an amplitude reflection event belongs to an interesting geological structure. The structures are delineated using the probability map. The delineated images are further used to post-train a segmentation model to extend our results to full-vertical sections. The results on vertical sections show that we can factorize a seismic volume into its corresponding structural components. Lastly, we showed that our deep learning framework could be modeled as an attribute extractor and we compared our attribute result with various existing attributes in literature and demonstrate competitive performance with them.

arXiv:2108.09408공식 코드 · Papers with Code

Multi-scale Edge-based U-shape Network for Salient Object Detection

Deep-learning based salient object detection methods achieve great improvements. However, there are still problems existing in the predictions, such as blurry boundary and inaccurate location, which is mainly caused by inadequate feature extraction and integration. In this paper, we propose a Multi-scale Edge-based U-shape Network (MEUN) to integrate various features at different scales to achieve better performance. To extract more useful information for boundary prediction, U-shape Edge Network modules are embedded in each decoder units. Besides, the additional down-sampling module alleviates the location inaccuracy. Experimental results on four benchmark datasets demonstrate the validity and reliability of the proposed method. Multi-scale Edge based U-shape Network also shows its superiority when compared with 15 state-of-the-art salient object detection methods.

arXiv:2108.10059

ZS-SLR: Zero-Shot Sign Language Recognition from RGB-D Videos

Sign Language Recognition (SLR) is a challenging research area in computer vision. To tackle the annotation bottleneck in SLR, we formulate the problem of Zero-Shot Sign Language Recognition (ZS-SLR) and propose a two-stream model from two input modalities: RGB and Depth videos. To benefit from the vision Transformer capabilities, we use two vision Transformer models, for human detection and visual features representation. We configure a transformer encoder-decoder architecture, as a fast and accurate human detection model, to overcome the challenges of the current human detection models. Considering the human keypoints, the detected human body is segmented into nine parts. A spatio-temporal representation from human body is obtained using a vision Transformer and a LSTM network. A semantic space maps the visual features to the lingual embedding of the class labels via a Bidirectional Encoder Representations from Transformers (BERT) model. We evaluated the proposed model on four datasets, Montalbano II, MSR Daily Activity 3D, CAD-60, and NTU-60, obtaining state-of-the-art results compared to state-of-the-art ZS-SLR models.

arXiv:2108.09894공식 코드 · Papers with Code

CANet: A Context-Aware Network for Shadow Removal

In this paper, we propose a novel two-stage context-aware network named CANet for shadow removal, in which the contextual information from non-shadow regions is transferred to shadow regions at the embedded feature spaces. At Stage-I, we propose a contextual patch matching (CPM) module to generate a set of potential matching pairs of shadow and non-shadow patches. Combined with the potential contextual relationships between shadow and non-shadow regions, our well-designed contextual feature transfer (CFT) mechanism can transfer contextual information from non-shadow to shadow regions at different scales. With the reconstructed feature maps, we remove shadows at L and A/B channels separately. At Stage-II, we use an encoder-decoder to refine current results and generate the final shadow removal results. We evaluate our proposed CANet on two benchmark datasets and some real-world shadow images with complex scenes. Extensive experimental results strongly demonstrate the efficacy of our proposed CANet and exhibit superior performance to state-of-the-arts.

arXiv:2108.09432공식 코드 · Papers with Code

ARAPReg: An As-Rigid-As Possible Regularization Loss for Learning Deformable Shape Generators

This paper introduces an unsupervised loss for training parametric deformation shape generators. The key idea is to enforce the preservation of local rigidity among the generated shapes. Our approach builds on an approximation of the as-rigid-as possible (or ARAP) deformation energy. We show how to develop the unsupervised loss via a spectral decomposition of the Hessian of the ARAP energy. Our loss nicely decouples pose and shape variations through a robust norm. The loss admits simple closed-form expressions. It is easy to train and can be plugged into any standard generation models, e.g., variational auto-encoder (VAE) and auto-decoder (AD). Experimental results show that our approach outperforms existing shape generation approaches considerably on public benchmark datasets of various shape categories such as human, animal and bone.

