Judging the veracity of a sentence making one or more claims is an important and challenging problem with many dimensions. The recent FEVER task asked participants to classify input sentences as either SUPPORTED, REFUTED or NotEnoughInfo using Wikipedia as a source of true facts. SURFACE does this task and explains its decision through a selection of sentences from the trusted source. Our multi-task neural approach uses semantic lexical frames from FrameNet to jointly (i) find relevant evidential sentences in the trusted source and (ii) use them to classify the input sentence's veracity. An evaluation of our efficient three-parameter model on the FEVER dataset showed an improvement of 90% over the state-of-the-art baseline on retrieving relevant sentences and a 70% relative improvement in classification.
Segmentation of the left atrium (LA) is crucial for assessing its anatomy in both pre-operative atrial fibrillation (AF) ablation planning and post-operative follow-up studies. In this paper, we present a fully automated framework for left atrial segmentation in gadolinium-enhanced magnetic resonance images (GE-MRI) based on deep learning. We propose a fully convolutional neural network and explore the benefits of multi-task learning for performing both atrial segmentation and pre/post ablation classification. Our results show that, by sharing features between related tasks, the network can gain additional anatomical information and achieve more accurate atrial segmentation, leading to a mean Dice score of 0.901 on a test set of 20 3D MRI images. Code of our proposed algorithm is available at https://github.com/cherise215/atria_segmentation_2018/.
Although the convolutional neural networks (CNNs) have become popular for various image processing and computer vision task recently, it remains a challenging problem to reduce the storage cost of the parameters for resource-limited platforms. In the previous studies, tensor decomposition (TD) has achieved promising compression performance by embedding the kernel of a convolutional layer into a low-rank subspace. However the employment of TD is naively on the kernel or its specified variants. Unlike the conventional approaches, this paper shows that the kernel can be embedded into more general or even random low-rank subspaces. We demonstrate this by compressing the convolutional layers via randomly-shuffled tensor decomposition (RsTD) for a standard classification task using CIFAR-10. In addition, we analyze how the spatial similarity of the training data influences the low-rank structure of the kernels. The experimental results show that the CNN can be significantly compressed even if the kernels are randomly shuffled. Furthermore, the RsTD-based method yields more stable classification accuracy than the conventional TD-based methods in a large range of compression ratios.
This study aims to automatically diagnose thoracic diseases depicted on the chest x-ray (CXR) images using deep convolutional neural networks. The existing methods generally used the entire CXR images for training purposes, but this strategy may suffer from two drawbacks. First, potential misalignment or the existence of irrelevant objects in the entire CXR images may cause unnecessary noise and thus limit the network performance. Second, the relatively low image resolution caused by the resizing operation, which is a common preprocessing procedure for training neural networks, may lead to the loss of image details, making it difficult to detect pathologies with small lesion regions. To address these issues, we present a novel method termed as segmentation-based deep fusion network (SDFN), which leverages the domain knowledge and the higherresolution information of local lung regions. Specifically, the local lung regions were identified and cropped by the Lung Region Generator (LRG). Two CNN-based classification models were then used as feature extractors to obtain the discriminative features of the entire CXR images and the cropped lung region images. Lastly, the obtained features were fused by the feature fusion module for disease classification. Evaluated by the NIH benchmark split on the Chest X-ray 14 Dataset, our experimental result demonstrated that the developed method achieved more accurate disease classification compared with the available approaches via the receiver operating characteristic (ROC) analyses. It was also found that the SDFN could localize the lesion regions more precisely as compared to the traditional method.
Recurrent Neural Network (RNN) has been successfully applied in many sequence learning problems. Such as handwriting recognition, image description, natural language processing and video motion analysis. After years of development, researchers have improved the internal structure of the RNN and introduced many variants. Among others, Gated Recurrent Unit (GRU) is one of the most widely used RNN model. However, GRU lacks the capability of adaptively paying attention to certain regions or locations, so that it may cause information redundancy or loss during leaning. In this paper, we propose a RNN model, called Recurrent Attention Unit (RAU), which seamlessly integrates the attention mechanism into the interior of GRU by adding an attention gate. The attention gate can enhance GRU's ability to remember long-term memory and help memory cells quickly discard unimportant content. RAU is capable of extracting information from the sequential data by adaptively selecting a sequence of regions or locations and pay more attention to the selected regions during learning. Extensive experiments on image classification, sentiment classification and language modeling show that RAU consistently outperforms GRU and other baseline methods.
