Triple extraction is an essential task in information extraction for natural language processing and knowledge graph construction. In this paper, we revisit the end-to-end triple extraction task for sequence generation. Since generative triple extraction may struggle to capture long-term dependencies and generate unfaithful triples, we introduce a novel model, contrastive triple extraction with a generative transformer. Specifically, we introduce a single shared transformer module for encoder-decoder-based generation. To generate faithful results, we propose a novel triplet contrastive training object. Moreover, we introduce two mechanisms to further improve model performance (i.e., batch-wise dynamic attention-masking and triple-wise calibration). Experimental results on three datasets (i.e., NYT, WebNLG, and MIE) show that our approach achieves better performance than that of baselines.
This paper presents a novel method to generate answers for non-extraction machine reading comprehension (MRC) tasks whose answers cannot be simply extracted as one span from the given passages. Using a pointer network-style extractive decoder for such type of MRC may result in unsatisfactory performance when the ground-truth answers are given by human annotators or highly re-paraphrased from parts of the passages. On the other hand, using generative decoder cannot well guarantee the resulted answers with well-formed syntax and semantics when encountering long sentences. Therefore, to alleviate the obvious drawbacks of both sides, we propose an answer making-up method from extracted multi-spans that are learned by our model as highly confident $n$-gram candidates in the given passage. That is, the returned answers are composed of discontinuous multi-spans but not just one consecutive span in the given passages anymore. The proposed method is simple but effective: empirical experiments on MS MARCO show that the proposed method has a better performance on accurately generating long answers, and substantially outperforms two competitive typical one-span and Seq2Seq baseline decoders.
We propose a self-supervised method for partial point set registration. While recent proposed learning-based methods have achieved impressive registration performance on the full shape observations, these methods mostly suffer from performance degradation when dealing with partial shapes. To bridge the performance gaps between partial point set registration with full point set registration, we proposed to incorporate a shape completion network to benefit the registration process. To achieve this, we design a latent code for each pair of shapes, which can be regarded as a geometric encoding of the target shape. By doing so, our model does need an explicit feature embedding network to learn the feature encodings. More importantly, both our shape completion network and the point set registration network take the shared latent codes as input, which are optimized along with the parameters of two decoder networks in the training process. Therefore, the point set registration process can thus benefit from the joint optimization process of latent codes, which are enforced to represent the information of full shape instead of partial ones. In the inference stage, we fix the network parameter and optimize the latent codes to get the optimal shape completion and registration results. Our proposed method is pure unsupervised and does not need any ground truth supervision. Experiments on the ModelNet40 dataset demonstrate the effectiveness of our model for partial point set registration.
The autoencoder concept has fostered the reinterpretation and the design of modern communication systems. It consists of an encoder, a channel, and a decoder block which modify their internal neural structure in an end-to-end learning fashion. However, the current approach to train an autoencoder relies on the use of the cross-entropy loss function. This approach can be prone to overfitting issues and often fails to learn an optimal system and signal representation (code). In addition, less is known about the autoencoder ability to design channel capacity-approaching codes, i.e., codes that maximize the input-output information under a certain power constraint. The task being even more formidable for an unknown channel for which the capacity is unknown and therefore it has to be learnt. In this paper, we address the challenge of designing capacity-approaching codes by incorporating the presence of the communication channel into a novel loss function for the autoencoder training. In particular, we exploit the mutual information between the transmitted and received signals as a regularization term in the cross-entropy loss function, with the aim of controlling the amount of information stored. By jointly maximizing the mutual information and minimizing the cross-entropy, we propose a methodology that a) computes an estimate of the channel capacity and b) constructs an optimal coded signal approaching it. Several simulation results offer evidence of the potentiality of the proposed method.
Software-defined networking (SDN) and the network function virtualization (NFV) led to great developments in software based control technology by decreasing expenditures. Service function chaining (SFC) is an important technology to find efficient paths in network servers to process all of the requested virtualized network functions (VNF). However, SFC is challenging since it has to maintain high Quality of Service (QoS) even for complicated situations. Although some works have been conducted for such tasks with high-level intelligent models like deep neural networks (DNNs), those approaches are not efficient in utilizing the topology information of networks and cannot be applied to networks with dynamically changing topology since their models assume that the topology is fixed. In this paper, we propose a new neural network architecture for SFC, which is based on graph neural network (GNN) considering the graph-structured properties of network topology. The proposed SFC model consists of an encoder and a decoder, where the encoder finds the representation of the network topology, and then the decoder estimates probabilities of neighborhood nodes and their probabilities to process a VNF. In the experiments, our proposed architecture outperformed previous performances of DNN based baseline model. Moreover, the GNN based model can be applied to a new network topology without re-designing and re-training.