arXiv:2108.09885

DTWSSE: Data Augmentation with a Siamese Encoder for Time Series

Access to labeled time series data is often limited in the real world, which constrains the performance of deep learning models in the field of time series analysis. Data augmentation is an effective way to solve the problem of small sample size and imbalance in time series datasets. The two key factors of data augmentation are the distance metric and the choice of interpolation method. SMOTE does not perform well on time series data because it uses a Euclidean distance metric and interpolates directly on the object. Therefore, we propose a DTW-based synthetic minority oversampling technique using siamese encoder for interpolation named DTWSSE. In order to reasonably measure the distance of the time series, DTW, which has been verified to be an effective method forts, is employed as the distance metric. To adapt the DTW metric, we use an autoencoder trained in an unsupervised self-training manner for interpolation. The encoder is a Siamese Neural Network for mapping the time series data from the DTW hidden space to the Euclidean deep feature space, and the decoder is used to map the deep feature space back to the DTW hidden space. We validate the proposed methods on a number of different balanced or unbalanced time series datasets. Experimental results show that the proposed method can lead to better performance of the downstream deep learning model.

arXiv:2108.08877

Sentence-T5: Scalable Sentence Encoders from Pre-trained Text-to-Text Models

We provide the first exploration of sentence embeddings from text-to-text transformers (T5). Sentence embeddings are broadly useful for language processing tasks. While T5 achieves impressive performance on language tasks cast as sequence-to-sequence mapping problems, it is unclear how to produce sentence embeddings from encoder-decoder models. We investigate three methods for extracting T5 sentence embeddings: two utilize only the T5 encoder and one uses the full T5 encoder-decoder model. To support our investigation, we establish a new sentence representation transfer benchmark, SentGLUE, which extends the SentEval toolkit to nine tasks from the GLUE benchmark. Our encoder-only models outperforms Sentence-BERT and SimCSE sentence embeddings on both SentEval and SentGLUE transfer tasks, including semantic textual similarity (STS). Scaling up T5 from millions to billions of parameters is found to produce consistent further improvements. Finally, our encoder-decoder method achieves a new state-of-the-art on STS when using sentence embeddings. Our models are released at https://tfhub.dev/google/collections/sentence-t5/1.

arXiv:2108.09174공식 코드 · Papers with Code

Trans4Trans: Efficient Transformer for Transparent Object and Semantic Scene Segmentation in Real-World Navigation Assistance

Transparent objects, such as glass walls and doors, constitute architectural obstacles hindering the mobility of people with low vision or blindness. For instance, the open space behind glass doors is inaccessible, unless it is correctly perceived and interacted with. However, traditional assistive technologies rarely cover the segmentation of these safety-critical transparent objects. In this paper, we build a wearable system with a novel dual-head Transformer for Transparency (Trans4Trans) perception model, which can segment general- and transparent objects. The two dense segmentation results are further combined with depth information in the system to help users navigate safely and assist them to negotiate transparent obstacles. We propose a lightweight Transformer Parsing Module (TPM) to perform multi-scale feature interpretation in the transformer-based decoder. Benefiting from TPM, the double decoders can perform joint learning from corresponding datasets to pursue robustness, meanwhile maintain efficiency on a portable GPU, with negligible calculation increase. The entire Trans4Trans model is constructed in a symmetrical encoder-decoder architecture, which outperforms state-of-the-art methods on the test sets of Stanford2D3D and Trans10K-v2 datasets, obtaining mIoU of 45.13% and 75.14%, respectively. Through a user study and various pre-tests conducted in indoor and outdoor scenes, the usability and reliability of our assistive system have been extensively verified. Meanwhile, the Tran4Trans model has outstanding performances on driving scene datasets. On Cityscapes, ACDC, and DADA-seg datasets corresponding to common environments, adverse weather, and traffic accident scenarios, mIoU scores of 81.5%, 76.3%, and 39.2% are obtained, demonstrating its high efficiency and robustness for real-world transportation applications.

arXiv:2108.08931

Augmenting Implicit Neural Shape Representations with Explicit Deformation Fields

Implicit neural representation is a recent approach to learn shape collections as zero level-sets of neural networks, where each shape is represented by a latent code. So far, the focus has been shape reconstruction, while shape generalization was mostly left to generic encoder-decoder or auto-decoder regularization. In this paper we advocate deformation-aware regularization for implicit neural representations, aiming at producing plausible deformations as latent code changes. The challenge is that implicit representations do not capture correspondences between different shapes, which makes it difficult to represent and regularize their deformations. Thus, we propose to pair the implicit representation of the shapes with an explicit, piecewise linear deformation field, learned as an auxiliary function. We demonstrate that, by regularizing these deformation fields, we can encourage the implicit neural representation to induce natural deformations in the learned shape space, such as as-rigid-as-possible deformations.