The antinuclear antibody detection with human epithelial cells is a popular approach for autoimmune diseases diagnosis. The manual evaluation demands time, effort and capital, and automation in screening can greatly aid the physicians in these respects. In this work, we employ simple, efficient and visually more interpretable, class-specific features which defined based on the visual characteristics of each class. We believe that defining features with a good visual interpretation, is indeed important in a scenario, where such an approach is used in an interactive CAD system for pathologists. Considering that problem consists of few classes, and our rather simplistic feature definitions, frameworks can be structured as hierarchies of various binary classifiers. These variants include frameworks which are earlier explored and some which are not explored for this task. We perform various experiments which include traditional texture features and demonstrate the effectiveness of class-specific features in various frameworks. We make insightful comparisons between different types of classification frameworks given their silent aspects and pros and cons over each other. We also demonstrate an experiment with only intermediates samples for testing. The proposed work yields encouraging results with respect to the state-of-the-art and highlights the role of class-specific features in different classification frameworks.
Deep learning is popular as an end-to-end framework extracting the prominent features and performing the classification also. In this paper, we extensively investigate deep networks as an alternate to feature encoding technique of low level descriptors for emotion recognition on the benchmark EmoDB dataset. Fusion performance with such obtained encoded features with other available features is also investigated. Highest performance to date in the literature is observed.
Convolutional neural networks (CNNs) attained a good performance in hyperspectral sensing image (HSI) classification, but CNNs consider spectra as orderless vectors. Therefore, considering the spectra as sequences, recurrent neural networks (RNNs) have been applied in HSI classification, for RNNs is skilled at dealing with sequential data. However, for a long-sequence task, RNNs is difficult for training and not as effective as we expected. Besides, spatial contextual features are not considered in RNNs. In this study, we propose a Shorten Spatial-spectral RNN with Parallel-GRU (St-SS-pGRU) for HSI classification. A shorten RNN is more efficient and easier for training than band-by-band RNN. By combining converlusion layer, the St-SSpGRU model considers not only spectral but also spatial feature, which results in a better performance. An architecture named parallel-GRU is also proposed and applied in St-SS-pGRU. With this architecture, the model gets a better performance and is more robust.
Class imbalance is a pervasive issue among classification models including deep learning, whose capacity to extract task-specific features is affected in imbalanced settings. However, the challenges of handling imbalance among a large number of classes, commonly addressed by deep learning, have not received a significant amount of attention in previous studies. In this paper, we propose an extension of the deep over-sampling framework, to exploit automatically-generated abstract-labels, i.e., a type of side-information used in weak-label learning, to enhance deep representation learning against class imbalance. We attempt to exploit the labels to guide the deep representation of instances towards different subspaces, to induce a soft-separation of inherent subtasks of the classification problem. Our empirical study shows that the proposed framework achieves a substantial improvement on image classification benchmarks with imbalanced among large and small numbers of classes.
A number of studies on lung nodule classification lack clinical/biological interpretations of the features extracted by convolutional neural network (CNN). The methods like class activation mapping (CAM) and gradient-based CAM (Grad-CAM) are tailored for interpreting localization and classification tasks while they ignored fine-grained features. Therefore, CAM and Grad-CAM cannot provide optimal interpretation for lung nodule categorization task in low-dose CT images, in that fine-grained pathological clues like discrete and irregular shape and margins of nodules are capable of enhancing sensitivity and specificity of nodule classification with regards to CNN. In this paper, we first develop a soft activation mapping (SAM) to enable fine-grained lung nodule shape \& margin (LNSM) feature analysis with a CNN so that it can access rich discrete features. Secondly, by combining high-level convolutional features with SAM, we further propose a high-level feature enhancement scheme (HESAM) to localize LNSM features. Experiments on the LIDC-IDRI dataset indicate that 1) SAM captures more fine-grained and discrete attention regions than existing methods, 2) HESAM localizes more accurately on LNSM features and obtains the state-of-the-art predictive performance, reducing the false positive rate, and 3) we design and conduct a visually matching experiment which incorporates radiologists study to increase the confidence level of applying our method to clinical diagnosis.
We introduce a bottleneck method for learning data representations based on information deficiency, rather than the more traditional information sufficiency. A variational upper bound allows us to implement this method efficiently. The bound itself is bounded above by the variational information bottleneck objective, and the two methods coincide in the regime of single-shot Monte Carlo approximations. The notion of deficiency provides a principled way of approximating complicated channels by relatively simpler ones. We show that the deficiency of one channel with respect to another has an operational interpretation in terms of the optimal risk gap of decision problems, capturing classification as a special case. Experiments demonstrate that the deficiency bottleneck can provide advantages in terms of minimal sufficiency as measured by information bottleneck curves, while retaining robust test performance in classification tasks.