We propose a novel method to sparsify attention in the Transformer model by learning to select the most-informative token representations during the training process, thus focusing on the task-specific parts of an input. A reduction of quadratic time and memory complexity to sublinear was achieved due to a robust trainable top-$k$ operator. Our experiments on a challenging long document summarization task show that even our simple baseline performs comparably to the current SOTA, and with trainable pooling, we can retain its top quality, while being $1.8\times$ faster during training, $4.5\times$ faster during inference, and up to $13\times$ more computationally efficient in the decoder.
Detection of burn marks due to wildfires in inaccessible rain forests is important for various disaster management and ecological studies. The fragmented nature of arable landscapes and diverse cropping patterns often thwart the precise mapping of burn scars. Recent advances in remote-sensing and availability of multimodal data offer a viable solution to this mapping problem. However, the task to segment burn marks is difficult because of its indistinguishably with similar looking land patterns, severe fragmented nature of burn marks and partially labelled noisy datasets. In this work we present AmazonNET -- a convolutional based network that allows extracting of burn patters from multimodal remote sensing images. The network consists of UNet: a well-known encoder decoder type of architecture with skip connections commonly used in biomedical segmentation. The proposed framework utilises stacked RGB-NIR channels to segment burn scars from the pastures by training on a new weakly labelled noisy dataset from Amazonia. Our model illustrates superior performance by correctly identifying partially labelled burn scars and rejecting incorrectly labelled samples, demonstrating our approach as one of the first to effectively utilise deep learning based segmentation models in multimodal burn scar identification.
Annotating seismic data is expensive, laborious and subjective due to the number of years required for seismic interpreters to attain proficiency in interpretation. In this paper, we develop a framework to automate annotating pixels of a seismic image to delineate geological structural elements given image-level labels assigned to each image. Our framework factorizes the latent space of a deep encoder-decoder network by projecting the latent space to learned sub-spaces. Using constraints in the pixel space, the seismic image is further factorized to reveal confidence values on pixels associated with the geological element of interest. Details of the annotated image are provided for analysis and qualitative comparison is made with similar frameworks.
State-of-the-art models for high-resolution image generation, such as BigGAN and VQVAE-2, require an incredible amount of compute resources and/or time (512 TPU-v3 cores) to train, putting them out of reach for the larger research community. On the other hand, GAN-based image super-resolution models, such as ESRGAN, can not only upscale images to high dimensions, but also are efficient to train. In this paper, we present not-so-big-GAN (nsb-GAN), a simple yet cost-effective two-step training framework for deep generative models (DGMs) of high-dimensional natural images. First, we generate images in low-frequency bands by training a sampler in the wavelet domain. Then, we super-resolve these images from the wavelet domain back to the pixel-space with our novel wavelet super-resolution decoder network. Wavelet-based down-sampling method preserves more structural information than pixel-based methods, leading to significantly better generative quality of the low-resolution sampler (e.g., 64x64). Since the sampler and decoder can be trained in parallel and operate on much lower dimensional spaces than end-to-end models, the training cost is substantially reduced. On ImageNet 512x512, our model achieves a Fr\'echet Inception Distance (FID) of 10.59 -- beating the baseline BigGAN model -- at half the compute (256 TPU-v3 cores).
In general, it is straightforward to render an offline handwriting image from an online handwriting pattern. However, it is challenging to reconstruct an online handwriting pattern given an offline handwriting image, especially for multiple-stroke character as Japanese kanji. The multiple-stroke character requires not only point coordinates but also stroke orders whose difficulty is exponential growth by the number of strokes. Besides, several crossed and touch points might increase the difficulty of the recovered task. We propose a deep neural network-based method to solve the recovered task using a large online handwriting database. Our proposed model has two main components: Convolutional Neural Network-based encoder and Long Short-Term Memory Network-based decoder with an attention layer. The encoder focuses on feature extraction while the decoder refers to the extracted features and generates the time-sequences of coordinates. We also demonstrate the effect of the attention layer to guide the decoder during the reconstruction. We evaluate the performance of the proposed method by both visual verification and handwritten character recognition. Although the visual verification reveals some problems, the recognition experiments demonstrate the effect of trajectory recovery in improving the accuracy of offline handwritten character recognition when online recognition for the recovered trajectories are combined.
