PyTorchKR Papers

최신 논문

총 9,642편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2008.10544

TORNADO-Net: mulTiview tOtal vaRiatioN semAntic segmentation with Diamond inceptiOn module

Semantic segmentation of point clouds is a key component of scene understanding for robotics and autonomous driving. In this paper, we introduce TORNADO-Net - a neural network for 3D LiDAR point cloud semantic segmentation. We incorporate a multi-view (bird-eye and range) projection feature extraction with an encoder-decoder ResNet architecture with a novel diamond context block. Current projection-based methods do not take into account that neighboring points usually belong to the same class. To better utilize this local neighbourhood information and reduce noisy predictions, we introduce a combination of Total Variation, Lovasz-Softmax, and Weighted Cross-Entropy losses. We also take advantage of the fact that the LiDAR data encompasses 360 degrees field of view and uses circular padding. We demonstrate state-of-the-art results on the SemanticKITTI dataset and also provide thorough quantitative evaluations and ablation results.

arXiv:2008.10487

EfficientFCN: Holistically-guided Decoding for Semantic Segmentation

Both performance and efficiency are important to semantic segmentation. State-of-the-art semantic segmentation algorithms are mostly based on dilated Fully Convolutional Networks (dilatedFCN), which adopt dilated convolutions in the backbone networks to extract high-resolution feature maps for achieving high-performance segmentation performance. However, due to many convolution operations are conducted on the high-resolution feature maps, such dilatedFCN-based methods result in large computational complexity and memory consumption. To balance the performance and efficiency, there also exist encoder-decoder structures that gradually recover the spatial information by combining multi-level feature maps from the encoder. However, the performances of existing encoder-decoder methods are far from comparable with the dilatedFCN-based methods. In this paper, we propose the EfficientFCN, whose backbone is a common ImageNet pre-trained network without any dilated convolution. A holistically-guided decoder is introduced to obtain the high-resolution semantic-rich feature maps via the multi-scale features from the encoder. The decoding task is converted to novel codebook generation and codeword assembly task, which takes advantages of the high-level and low-level features from the encoder. Such a framework achieves comparable or even better performance than state-of-the-art methods with only 1/3 of the computational cost. Extensive experiments on PASCAL Context, PASCAL VOC, ADE20K validate the effectiveness of the proposed EfficientFCN.

arXiv:2008.11151공식 코드 · Papers with Code

FastSal: a Computationally Efficient Network for Visual Saliency Prediction

This paper focuses on the problem of visual saliency prediction, predicting regions of an image that tend to attract human visual attention, under a constrained computational budget. We modify and test various recent efficient convolutional neural network architectures like EfficientNet and MobileNetV2 and compare them with existing state-of-the-art saliency models such as SalGAN and DeepGaze II both in terms of standard accuracy metrics like AUC and NSS, and in terms of the computational complexity and model size. We find that MobileNetV2 makes an excellent backbone for a visual saliency model and can be effective even without a complex decoder. We also show that knowledge transfer from a more computationally expensive model like DeepGaze II can be achieved via pseudo-labelling an unlabelled dataset, and that this approach gives result on-par with many state-of-the-art algorithms with a fraction of the computational cost and model size. Source code is available at https://github.com/feiyanhu/FastSal.

arXiv:2008.09151

Multi-modal Cooking Workflow Construction for Food Recipes

Understanding food recipe requires anticipating the implicit causal effects of cooking actions, such that the recipe can be converted into a graph describing the temporal workflow of the recipe. This is a non-trivial task that involves common-sense reasoning. However, existing efforts rely on hand-crafted features to extract the workflow graph from recipes due to the lack of large-scale labeled datasets. Moreover, they fail to utilize the cooking images, which constitute an important part of food recipes. In this paper, we build MM-ReS, the first large-scale dataset for cooking workflow construction, consisting of 9,850 recipes with human-labeled workflow graphs. Cooking steps are multi-modal, featuring both text instructions and cooking images. We then propose a neural encoder-decoder model that utilizes both visual and textual information to construct the cooking workflow, which achieved over 20% performance gain over existing hand-crafted baselines.