arXiv:2108.08839공식 코드 · Papers with Code

PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformers

Point clouds captured in real-world applications are often incomplete due to the limited sensor resolution, single viewpoint, and occlusion. Therefore, recovering the complete point clouds from partial ones becomes an indispensable task in many practical applications. In this paper, we present a new method that reformulates point cloud completion as a set-to-set translation problem and design a new model, called PoinTr that adopts a transformer encoder-decoder architecture for point cloud completion. By representing the point cloud as a set of unordered groups of points with position embeddings, we convert the point cloud to a sequence of point proxies and employ the transformers for point cloud generation. To facilitate transformers to better leverage the inductive bias about 3D geometric structures of point clouds, we further devise a geometry-aware block that models the local geometric relationships explicitly. The migration of transformers enables our model to better learn structural knowledge and preserve detailed information for point cloud completion. Furthermore, we propose two more challenging benchmarks with more diverse incomplete point clouds that can better reflect the real-world scenarios to promote future research. Experimental results show that our method outperforms state-of-the-art methods by a large margin on both the new benchmarks and the existing ones. Code is available at https://github.com/yuxumin/PoinTr

arXiv:2108.08760공식 코드 · Papers with Code

Robust outlier detection by de-biasing VAE likelihoods

Deep networks often make confident, yet, incorrect, predictions when tested with outlier data that is far removed from their training distributions. Likelihoods computed by deep generative models (DGMs) are a candidate metric for outlier detection with unlabeled data. Yet, previous studies have shown that DGM likelihoods are unreliable and can be easily biased by simple transformations to input data. Here, we examine outlier detection with variational autoencoders (VAEs), among the simplest of DGMs. We propose novel analytical and algorithmic approaches to ameliorate key biases with VAE likelihoods. Our bias corrections are sample-specific, computationally inexpensive, and readily computed for various decoder visible distributions. Next, we show that a well-known image pre-processing technique -- contrast stretching -- extends the effectiveness of bias correction to further improve outlier detection. Our approach achieves state-of-the-art accuracies with nine grayscale and natural image datasets, and demonstrates significant advantages -- both with speed and performance -- over four recent, competing approaches. In summary, lightweight remedies suffice to achieve robust outlier detection with VAEs.

arXiv:2108.08669공식 코드 · Papers with Code

Video Relation Detection via Tracklet based Visual Transformer

Video Visual Relation Detection (VidVRD), has received significant attention of our community over recent years. In this paper, we apply the state-of-the-art video object tracklet detection pipeline MEGA and deepSORT to generate tracklet proposals. Then we perform VidVRD in a tracklet-based manner without any pre-cutting operations. Specifically, we design a tracklet-based visual Transformer. It contains a temporal-aware decoder which performs feature interactions between the tracklets and learnable predicate query embeddings, and finally predicts the relations. Experimental results strongly demonstrate the superiority of our method, which outperforms other methods by a large margin on the Video Relation Understanding (VRU) Grand Challenge in ACM Multimedia 2021. Codes are released at https://github.com/Dawn-LX/VidVRD-tracklets.

arXiv:2108.07980

A Multi-level Acoustic Feature Extraction Framework for Transformer Based End-to-End Speech Recognition

Transformer based end-to-end modelling approaches with multiple stream inputs have been achieved great success in various automatic speech recognition (ASR) tasks. An important issue associated with such approaches is that the intermediate features derived from each stream might have similar representations and thus it is lacking of feature diversity, such as the descriptions related to speaker characteristics. To address this issue, this paper proposed a novel multi-level acoustic feature extraction framework that can be easily combined with Transformer based ASR models. The framework consists of two input streams: a shallow stream with high-resolution spectrograms and a deep stream with low-resolution spectrograms. The shallow stream is used to acquire traditional shallow features that is beneficial for the classification of phones or words while the deep stream is used to obtain utterance-level speaker-invariant deep features for improving the feature diversity. A feature correlation based fusion strategy is used to aggregate both features across the frequency and time domains and then fed into the Transformer encoder-decoder module. By using the proposed multi-level acoustic feature extraction framework, state-of-the-art word error rate of 21.7% and 2.5% were obtained on the HKUST Mandarin telephone and Librispeech speech recognition tasks respectively.