Facial attribute editing has mainly two objectives: 1) translating image from a source domain to a target one, and 2) only changing the facial regions related to a target attribute and preserving the attribute-excluding details. In this work, we propose a Multi-attention U-Net-based Generative Adversarial Network (MU-GAN). First, we replace a classic convolutional encoder-decoder with a symmetric U-Net-like structure in a generator, and then apply an additive attention mechanism to build attention-based U-Net connections for adaptively transferring encoder representations to complement a decoder with attribute-excluding detail and enhance attribute editing ability. Second, a self-attention mechanism is incorporated into convolutional layers for modeling long-range and multi-level dependencies across image regions. experimental results indicate that our method is capable of balancing attribute editing ability and details preservation ability, and can decouple the correlation among attributes. It outperforms the state-of-the-art methods in terms of attribute manipulation accuracy and image quality.
We present an approach called VisCode for embedding information into visualization images. This technology can implicitly embed data information specified by the user into a visualization while ensuring that the encoded visualization image is not distorted. The VisCode framework is based on a deep neural network. We propose to use visualization images and QR codes data as training data and design a robust deep encoder-decoder network. The designed model considers the salient features of visualization images to reduce the explicit visual loss caused by encoding. To further support large-scale encoding and decoding, we consider the characteristics of information visualization and propose a saliency-based QR code layout algorithm. We present a variety of practical applications of VisCode in the context of information visualization and conduct a comprehensive evaluation of the perceptual quality of encoding, decoding success rate, anti-attack capability, time performance, etc. The evaluation results demonstrate the effectiveness of VisCode.
In this paper, we propose a deep neural network based model to predict the time evolution of field values in transient electrodynamics. The key component of our model is a recurrent neural network, which learns representations of long-term spatial-temporal dependencies in the sequence of its input data. We develop an encoder-recurrent-decoder architecture, which is trained with finite difference time domain simulations of plane wave scattering from distributed, perfect electric conducting objects. We demonstrate that, the trained network can emulate a transient electrodynamics problem with more than 17 times speed-up in simulation time compared to traditional finite difference time domain solvers.
Reverse-engineering bar charts extracts textual and numeric information from the visual representations of bar charts to support application scenarios that require the underlying information. In this paper, we propose a neural network-based method for reverse-engineering bar charts. We adopt a neural network-based object detection model to simultaneously localize and classify textual information. This approach improves the efficiency of textual information extraction. We design an encoder-decoder framework that integrates convolutional and recurrent neural networks to extract numeric information. We further introduce an attention mechanism into the framework to achieve high accuracy and robustness. Synthetic and real-world datasets are used to evaluate the effectiveness of the method. To the best of our knowledge, this work takes the lead in constructing a complete neural network-based method of reverse-engineering bar charts.
We propose the first stochastic framework to employ uncertainty for RGB-D saliency detection by learning from the data labeling process. Existing RGB-D saliency detection models treat this task as a point estimation problem by predicting a single saliency map following a deterministic learning pipeline. We argue that, however, the deterministic solution is relatively ill-posed. Inspired by the saliency data labeling process, we propose a generative architecture to achieve probabilistic RGB-D saliency detection which utilizes a latent variable to model the labeling variations. Our framework includes two main models: 1) a generator model, which maps the input image and latent variable to stochastic saliency prediction, and 2) an inference model, which gradually updates the latent variable by sampling it from the true or approximate posterior distribution. The generator model is an encoder-decoder saliency network. To infer the latent variable, we introduce two different solutions: i) a Conditional Variational Auto-encoder with an extra encoder to approximate the posterior distribution of the latent variable; and ii) an Alternating Back-Propagation technique, which directly samples the latent variable from the true posterior distribution. Qualitative and quantitative results on six challenging RGB-D benchmark datasets show our approach's superior performance in learning the distribution of saliency maps. The source code is publicly available via our project page: https://github.com/JingZhang617/UCNet.