arXiv:2008.09396공식 코드 · Papers with Code

Neural Machine Translation without Embeddings

Many NLP models operate over sequences of subword tokens produced by hand-crafted tokenization rules and heuristic subword induction algorithms. A simple universal alternative is to represent every computerized text as a sequence of bytes via UTF-8, obviating the need for an embedding layer since there are fewer token types (256) than dimensions. Surprisingly, replacing the ubiquitous embedding layer with one-hot representations of each byte does not hurt performance; experiments on byte-to-byte machine translation from English to 10 different languages show a consistent improvement in BLEU, rivaling character-level and even standard subword-level models. A deeper investigation reveals that the combination of embeddingless models with decoder-input dropout amounts to token dropout, which benefits byte-to-byte models in particular.

arXiv:2008.09333

Tweet to News Conversion: An Investigation into Unsupervised Controllable Text Generation

Text generator systems have become extremely popular with the advent of recent deep learning models such as encoder-decoder. Controlling the information and style of the generated output without supervision is an important and challenging Natural Language Processing (NLP) task. In this paper, we define the task of constructing a coherent paragraph from a set of disaster domain tweets, without any parallel data. We tackle the problem by building two systems in pipeline. The first system focuses on unsupervised style transfer and converts the individual tweets into news sentences. The second system stitches together the outputs from the first system to form a coherent news paragraph. We also propose a novel training mechanism, by splitting the sentences into propositions and training the second system to merge the sentences. We create a validation and test set consisting of tweet-sets and their equivalent news paragraphs to perform empirical evaluation. In a completely unsupervised setting, our model was able to achieve a BLEU score of 19.32, while successfully transferring styles and joining tweets to form a meaningful news paragraph.

arXiv:2008.09249공식 코드 · Papers with Code

GRIT: Generative Role-filler Transformers for Document-level Event Entity Extraction

We revisit the classic problem of document-level role-filler entity extraction (REE) for template filling. We argue that sentence-level approaches are ill-suited to the task and introduce a generative transformer-based encoder-decoder framework (GRIT) that is designed to model context at the document level: it can make extraction decisions across sentence boundaries; is implicitly aware of noun phrase coreference structure, and has the capacity to respect cross-role dependencies in the template structure. We evaluate our approach on the MUC-4 dataset, and show that our model performs substantially better than prior work. We also show that our modeling choices contribute to model performance, e.g., by implicitly capturing linguistic knowledge such as recognizing coreferent entity mentions.

arXiv:2008.09090공식 코드 · Papers with Code

TRU-NET: A Deep Learning Approach to High Resolution Prediction of Rainfall

Climate models (CM) are used to evaluate the impact of climate change on the risk of floods and strong precipitation events. However, these numerical simulators have difficulties representing precipitation events accurately, mainly due to limited spatial resolution when simulating multi-scale dynamics in the atmosphere. To improve the prediction of high resolution precipitation we apply a Deep Learning (DL) approach using an input of CM simulations of the model fields (weather variables) that are more predictable than local precipitation. To this end, we present TRU-NET (Temporal Recurrent U-Net), an encoder-decoder model featuring a novel 2D cross attention mechanism between contiguous convolutional-recurrent layers to effectively model multi-scale spatio-temporal weather processes. We use a conditional-continuous loss function to capture the zero-skewed %extreme event patterns of rainfall. Experiments show that our model consistently attains lower RMSE and MAE scores than a DL model prevalent in short term precipitation prediction and improves upon the rainfall predictions of a state-of-the-art dynamical weather model. Moreover, by evaluating the performance of our model under various, training and testing, data formulation strategies, we show that there is enough data for our deep learning approach to output robust, high-quality results across seasons and varying regions.

arXiv:2008.09049

Discovering Useful Sentence Representations from Large Pretrained Language Models