Extreme multi-label classification (XMLC) is a problem of tagging an instance with a small subset of relevant labels chosen from an extremely large pool of possible labels. Large label spaces can be efficiently handled by organizing labels as a tree, like in the hierarchical softmax (HSM) approach commonly used for multi-class problems. In this paper, we investigate probabilistic label trees (PLTs) that have been recently devised for tackling XMLC problems. We show that PLTs are a no-regret multi-label generalization of HSM when precision@k is used as a model evaluation metric. Critically, we prove that pick-one-label heuristic - a reduction technique from multi-label to multi-class that is routinely used along with HSM - is not consistent in general. We also show that our implementation of PLTs, referred to as extremeText (XT), obtains significantly better results than HSM with the pick-one-label heuristic and XML-CNN, a deep network specifically designed for XMLC problems. Moreover, XT is competitive to many state-of-the-art approaches in terms of statistical performance, model size and prediction time which makes it amenable to deploy in an online system.
GANs involve training two networks in an adversarial game, where each network's task depends on its adversary. Recently, several works have framed GAN training as an online or continual learning problem. We focus on the discriminator, which must perform classification under an (adversarially) shifting data distribution. When trained on sequential tasks, neural networks exhibit \emph{forgetting}. For GANs, discriminator forgetting leads to training instability. To counter forgetting, we encourage the discriminator to maintain useful representations by adding a self-supervision. Conditional GANs have a similar effect using labels. However, our self-supervised GAN does not require labels, and closes the performance gap between conditional and unconditional models. We show that, in doing so, the self-supervised discriminator learns better representations than regular GANs.
Adversarial sample attacks perturb benign inputs to induce DNN misbehaviors. Recent research has demonstrated the widespread presence and the devastating consequences of such attacks. Existing defense techniques either assume prior knowledge of specific attacks or may not work well on complex models due to their underlying assumptions. We argue that adversarial sample attacks are deeply entangled with interpretability of DNN models: while classification results on benign inputs can be reasoned based on the human perceptible features/attributes, results on adversarial samples can hardly be explained. Therefore, we propose a novel adversarial sample detection technique for face recognition models, based on interpretability. It features a novel bi-directional correspondence inference between attributes and internal neurons to identify neurons critical for individual attributes. The activation values of critical neurons are enhanced to amplify the reasoning part of the computation and the values of other neurons are weakened to suppress the uninterpretable part. The classification results after such transformation are compared with those of the original model to detect adversaries. Results show that our technique can achieve 94% detection accuracy for 7 different kinds of attacks with 9.91% false positives on benign inputs. In contrast, a state-of-the-art feature squeezing technique can only achieve 55% accuracy with 23.3% false positives.
This paper proposes a framework based on deep convolutional neural networks (CNNs) for automatic heart sound classification using short-segments of individual heart beats. We design a 1D-CNN that directly learns features from raw heart-sound signals, and a 2D-CNN that takes inputs of two- dimensional time-frequency feature maps based on Mel-frequency cepstral coefficients (MFCC). We further develop a time-frequency CNN ensemble (TF-ECNN) combining the 1D-CNN and 2D-CNN based on score-level fusion of the class probabilities. On the large PhysioNet CinC challenge 2016 database, the proposed CNN models outperformed traditional classifiers based on support vector machine and hidden Markov models with various hand-crafted time- and frequency-domain features. Best classification scores with 89.22% accuracy and 89.94% sensitivity were achieved by the ECNN, and 91.55% specificity and 88.82% modified accuracy by the 2D-CNN alone on the test set.
We study consistency properties of machine learning methods based on minimizing convex surrogates. We extend the recent framework of Osokin et al. (2017) for the quantitative analysis of consistency properties to the case of inconsistent surrogates. Our key technical contribution consists in a new lower bound on the calibration function for the quadratic surrogate, which is non-trivial (not always zero) for inconsistent cases. The new bound allows to quantify the level of inconsistency of the setting and shows how learning with inconsistent surrogates can have guarantees on sample complexity and optimization difficulty. We apply our theory to two concrete cases: multi-class classification with the tree-structured loss and ranking with the mean average precision loss. The results show the approximation-computation trade-offs caused by inconsistent surrogates and their potential benefits.
