PyTorchKR Papers

최신 논문

총 9,642편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2102.03111

3D Medical Multi-modal Segmentation Network Guided by Multi-source Correlation Constraint

In the field of multimodal segmentation, the correlation between different modalities can be considered for improving the segmentation results. In this paper, we propose a multi-modality segmentation network with a correlation constraint. Our network includes N model-independent encoding paths with N image sources, a correlation constraint block, a feature fusion block, and a decoding path. The model independent encoding path can capture modality-specific features from the N modalities. Since there exists a strong correlation between different modalities, we first propose a linear correlation block to learn the correlation between modalities, then a loss function is used to guide the network to learn the correlated features based on the linear correlation block. This block forces the network to learn the latent correlated features which are more relevant for segmentation. Considering that not all the features extracted from the encoders are useful for segmentation, we propose to use dual attention based fusion block to recalibrate the features along the modality and spatial paths, which can suppress less informative features and emphasize the useful ones. The fused feature representation is finally projected by the decoder to obtain the segmentation result. Our experiment results tested on BraTS-2018 dataset for brain tumor segmentation demonstrate the effectiveness of our proposed method.

논문 ID: 4171.34138

Beyond the Attention: Distinguish the Discriminative and Confusable Features For Fine-grained Image Classification

Learning subtle discriminative features plays a significant role in fine-grained image classification. Existing methods usually extract the distinguishable parts through the attention module for classification. Although these learned distinguishable parts contain valuable features that are beneficial for classification, part of irrelevant features are also preserved, which may confuse the model to make a correct classification, especially for the fine-grained tasks due to their similarities. How to keep the discriminative features while removing confusable features from the distinguishable parts is an interesting yet changeling task. In this paper, we introduce a novel classification approach, named Logical-based Feature Extraction Model (LAFE for short) to address this issue. The main advantage of LAFE lies in the fact that it can explicitly add the significance of discriminative features and subtract the confusable features. Specifically, LAFE utilizes the region attention modules and channel attention modules to extract discriminative features and confusable features respectively. Based on this, two novel loss functions are designed to automatically induce attention over these features for fine-grained image classification. Our approach demonstrates its robustness, efficiency, and state-of-the-art performance on three benchmark datasets.

arXiv:2102.03285

Unsupervised Novel View Synthesis from a Single Image

Novel view synthesis from a single image aims at generating novel views from a single input image of an object. Several works recently achieved remarkable results, though require some form of multi-view supervision at training time, therefore limiting their deployment in real scenarios. This work aims at relaxing this assumption enabling training of conditional generative model for novel view synthesis in a completely unsupervised manner. We first pre-train a purely generative decoder model using a GAN formulation while at the same time training an encoder network to invert the mapping from latent code to images. Then we swap encoder and decoder and train the network as a conditioned GAN with a mixture of auto-encoder-like objective and self-distillation. At test time, given a view of an object, our model first embeds the image content in a latent code and regresses its pose w.r.t. a canonical reference system, then generates novel views of it by keeping the code and varying the pose. We show that our framework achieves results comparable to the state of the art on ShapeNet and that it can be employed on unconstrained collections of natural images, where no competing method can be trained.

arXiv:2102.03216

Intermediate Loss Regularization for CTC-based Speech Recognition

We present a simple and efficient auxiliary loss function for automatic speech recognition (ASR) based on the connectionist temporal classification (CTC) objective. The proposed objective, an intermediate CTC loss, is attached to an intermediate layer in the CTC encoder network. This intermediate CTC loss well regularizes CTC training and improves the performance requiring only small modification of the code and small and no overhead during training and inference, respectively. In addition, we propose to combine this intermediate CTC loss with stochastic depth training, and apply this combination to a recently proposed Conformer network. We evaluate the proposed method on various corpora, reaching word error rate (WER) 9.9% on the WSJ corpus and character error rate (CER) 5.2% on the AISHELL-1 corpus respectively, based on CTC greedy search without a language model. Especially, the AISHELL-1 task is comparable to other state-of-the-art ASR systems based on auto-regressive decoder with beam search.