Despite the extensive success of pretrained language models as encoders for building NLP systems, they haven't seen prominence as decoders for sequence generation tasks. We explore the question of whether these models can be adapted to be used as universal decoders. To be considered "universal," a decoder must have an implicit representation for any target sentence $s$, such that it can recover that sentence exactly when conditioned on its representation. For large transformer-based language models trained on vast amounts of English text, we investigate whether such representations can be easily discovered using standard optimization methods. We present and compare three representation injection techniques for transformer-based models and three accompanying methods which map sentences to and from this representation space. Experiments show that not only do representations exist for sentences from a variety of genres. More importantly, without needing complex optimization algorithms, our methods recover these sentences almost perfectly without fine-tuning the underlying language model at all.

arXiv:2008.08213공식 코드 · Papers with Code

DeepHandMesh: A Weakly-supervised Deep Encoder-Decoder Framework for High-fidelity Hand Mesh Modeling

Human hands play a central role in interacting with other people and objects. For realistic replication of such hand motions, high-fidelity hand meshes have to be reconstructed. In this study, we firstly propose DeepHandMesh, a weakly-supervised deep encoder-decoder framework for high-fidelity hand mesh modeling. We design our system to be trained in an end-to-end and weakly-supervised manner; therefore, it does not require groundtruth meshes. Instead, it relies on weaker supervisions such as 3D joint coordinates and multi-view depth maps, which are easier to get than groundtruth meshes and do not dependent on the mesh topology. Although the proposed DeepHandMesh is trained in a weakly-supervised way, it provides significantly more realistic hand mesh than previous fully-supervised hand models. Our newly introduced penetration avoidance loss further improves results by replicating physical interaction between hand parts. Finally, we demonstrate that our system can also be applied successfully to the 3D hand mesh estimation from general images. Our hand model, dataset, and codes are publicly available at https://mks0601.github.io/DeepHandMesh/.

arXiv:2008.07599공식 코드 · Papers with Code

Learning from Irregularly-Sampled Time Series: A Missing Data Perspective

Irregularly-sampled time series occur in many domains including healthcare. They can be challenging to model because they do not naturally yield a fixed-dimensional representation as required by many standard machine learning models. In this paper, we consider irregular sampling from the perspective of missing data. We model observed irregularly-sampled time series data as a sequence of index-value pairs sampled from a continuous but unobserved function. We introduce an encoder-decoder framework for learning from such generic indexed sequences. We propose learning methods for this framework based on variational autoencoders and generative adversarial networks. For continuous irregularly-sampled time series, we introduce continuous convolutional layers that can efficiently interface with existing neural network architectures. Experiments show that our models are able to achieve competitive or better classification results on irregularly-sampled multivariate time series compared to recent RNN models while offering significantly faster training times.

arXiv:2008.07588공식 코드 · Papers with Code

Uncertainty Quantification using Variational Inference for Biomedical Image Segmentation

Deep learning motivated by convolutional neural networks has been highly successful in a range of medical imaging problems like image classification, image segmentation, image synthesis etc. However for validation and interpretability, not only do we need the predictions made by the model but also how confident it is while making those predictions. This is important in safety critical applications for the people to accept it. In this work, we used an encoder decoder architecture based on variational inference techniques for segmenting brain tumour images. We evaluate our work on the publicly available BRATS dataset using Dice Similarity Coefficient (DSC) and Intersection Over Union (IOU) as the evaluation metrics. Our model is able to segment brain tumours while taking into account both aleatoric uncertainty and epistemic uncertainty in a principled bayesian manner.