arXiv:2108.08162공식 코드 · Papers with Code

Specificity-preserving RGB-D Saliency Detection

Salient object detection (SOD) on RGB and depth images has attracted more and more research interests, due to its effectiveness and the fact that depth cues can now be conveniently captured. Existing RGB-D SOD models usually adopt different fusion strategies to learn a shared representation from the two modalities (\ie, RGB and depth), while few methods explicitly consider how to preserve modality-specific characteristics. In this study, we propose a novel framework, termed SPNet} (Specificity-preserving network), which benefits SOD performance by exploring both the shared information and modality-specific properties (\eg, specificity). Specifically, we propose to adopt two modality-specific networks and a shared learning network to generate individual and shared saliency prediction maps, respectively. To effectively fuse cross-modal features in the shared learning network, we propose a cross-enhanced integration module (CIM) and then propagate the fused feature to the next layer for integrating cross-level information. Moreover, to capture rich complementary multi-modal information for boosting the SOD performance, we propose a multi-modal feature aggregation (MFA) module to integrate the modality-specific features from each individual decoder into the shared decoder. By using a skip connection, the hierarchical features between the encoder and decoder layers can be fully combined. Extensive experiments demonstrate that our~\ours~outperforms cutting-edge approaches on six popular RGB-D SOD and three camouflaged object detection benchmarks. The project is publicly available at: https://github.com/taozh2017/SPNet.

arXiv:2108.08157공식 코드 · Papers with Code

Towards Deep and Efficient: A Deep Siamese Self-Attention Fully Efficient Convolutional Network for Change Detection in VHR Images

Recently, FCNs have attracted widespread attention in the CD field. In pursuit of better CD performance, it has become a tendency to design deeper and more complicated FCNs, which inevitably brings about huge numbers of parameters and an unbearable computational burden. With the goal of designing a quite deep architecture to obtain more precise CD results while simultaneously decreasing parameter numbers to improve efficiency, in this work, we present a very deep and efficient CD network, entitled EffCDNet. In EffCDNet, to reduce the numerous parameters associated with deep architecture, an efficient convolution consisting of depth-wise convolution and group convolution with a channel shuffle mechanism is introduced to replace standard convolutional layers. In terms of the specific network architecture, EffCDNet does not use mainstream UNet-like architecture, but rather adopts the architecture with a very deep encoder and a lightweight decoder. In the very deep encoder, two very deep siamese streams stacked by efficient convolution first extract two highly representative and informative feature maps from input image-pairs. Subsequently, an efficient ASPP module is designed to capture multi-scale change information. In the lightweight decoder, a recurrent criss-cross self-attention (RCCA) module is applied to efficiently utilize non-local similar feature representations to enhance discriminability for each pixel, thus effectively separating the changed and unchanged regions. Moreover, to tackle the optimization problem in confused pixels, two novel loss functions based on information entropy are presented. On two challenging CD datasets, our approach outperforms other SOTA FCN-based methods, with only benchmark-level parameter numbers and quite low computational overhead.

arXiv:2108.08042

Joint Multiple Intent Detection and Slot Filling via Self-distillation

Intent detection and slot filling are two main tasks in natural language understanding (NLU) for identifying users' needs from their utterances. These two tasks are highly related and often trained jointly. However, most previous works assume that each utterance only corresponds to one intent, ignoring the fact that a user utterance in many cases could include multiple intents. In this paper, we propose a novel Self-Distillation Joint NLU model (SDJN) for multi-intent NLU. First, we formulate multiple intent detection as a weakly supervised problem and approach with multiple instance learning (MIL). Then, we design an auxiliary loop via self-distillation with three orderly arranged decoders: Initial Slot Decoder, MIL Intent Decoder, and Final Slot Decoder. The output of each decoder will serve as auxiliary information for the next decoder. With the auxiliary knowledge provided by the MIL Intent Decoder, we set Final Slot Decoder as the teacher model that imparts knowledge back to Initial Slot Decoder to complete the loop. The auxiliary loop enables intents and slots to guide mutually in-depth and further boost the overall NLU performance. Experimental results on two public multi-intent datasets indicate that our model achieves strong performance compared to others.

arXiv:2108.07851공식 코드 · Papers with Code

Boosting Salient Object Detection with Transformer-based Asymmetric Bilateral U-Net