arXiv:2102.02636

Deep Autoencoder-based Fuzzy C-Means for Topic Detection

Topic detection is a process for determining topics from a collection of textual data. One of the topic detection methods is a clustering-based method, which assumes that the centroids are topics. The clustering method has the advantage that it can process data with negative representations. Therefore, the clustering method allows a combination with a broader representation learning method. In this paper, we adopt deep learning for topic detection by using a deep autoencoder and fuzzy c-means called deep autoencoder-based fuzzy c-means (DFCM). The encoder of the autoencoder performs a lower-dimensional representation learning. Fuzzy c-means groups the lower-dimensional representation to identify the centroids. The autoencoder's decoder transforms back the centroids into the original representation to be interpreted as the topics. Our simulation shows that DFCM improves the coherence score of eigenspace-based fuzzy c-means (EFCM) and is comparable to the leading standard methods, i.e., nonnegative matrix factorization (NMF) or latent Dirichlet allocation (LDA).

arXiv:2102.02810공식 코드 · Papers with Code

Controlling Hallucinations at Word Level in Data-to-Text Generation

Data-to-Text Generation (DTG) is a subfield of Natural Language Generation aiming at transcribing structured data in natural language descriptions. The field has been recently boosted by the use of neural-based generators which exhibit on one side great syntactic skills without the need of hand-crafted pipelines; on the other side, the quality of the generated text reflects the quality of the training data, which in realistic settings only offer imperfectly aligned structure-text pairs. Consequently, state-of-art neural models include misleading statements - usually called hallucinations - in their outputs. The control of this phenomenon is today a major challenge for DTG, and is the problem addressed in the paper. Previous work deal with this issue at the instance level: using an alignment score for each table-reference pair. In contrast, we propose a finer-grained approach, arguing that hallucinations should rather be treated at the word level. Specifically, we propose a Multi-Branch Decoder which is able to leverage word-level labels to learn the relevant parts of each training instance. These labels are obtained following a simple and efficient scoring procedure based on co-occurrence analysis and dependency parsing. Extensive evaluations, via automated metrics and human judgment on the standard WikiBio benchmark, show the accuracy of our alignment labels and the effectiveness of the proposed Multi-Branch Decoder. Our model is able to reduce and control hallucinations, while keeping fluency and coherence in generated texts. Further experiments on a degraded version of ToTTo show that our model could be successfully used on very noisy settings.

arXiv:2102.02808공식 코드 · Papers with Code

Multi-Stage Progressive Image Restoration

Image restoration tasks demand a complex balance between spatial details and high-level contextualized information while recovering images. In this paper, we propose a novel synergistic design that can optimally balance these competing goals. Our main proposal is a multi-stage architecture, that progressively learns restoration functions for the degraded inputs, thereby breaking down the overall recovery process into more manageable steps. Specifically, our model first learns the contextualized features using encoder-decoder architectures and later combines them with a high-resolution branch that retains local information. At each stage, we introduce a novel per-pixel adaptive design that leverages in-situ supervised attention to reweight the local features. A key ingredient in such a multi-stage architecture is the information exchange between different stages. To this end, we propose a two-faceted approach where the information is not only exchanged sequentially from early to late stages, but lateral connections between feature processing blocks also exist to avoid any loss of information. The resulting tightly interlinked multi-stage architecture, named as MPRNet, delivers strong performance gains on ten datasets across a range of tasks including image deraining, deblurring, and denoising. The source code and pre-trained models are available at https://github.com/swz30/MPRNet.