arXiv:2008.07783

Mesh Guided One-shot Face Reenactment using Graph Convolutional Networks

Face reenactment aims to animate a source face image to a different pose and expression provided by a driving image. Existing approaches are either designed for a specific identity, or suffer from the identity preservation problem in the one-shot or few-shot scenarios. In this paper, we introduce a method for one-shot face reenactment, which uses the reconstructed 3D meshes (i.e., the source mesh and driving mesh) as guidance to learn the optical flow needed for the reenacted face synthesis. Technically, we explicitly exclude the driving face's identity information in the reconstructed driving mesh. In this way, our network can focus on the motion estimation for the source face without the interference of driving face shape. We propose a motion net to learn the face motion, which is an asymmetric autoencoder. The encoder is a graph convolutional network (GCN) that learns a latent motion vector from the meshes, and the decoder serves to produce an optical flow image from the latent vector with CNNs. Compared to previous methods using sparse keypoints to guide the optical flow learning, our motion net learns the optical flow directly from 3D dense meshes, which provide the detailed shape and pose information for the optical flow, so it can achieve more accurate expression and pose on the reenacted face. Extensive experiments show that our method can generate high-quality results and outperforms state-of-the-art methods in both qualitative and quantitative comparisons.

arXiv:2008.07772공식 코드 · Papers with Code

Very Deep Transformers for Neural Machine Translation

We explore the application of very deep Transformer models for Neural Machine Translation (NMT). Using a simple yet effective initialization technique that stabilizes training, we show that it is feasible to build standard Transformer-based models with up to 60 encoder layers and 12 decoder layers. These deep models outperform their baseline 6-layer counterparts by as much as 2.5 BLEU, and achieve new state-of-the-art benchmark results on WMT14 English-French (43.8 BLEU and 46.4 BLEU with back-translation) and WMT14 English-German (30.1 BLEU).The code and trained models will be publicly available at: https://github.com/namisan/exdeep-nmt.

arXiv:2008.07770

Fully automated deep learning based segmentation of normal, infarcted and edema regions from multiple cardiac MRI sequences

Myocardial characterization is essential for patients with myocardial infarction and other myocardial diseases, and the assessment is often performed using cardiac magnetic resonance (CMR) sequences. In this study, we propose a fully automated approach using deep convolutional neural networks (CNN) for cardiac pathology segmentation, including left ventricular (LV) blood pool, right ventricular blood pool, LV normal myocardium, LV myocardial edema (ME) and LV myocardial scars (MS). The input to the network consists of three CMR sequences, namely, late gadolinium enhancement (LGE), T2 and balanced steady state free precession (bSSFP). The proposed approach utilized the data provided by the MyoPS challenge hosted by MICCAI 2020 in conjunction with STACOM. The training set for the CNN model consists of images acquired from 25 cases, and the gold standard labels are provided by trained raters and validated by radiologists. The proposed approach introduces a data augmentation module, linear encoder and decoder module and a network module to increase the number of training samples and improve the prediction accuracy for LV ME and MS. The proposed approach is evaluated by the challenge organizers with a test set including 20 cases and achieves a mean dice score of $46.8\%$ for LV MS and $55.7\%$ for LV ME+MS

arXiv:2008.08062

Compute, Time and Energy Characterization of Encoder-Decoder Networks with Automatic Mixed Precision Training

Deep neural networks have shown great success in many diverse fields. The training of these networks can take significant amounts of time, compute and energy. As datasets get larger and models become more complex, the exploration of model architectures becomes prohibitive. In this paper we examine the compute, energy and time costs of training a UNet based deep neural network for the problem of predicting short term weather forecasts (called precipitation Nowcasting). By leveraging a combination of data distributed and mixed-precision training, we explore the design space for this problem. We also show that larger models with better performance come at a potentially incremental cost if appropriate optimizations are used. We show that it is possible to achieve a significant improvement in training time by leveraging mixed-precision training without sacrificing model performance. Additionally, we find that a 1549% increase in the number of trainable parameters for a network comes at a relatively smaller 63.22% increase in energy usage for a UNet with 4 encoding layers.