Existing salient object detection (SOD) methods mainly rely on U-shaped convolution neural networks (CNNs) with skip connections to combine the global contexts and local spatial details that are crucial for locating salient objects and refining object details, respectively. Despite great successes, the ability of CNNs in learning global contexts is limited. Recently, the vision transformer has achieved revolutionary progress in computer vision owing to its powerful modeling of global dependencies. However, directly applying the transformer to SOD is suboptimal because the transformer lacks the ability to learn local spatial representations. To this end, this paper explores the combination of transformers and CNNs to learn both global and local representations for SOD. We propose a transformer-based Asymmetric Bilateral U-Net (ABiU-Net). The asymmetric bilateral encoder has a transformer path and a lightweight CNN path, where the two paths communicate at each encoder stage to learn complementary global contexts and local spatial details, respectively. The asymmetric bilateral decoder also consists of two paths to process features from the transformer and CNN encoder paths, with communication at each decoder stage for decoding coarse salient object locations and fine-grained object details, respectively. Such communication between the two encoder/decoder paths enables AbiU-Net to learn complementary global and local representations, taking advantage of the natural merits of transformers and CNNs, respectively. Hence, ABiU-Net provides a new perspective for transformer-based SOD. Extensive experiments demonstrate that ABiU-Net performs favorably against previous state-of-the-art SOD methods. The code is available at https://github.com/yuqiuyuqiu/ABiU-Net.

arXiv:2108.08241

Semi-Supervised Learning for Channel Charting-Aided IoT Localization in Millimeter Wave Networks

In this paper, a novel framework is proposed for channel charting (CC)-aided localization in millimeter wave networks. In particular, a convolutional autoencoder model is proposed to estimate the three-dimensional location of wireless user equipment (UE), based on multipath channel state information (CSI), received by different base stations. In order to learn the radio-geometry map and capture the relative position of each UE, an autoencoder-based channel chart is constructed in an unsupervised manner, such that neighboring UEs in the physical space will remain close in the channel chart. Next, the channel charting model is extended to a semi-supervised framework, where the autoencoder is divided into two components: an encoder and a decoder, and each component is optimized individually, using the labeled CSI dataset with associated location information, to further improve positioning accuracy. Simulation results show that the proposed CC-aided semi-supervised localization yields a higher accuracy, compared with existing supervised positioning and conventional unsupervised CC approaches.

arXiv:2108.08217공식 코드 · Papers with Code

X-modaler: A Versatile and High-performance Codebase for Cross-modal Analytics

With the rise and development of deep learning over the past decade, there has been a steady momentum of innovation and breakthroughs that convincingly push the state-of-the-art of cross-modal analytics between vision and language in multimedia field. Nevertheless, there has not been an open-source codebase in support of training and deploying numerous neural network models for cross-modal analytics in a unified and modular fashion. In this work, we propose X-modaler -- a versatile and high-performance codebase that encapsulates the state-of-the-art cross-modal analytics into several general-purpose stages (e.g., pre-processing, encoder, cross-modal interaction, decoder, and decode strategy). Each stage is empowered with the functionality that covers a series of modules widely adopted in state-of-the-arts and allows seamless switching in between. This way naturally enables a flexible implementation of state-of-the-art algorithms for image captioning, video captioning, and vision-language pre-training, aiming to facilitate the rapid development of research community. Meanwhile, since the effective modular designs in several stages (e.g., cross-modal interaction) are shared across different vision-language tasks, X-modaler can be simply extended to power startup prototypes for other tasks in cross-modal analytics, including visual question answering, visual commonsense reasoning, and cross-modal retrieval. X-modaler is an Apache-licensed codebase, and its source codes, sample projects and pre-trained models are available on-line: https://github.com/YehLi/xmodaler.

arXiv:2108.07958공식 코드 · Papers with Code

Semantic Perturbations with Normalizing Flows for Improved Generalization

Data augmentation is a widely adopted technique for avoiding overfitting when training deep neural networks. However, this approach requires domain-specific knowledge and is often limited to a fixed set of hard-coded transformations. Recently, several works proposed to use generative models for generating semantically meaningful perturbations to train a classifier. However, because accurate encoding and decoding are critical, these methods, which use architectures that approximate the latent-variable inference, remained limited to pilot studies on small datasets. Exploiting the exactly reversible encoder-decoder structure of normalizing flows, we perform on-manifold perturbations in the latent space to define fully unsupervised data augmentations. We demonstrate that such perturbations match the performance of advanced data augmentation techniques -- reaching 96.6% test accuracy for CIFAR-10 using ResNet-18 and outperform existing methods, particularly in low data regimes -- yielding 10--25% relative improvement of test accuracy from classical training. We find that our latent adversarial perturbations adaptive to the classifier throughout its training are most effective, yielding the first test accuracy improvement results on real-world datasets -- CIFAR-10/100 -- via latent-space perturbations.