arXiv:2102.02779공식 코드 · Papers with Code

Unifying Vision-and-Language Tasks via Text Generation

Existing methods for vision-and-language learning typically require designing task-specific architectures and objectives for each task. For example, a multi-label answer classifier for visual question answering, a region scorer for referring expression comprehension, and a language decoder for image captioning, etc. To alleviate these hassles, in this work, we propose a unified framework that learns different tasks in a single architecture with the same language modeling objective, i.e., multimodal conditional text generation, where our models learn to generate labels in text based on the visual and textual inputs. On 7 popular vision-and-language benchmarks, including visual question answering, referring expression comprehension, visual commonsense reasoning, most of which have been previously modeled as discriminative tasks, our generative approach (with a single unified architecture) reaches comparable performance to recent task-specific state-of-the-art vision-and-language models. Moreover, our generative approach shows better generalization ability on questions that have rare answers. Also, we show that our framework allows multi-task learning in a single architecture with a single set of parameters, achieving similar performance to separately optimized single-task models. Our code is publicly available at: https://github.com/j-min/VL-T5

arXiv:2102.02723공식 코드 · Papers with Code

Data-to-text Generation with Macro Planning

Recent approaches to data-to-text generation have adopted the very successful encoder-decoder architecture or variants thereof. These models generate text which is fluent (but often imprecise) and perform quite poorly at selecting appropriate content and ordering it coherently. To overcome some of these issues, we propose a neural model with a macro planning stage followed by a generation stage reminiscent of traditional methods which embrace separate modules for planning and surface realization. Macro plans represent high level organization of important content such as entities, events and their interactions; they are learnt from data and given as input to the generator. Extensive experiments on two data-to-text benchmarks (RotoWire and MLB) show that our approach outperforms competitive baselines in terms of automatic and human evaluation.

arXiv:2102.01993공식 코드 · Papers with Code

Monaural Speech Enhancement with Complex Convolutional Block Attention Module and Joint Time Frequency Losses

Deep complex U-Net structure and convolutional recurrent network (CRN) structure achieve state-of-the-art performance for monaural speech enhancement. Both deep complex U-Net and CRN are encoder and decoder structures with skip connections, which heavily rely on the representation power of the complex-valued convolutional layers. In this paper, we propose a complex convolutional block attention module (CCBAM) to boost the representation power of the complex-valued convolutional layers by constructing more informative features. The CCBAM is a lightweight and general module which can be easily integrated into any complex-valued convolutional layers. We integrate CCBAM with the deep complex U-Net and CRN to enhance their performance for speech enhancement. We further propose a mixed loss function to jointly optimize the complex models in both time-frequency (TF) domain and time domain. By integrating CCBAM and the mixed loss, we form a new end-to-end (E2E) complex speech enhancement framework. Ablation experiments and objective evaluations show the superior performance of the proposed approaches (https://github.com/modelscope/ClearerVoice-Studio).

arXiv:2102.02080공식 코드 · Papers with Code

Top-down Discourse Parsing via Sequence Labelling

We introduce a top-down approach to discourse parsing that is conceptually simpler than its predecessors (Kobayashi et al., 2020; Zhang et al., 2020). By framing the task as a sequence labelling problem where the goal is to iteratively segment a document into individual discourse units, we are able to eliminate the decoder and reduce the search space for splitting points. We explore both traditional recurrent models and modern pre-trained transformer models for the task, and additionally introduce a novel dynamic oracle for top-down parsing. Based on the Full metric, our proposed LSTM model sets a new state-of-the-art for RST parsing.

arXiv:2102.01822

Multi-class probabilistic atlas-based whole heart segmentation method in cardiac CT and MRI

Accurate and robust whole heart substructure segmentation is crucial in developing clinical applications, such as computer-aided diagnosis and computer-aided surgery. However, segmentation of different heart substructures is challenging because of inadequate edge or boundary information, the complexity of the background and texture, and the diversity in different substructures' sizes and shapes. This article proposes a framework for multi-class whole heart segmentation employing non-rigid registration-based probabilistic atlas incorporating the Bayesian framework. We also propose a non-rigid registration pipeline utilizing a multi-resolution strategy for obtaining the highest attainable mutual information between the moving and fixed images. We further incorporate non-rigid registration into the expectation-maximization algorithm and implement different deep convolutional neural network-based encoder-decoder networks for ablation studies. All the extensive experiments are conducted utilizing the publicly available dataset for the whole heart segmentation containing 20 MRI and 20 CT cardiac images. The proposed approach exhibits an encouraging achievement, yielding a mean volume overlapping error of 14.5 % for CT scans exceeding the state-of-the-art results by a margin of 1.3 % in terms of the same metric. As the proposed approach provides better-results to delineate the different substructures of the heart, it can be a medical diagnostic aiding tool for helping experts with quicker and more accurate results.