arXiv:2008.06266

Optimized Deep Encoder-Decoder Methods for Crack Segmentation

Surface crack segmentation poses a challenging computer vision task as background, shape, colour and size of cracks vary. In this work we propose optimized deep encoder-decoder methods consisting of a combination of techniques which yield an increase in crack segmentation performance. Specifically we propose a decoder-part for an encoder-decoder based deep learning architecture for semantic segmentation and study its components to achieve increased performance. We also examine the use of different encoder strategies and introduce a data augmentation policy to increase the amount of available training data. The performance evaluation of our method is carried out on four publicly available crack segmentation datasets. Additionally, we introduce two techniques into the field of surface crack segmentation, previously not used there: Generating results using test-time-augmentation and performing a statistical result analysis over multiple training runs. The former approach generally yields increased performance results, whereas the latter allows for more reproducible and better representability of a methods results. Using those aforementioned strategies with our proposed encoder-decoder architecture we are able to achieve new state of the art results in all datasets.

arXiv:2008.06687공식 코드 · Papers with Code

Natural Reweighted Wake-Sleep

Helmholtz Machines (HMs) are a class of generative models composed of two Sigmoid Belief Networks (SBNs), acting respectively as an encoder and a decoder. These models are commonly trained using a two-step optimization algorithm called Wake-Sleep (WS) and more recently by improved versions, such as Reweighted Wake-Sleep (RWS) and Bidirectional Helmholtz Machines (BiHM). The locality of the connections in an SBN induces sparsity in the Fisher Information Matrices associated to the probabilistic models, in the form of a finely-grained block-diagonal structure. In this paper we exploit this property to efficiently train SBNs and HMs using the natural gradient. We present a novel algorithm, called Natural Reweighted Wake-Sleep (NRWS), that corresponds to the geometric adaptation of its standard version. In a similar manner, we also introduce Natural Bidirectional Helmholtz Machine (NBiHM). Differently from previous work, we will show how for HMs the natural gradient can be efficiently computed without the need of introducing any approximation in the structure of the Fisher information matrix. The experiments performed on standard datasets from the literature show a consistent improvement of NRWS and NBiHM not only with respect to their non-geometric baselines but also with respect to state-of-the-art training algorithms for HMs. The improvement is quantified both in terms of speed of convergence as well as value of the log-likelihood reached after training.

arXiv:2008.06632

Dehaze-GLCGAN: Unpaired Single Image De-hazing via Adversarial Training

Single image de-hazing is a challenging problem, and it is far from solved. Most current solutions require paired image datasets that include both hazy images and their corresponding haze-free ground-truth images. However, in reality, lighting conditions and other factors can produce a range of haze-free images that can serve as ground truth for a hazy image, and a single ground truth image cannot capture that range. This limits the scalability and practicality of paired image datasets in real-world applications. In this paper, we focus on unpaired single image de-hazing and we do not rely on the ground truth image or physical scattering model. We reduce the image de-hazing problem to an image-to-image translation problem and propose a dehazing Global-Local Cycle-consistent Generative Adversarial Network (Dehaze-GLCGAN). Generator network of Dehaze-GLCGAN combines an encoder-decoder architecture with residual blocks to better recover the haze free scene. We also employ a global-local discriminator structure to deal with spatially varying haze. Through ablation study, we demonstrate the effectiveness of different factors in the performance of the proposed network. Our extensive experiments over three benchmark datasets show that our network outperforms previous work in terms of PSNR and SSIM while being trained on smaller amount of data compared to other methods.

arXiv:2008.06908공식 코드 · Papers with Code

Visually Aware Skip-Gram for Image Based Recommendations

The visual appearance of a product significantly influences purchase decisions on e-commerce websites. We propose a novel framework VASG (Visually Aware Skip-Gram) for learning user and product representations in a common latent space using product image features. Our model is an amalgamation of the Skip-Gram architecture and a deep neural network based Decoder. Here the Skip-Gram attempts to capture user preference by optimizing user-product co-occurrence in a Heterogeneous Information Network while the Decoder simultaneously learns a mapping to transform product image features to the Skip-Gram embedding space. This architecture is jointly optimized in an end-to-end, multitask fashion. The proposed framework enables us to make personalized recommendations for cold-start products which have no purchase history. Experiments conducted on large real-world datasets show that the learned embeddings can generate effective recommendations using nearest neighbour searches.