Recent advances in natural language generation have introduced powerful language models with high-quality output text. However, this raises concerns about the potential misuse of such models for malicious purposes. In this paper, we study natural language watermarking as a defense to help better mark and trace the provenance of text. We introduce the Adversarial Watermarking Transformer (AWT) with a jointly trained encoder-decoder and adversarial training that, given an input text and a binary message, generates an output text that is unobtrusively encoded with the given message. We further study different training and inference strategies to achieve minimal changes to the semantics and correctness of the input text. AWT is the first end-to-end model to hide data in text by automatically learning -- without ground truth -- word substitutions along with their locations in order to encode the message. We empirically show that our model is effective in largely preserving text utility and decoding the watermark while hiding its presence against adversaries. Additionally, we demonstrate that our method is robust against a range of attacks.
We present a simple yet effective general-purpose framework for modeling 3D shapes by leveraging recent advances in 2D image generation using CNNs. Using just a single depth image of the object, we can output a dense multi-view depth map representation of 3D objects. Our simple encoder-decoder framework, comprised of a novel identity encoder and class-conditional viewpoint generator, generates 3D consistent depth maps. Our experimental results demonstrate the two-fold advantage of our approach. First, we can directly borrow architectures that work well in the 2D image domain to 3D. Second, we can effectively generate high-resolution 3D shapes with low computational memory. Our quantitative evaluations show that our method is superior to existing depth map methods for reconstructing and synthesizing 3D objects and is competitive with other representations, such as point clouds, voxel grids, and implicit functions.
Style transfer is the process of changing the style of an image, video, audio clip or musical piece so as to match the style of a given example. Even though the task has interesting practical applications within the music industry, it has so far received little attention from the audio and music processing community. In this paper, we present Groove2Groove, a one-shot style transfer method for symbolic music, focusing on the case of accompaniment styles in popular music and jazz. We propose an encoder-decoder neural network for the task, along with a synthetic data generation scheme to supply it with parallel training examples. This synthetic parallel data allows us to tackle the style transfer problem using end-to-end supervised learning, employing powerful techniques used in natural language processing. We experimentally demonstrate the performance of the model on style transfer using existing and newly proposed metrics, and also explore the possibility of style interpolation.
In many ultrasonic imaging systems, data acquisition and image formation are performed on separate computing devices. Data transmission is becoming a bottleneck, thus, efficient data compression is essential. Compression rates can be improved by considering the fact that many image formation methods rely on approximations of wave-matter interactions, and only use the corresponding part of the data. Tailored data compression could exploit this, but extracting the useful part of the data efficiently is not always trivial. In this work, we tackle this problem using deep neural networks, optimized to preserve the image quality of a particular image formation method. The Delay-And-Sum (DAS) algorithm is examined which is used in reflectivity-based ultrasonic imaging. We propose a novel encoder-decoder architecture with vector quantization and formulate image formation as a network layer for end-to-end training. Experiments demonstrate that our proposed data compression tailored for a specific image formation method obtains significantly better results as opposed to compression agnostic to subsequent imaging. We maintain high image quality at much higher compression rates than the theoretical lossless compression rate derived from the rank of the linear imaging operator. This demonstrates the great potential of deep ultrasonic data compression tailored for a specific image formation method.
Multimodal machine translation (MMT), which mainly focuses on enhancing text-only translation with visual features, has attracted considerable attention from both computer vision and natural language processing communities. Most current MMT models resort to attention mechanism, global context modeling or multimodal joint representation learning to utilize visual features. However, the attention mechanism lacks sufficient semantic interactions between modalities while the other two provide fixed visual context, which is unsuitable for modeling the observed variability when generating translation. To address the above issues, in this paper, we propose a novel Dynamic Context-guided Capsule Network (DCCN) for MMT. Specifically, at each timestep of decoding, we first employ the conventional source-target attention to produce a timestep-specific source-side context vector. Next, DCCN takes this vector as input and uses it to guide the iterative extraction of related visual features via a context-guided dynamic routing mechanism. Particularly, we represent the input image with global and regional visual features, we introduce two parallel DCCNs to model multimodal context vectors with visual features at different granularities. Finally, we obtain two multimodal context vectors, which are fused and incorporated into the decoder for the prediction of the target word. Experimental results on the Multi30K dataset of English-to-German and English-to-French translation demonstrate the superiority of DCCN. Our code is available on https://github.com/DeepLearnXMU/MM-DCCN.