arXiv:2102.01380

Internal Language Model Training for Domain-Adaptive End-to-End Speech Recognition

The efficacy of external language model (LM) integration with existing end-to-end (E2E) automatic speech recognition (ASR) systems can be improved significantly using the internal language model estimation (ILME) method. In this method, the internal LM score is subtracted from the score obtained by interpolating the E2E score with the external LM score, during inference. To improve the ILME-based inference, we propose an internal LM training (ILMT) method to minimize an additional internal LM loss by updating only the E2E model components that affect the internal LM estimation. ILMT encourages the E2E model to form a standalone LM inside its existing components, without sacrificing ASR accuracy. After ILMT, the more modular E2E model with matched training and inference criteria enables a more thorough elimination of the source-domain internal LM, and therefore leads to a more effective integration of the target-domain external LM. Experimented with 30K-hour trained recurrent neural network transducer and attention-based encoder-decoder models, ILMT with ILME-based inference achieves up to 31.5% and 11.4% relative word error rate reductions from standard E2E training with Shallow Fusion on out-of-domain LibriSpeech and in-domain Microsoft production test sets, respectively.

arXiv:2102.01120

RectiNet-v2: A stacked network architecture for document image dewarping

With the advent of mobile and hand-held cameras, document images have found their way into almost every domain. Dewarping of these images for the removal of perspective distortions and folds is essential so that they can be understood by document recognition algorithms. For this, we propose an end-to-end CNN architecture that can produce distortion free document images from warped documents it takes as input. We train this model on warped document images simulated synthetically to compensate for lack of enough natural data. Our method is novel in the use of a bifurcated decoder with shared weights to prevent intermingling of grid coordinates, in the use of residual networks in the U-Net skip connections to allow flow of data from different receptive fields in the model, and in the use of a gated network to help the model focus on structure and line level detail of the document image. We evaluate our method on the DocUNet dataset, a benchmark in this domain, and obtain results comparable to state-of-the-art methods.

arXiv:2102.01547공식 코드 · Papers with Code

WeNet: Production oriented Streaming and Non-streaming End-to-End Speech Recognition Toolkit

In this paper, we propose an open source, production first, and production ready speech recognition toolkit called WeNet in which a new two-pass approach is implemented to unify streaming and non-streaming end-to-end (E2E) speech recognition in a single model. The main motivation of WeNet is to close the gap between the research and the production of E2E speechrecognition models. WeNet provides an efficient way to ship ASR applications in several real-world scenarios, which is the main difference and advantage to other open source E2E speech recognition toolkits. In our toolkit, a new two-pass method is implemented. Our method propose a dynamic chunk-based attention strategy of the the transformer layers to allow arbitrary right context length modifies in hybrid CTC/attention architecture. The inference latency could be easily controlled by only changing the chunk size. The CTC hypotheses are then rescored by the attention decoder to get the final result. Our experiments on the AISHELL-1 dataset using WeNet show that, our model achieves 5.03\% relative character error rate (CER) reduction in non-streaming ASR compared to a standard non-streaming transformer. After model quantification, our model perform reasonable RTF and latency.

arXiv:2102.01192공식 코드 · Papers with Code

Generative Spoken Language Modeling from Raw Audio

We introduce Generative Spoken Language Modeling, the task of learning the acoustic and linguistic characteristics of a language from raw audio (no text, no labels), and a set of metrics to automatically evaluate the learned representations at acoustic and linguistic levels for both encoding and generation. We set up baseline systems consisting of a discrete speech encoder (returning pseudo-text units), a generative language model (trained on pseudo-text), and a speech decoder (generating a waveform from pseudo-text) all trained without supervision and validate the proposed metrics with human evaluation. Across 3 speech encoders (CPC, wav2vec 2.0, HuBERT), we find that the number of discrete units (50, 100, or 200) matters in a task-dependent and encoder-dependent way, and that some combinations approach text-based systems.