This paper presents a new technique for training networks for low-precision communication. Targeting minimal communication between nodes not only enables the use of emerging spiking neuromorphic platforms, but may additionally streamline processing conventionally. Low-power and embedded neuromorphic processors potentially offer dramatic performance-per-Watt improvements over traditional von Neumann processors, however programming these brain-inspired platforms generally requires platform-specific expertise which limits their applicability. To date, the majority of artificial neural networks have not operated using discrete spike-like communication. We present a method for training deep spiking neural networks using an iterative modification of the backpropagation optimization algorithm. This method, which we call Whetstone, effectively and reliably configures a network for a spiking hardware target with little, if any, loss in performance. Whetstone networks use single time step binary communication and do not require a rate code or other spike-based coding scheme, thus producing networks comparable in timing and size to conventional ANNs, albeit with binarized communication. We demonstrate Whetstone on a number of image classification networks, describing how the sharpening process interacts with different training optimizers and changes the distribution of activity within the network. We further note that Whetstone is compatible with several non-classification neural network applications, such as autoencoders and semantic segmentation. Whetstone is widely extendable and currently implemented using custom activation functions within the Keras wrapper to the popular TensorFlow machine learning framework.
We introduce an extension of the multi-instance learning problem where examples are organized as nested bags of instances (e.g., a document could be represented as a bag of sentences, which in turn are bags of words). This framework can be useful in various scenarios, such as text and image classification, but also supervised learning over graphs. As a further advantage, multi-multi instance learning enables a particular way of interpreting predictions and the decision function. Our approach is based on a special neural network layer, called bag-layer, whose units aggregate bags of inputs of arbitrary size. We prove theoretically that the associated class of functions contains all Boolean functions over sets of sets of instances and we provide empirical evidence that functions of this kind can be actually learned on semi-synthetic datasets. We finally present experiments on text classification, on citation graphs, and social graph data, which show that our model obtains competitive results with respect to accuracy when compared to other approaches such as convolutional networks on graphs, while at the same time it supports a general approach to interpret the learnt model, as well as explain individual predictions.
Approximate Bayesian computation (ABC) methods can be used to sample from posterior distributions when the likelihood function is unavailable or intractable, as is often the case in biological systems. ABC methods suffer from inefficient particle proposals in high dimensions, and subjectivity in the choice of summary statistics, discrepancy measure, and error tolerance. Sequential Monte Carlo (SMC) methods have been combined with ABC to improve the efficiency of particle proposals, but suffer from subjectivity and require many simulations from the likelihood function. Likelihood-Free Inference by Ratio Estimation (LFIRE) leverages classification to estimate the posterior density directly but does not explore the parameter space efficiently. This work proposes a classification approach that approximates population Monte Carlo (PMC), where model class probabilities from classification are used to update particle weights. This approach, called Classification-PMC, blends adaptive proposals and classification, efficiently producing samples from the posterior without subjectivity. We show through a simulation study that Classification-PMC outperforms two state-of-the-art methods: ratio estimation and SMC ABC when it is computationally difficult to simulate from the likelihood.
Convolutional Neural Networks (CNNs) are a cornerstone of the Deep Learning toolbox and have led to many breakthroughs in Artificial Intelligence. These networks have mostly been developed for regular Euclidean domains such as those supporting images, audio, or video. Because of their success, CNN-based methods are becoming increasingly popular in Cosmology. Cosmological data often comes as spherical maps, which make the use of the traditional CNNs more complicated. The commonly used pixelization scheme for spherical maps is the Hierarchical Equal Area isoLatitude Pixelisation (HEALPix). We present a spherical CNN for analysis of full and partial HEALPix maps, which we call DeepSphere. The spherical CNN is constructed by representing the sphere as a graph. Graphs are versatile data structures that can act as a discrete representation of a continuous manifold. Using the graph-based representation, we define many of the standard CNN operations, such as convolution and pooling. With filters restricted to being radial, our convolutions are equivariant to rotation on the sphere, and DeepSphere can be made invariant or equivariant to rotation. This way, DeepSphere is a special case of a graph CNN, tailored to the HEALPix sampling of the sphere. This approach is computationally more efficient than using spherical harmonics to perform convolutions. We demonstrate the method on a classification problem of weak lensing mass maps from two cosmological models and compare the performance of the CNN with that of two baseline classifiers. The results show that the performance of DeepSphere is always superior or equal to both of these baselines. For high noise levels and for data covering only a smaller fraction of the sphere, DeepSphere achieves typically 10% better classification accuracy than those baselines. Finally, we show how learned filters can be visualized to introspect the neural network.