arXiv:2102.00769

GTAE: Graph-Transformer based Auto-Encoders for Linguistic-Constrained Text Style Transfer

Non-parallel text style transfer has attracted increasing research interests in recent years. Despite successes in transferring the style based on the encoder-decoder framework, current approaches still lack the ability to preserve the content and even logic of original sentences, mainly due to the large unconstrained model space or too simplified assumptions on latent embedding space. Since language itself is an intelligent product of humans with certain grammars and has a limited rule-based model space by its nature, relieving this problem requires reconciling the model capacity of deep neural networks with the intrinsic model constraints from human linguistic rules. To this end, we propose a method called Graph Transformer based Auto Encoder (GTAE), which models a sentence as a linguistic graph and performs feature extraction and style transfer at the graph level, to maximally retain the content and the linguistic structure of original sentences. Quantitative experiment results on three non-parallel text style transfer tasks show that our model outperforms state-of-the-art methods in content preservation, while achieving comparable performance on transfer accuracy and sentence naturalness.

arXiv:2102.00676

Underwater Image Enhancement via Learning Water Type Desensitized Representations

We present a novel underwater image enhancement method termed SCNet to improve the image quality meanwhile cope with the degradation diversity caused by the water. SCNet is based on normalization schemes across both spatial and channel dimensions with the key idea of learning water type desensitized features. Specifically, we apply whitening to de-correlate activations across spatial dimensions for each instance in a mini-batch. We also eliminate channel-wise correlation by standardizing and re-injecting the first two moments of the activations across channels. The normalization schemes of spatial and channel dimensions are performed at each scale of the U-Net to obtain multi-scale representations. With such water type irrelevant encodings, the decoder can easily reconstruct the clean signal and be unaffected by the distortion types. Experimental results on two real-world underwater image datasets show that our approach can successfully enhance images with diverse water types, and achieves competitive performance in visual quality improvement.

arXiv:2102.00364

OAS-Net: Occlusion Aware Sampling Network for Accurate Optical Flow

Optical flow estimation is an essential step for many real-world computer vision tasks. Existing deep networks have achieved satisfactory results by mostly employing a pyramidal coarse-to-fine paradigm, where a key process is to adopt warped target feature based on previous flow prediction to correlate with source feature for building 3D matching cost volume. However, the warping operation can lead to troublesome ghosting problem that results in ambiguity. Moreover, occluded areas are treated equally with non occluded regions in most existing works, which may cause performance degradation. To deal with these challenges, we propose a lightweight yet efficient optical flow network, named OAS-Net (occlusion aware sampling network) for accurate optical flow. First, a new sampling based correlation layer is employed without noisy warping operation. Second, a novel occlusion aware module is presented to make raw cost volume conscious of occluded regions. Third, a shared flow and occlusion awareness decoder is adopted for structure compactness. Experiments on Sintel and KITTI datasets demonstrate the effectiveness of proposed approaches.

arXiv:2102.00228

MUSE: Multi-Scale Temporal Features Evolution for Knowledge Tracing

Transformer based knowledge tracing model is an extensively studied problem in the field of computer-aided education. By integrating temporal features into the encoder-decoder structure, transformers can processes the exercise information and student response information in a natural way. However, current state-of-the-art transformer-based variants still share two limitations. First, extremely long temporal features cannot well handled as the complexity of self-attention mechanism is O(n2). Second, existing approaches track the knowledge drifts under fixed a window size, without considering different temporal-ranges. To conquer these problems, we propose MUSE, which is equipped with multi-scale temporal sensor unit, that takes either local or global temporal features into consideration. The proposed model is capable to capture the dynamic changes in users knowledge states at different temporal-ranges, and provides an efficient and powerful way to combine local and global features to make predictions. Our method won the 5-th place over 3,395 teams in the Riiid AIEd Challenge 2020.