Deep hashing methods have received much attention recently, which achieve promising results by taking advantage of the strong representation power of deep networks. However, most existing deep hashing methods learn a whole set of hashing functions independently, while ignore the correlations between different hashing functions that can promote the retrieval accuracy greatly. Inspired by the sequential decision ability of deep reinforcement learning, we propose a new Deep Reinforcement Learning approach for Image Hashing (DRLIH). Our proposed DRLIH approach models the hashing learning problem as a sequential decision process, which learns each hashing function by correcting the errors imposed by previous ones and promotes retrieval accuracy. To the best of our knowledge, this is the first work to address hashing problem from deep reinforcement learning perspective. The main contributions of our proposed DRLIH approach can be summarized as follows: (1) We propose a deep reinforcement learning hashing network. In the proposed network, we utilize recurrent neural network (RNN) as agents to model the hashing functions, which take actions of projecting images into binary codes sequentially, so that the current hashing function learning can take previous hashing functions’ error into account. (2) We propose a sequential learning strategy based on proposed DRLIH. We define the state as a tuple of internal features of RNN’s hidden layers and image features, which can reflect history decisions made by the agents. We also propose an action group method to enhance the correlation of hash functions in the same group. Experiments on three widely-used datasets demonstrate the effectiveness of our proposed DRLIH approach.
http://arxiv.org/abs/1802.02904
A preferential domain is a collection of sets of preferences which are linear orders over a set of alternatives. These domains have been studied extensively in social choice theory due to both its practical importance and theoretical elegance. Examples of some extensively studied preferential domains include single peaked, single crossing, Euclidean, etc. In this paper, we study the sample complexity of testing whether a given preference profile is close to some specific domain. We consider two notions of closeness: (a) closeness via preferences, and (b) closeness via alternatives. We further explore the effect of assuming that the {\em outlier} preferences/alternatives to be random (instead of arbitrary) on the sample complexity of the testing problem. In most cases, we show that the above testing problem can be solved with high probability for all commonly used domains by observing only a small number of samples (independent of the number of preferences, $n$, and often the number of alternatives, $m$). In the remaining few cases, we prove either impossibility results or $\Omega(n)$ lower bound on the sample complexity. We complement our theoretical findings with extensive simulations to figure out the actual constant factors of our asymptotic sample complexity bounds.
http://arxiv.org/abs/1902.08930
We propose a novel iterative channel estimation (ICE) algorithm that essentially removes the critical known noisy channel assumption for universal discrete denoising problem. Our algorithm is based on Neural DUDE (N-DUDE), a recently proposed neural network-based discrete denoiser, and it estimates the channel transition matrix as well as the neural network parameters in an alternating manner until convergence. While we do not make any probabilistic assumption on the underlying clean data, our ICE resembles Expectation-Maximization (EM) with variational approximation, and it takes advantage of the property of N-DUDE being locally robust around the true channel. With extensive experiments on several radically different types of data, we show that the ICE equipped N-DUDE (dubbed as ICE-N-DUDE) can perform \emph{universally} well regardless of the uncertainties in both the channel and the clean source. Moreover, we show ICE-N-DUDE becomes extremely robust to its hyperparameters and significantly outperforms the strong baseline that can deal with the channel uncertainties for denoising, the widely used Baum-Welch (BW) algorithm for hidden Markov models (HMM).
http://arxiv.org/abs/1902.08921
A fast and effective motion deblurring method has great application values in real life. This work presents an innovative approach in which a self-paced learning is combined with GAN to deblur image. First, We explain that a proper generator can be used as deep priors and point out that the solution for pixel-based loss is not same with the one for perception-based loss. By using these ideas as starting points, a Bi-Skip network is proposed to improve the generating ability and a bi-level loss is adopted to solve the problem that common conditions are non-identical. Second, considering that the complex motion blur will perturb the network in the training process, a self-paced mechanism is adopted to enhance the robustness of the network. Through extensive evaluations on both qualitative and quantitative criteria, it is demonstrated that our approach has a competitive advantage over state-of-the-art methods.
http://arxiv.org/abs/1902.08915
The fusion of color and lidar data plays a critical role in object detection for autonomous vehicles, which base their decision making on these inputs. While existing methods exploit redundant and complimentary information under good imaging conditions, they fail to do this in adverse weather and imaging conditions where the sensory streams can be asymmetrically distorted. These rare “edge-case” scenarios are not represented in available data sets, and existing fusion architectures are not designed to handle severe asymmetric distortions. We present a deep fusion architecture that allows for robust fusion in fog and snow without having large labeled training data available for these scenarios. Departing from proposal-level fusion, we propose a real-time single-shot model that adaptively fuses features driven by temporal coherence of the distortions. We validate the proposed method, trained on clean data, in simulation and on unseen conditions of in-the-wild driving scenarios.
http://arxiv.org/abs/1902.08913
Lexicalized parsing models are based on the assumptions that (i) constituents are organized around a lexical head (ii) bilexical statistics are crucial to solve ambiguities. In this paper, we introduce an unlexicalized transition-based parser for discontinuous constituency structures, based on a structure-label transition system and a bi-LSTM scoring system. We compare it to lexicalized parsing models in order to address the question of lexicalization in the context of discontinuous constituency parsing. Our experiments show that unlexicalized models systematically achieve higher results than lexicalized models, and provide additional empirical evidence that lexicalization is not necessary to achieve strong parsing results. Our best unlexicalized model sets a new state of the art on English and German discontinuous constituency treebanks. We further provide a per-phenomenon analysis of its errors on discontinuous constituents.
http://arxiv.org/abs/1902.08912
There are several distinct failure modes for overoptimization of systems on the basis of metrics. This occurs when a metric which can be used to improve a system is used to an extent that further optimization is ineffective or harmful, and is sometimes termed Goodhart’s Law. This class of failure is often poorly understood, partly because terminology for discussing them is ambiguous, and partly because discussion using this ambiguous terminology ignores distinctions between different failure modes of this general type. This paper expands on an earlier discussion by Garrabrant, which notes there are “(at least) four different mechanisms” that relate to Goodhart’s Law. This paper is intended to explore these mechanisms further, and specify more clearly how they occur. This discussion should be helpful in better understanding these types of failures in economic regulation, in public policy, in machine learning, and in Artificial Intelligence alignment. The importance of Goodhart effects depends on the amount of power directed towards optimizing the proxy, and so the increased optimization power offered by artificial intelligence makes it especially critical for that field.
http://arxiv.org/abs/1803.04585
Nowadays, the automatic detection of emotions is employed by many applications in different fields like security informatics, e-learning, humor detection, targeted advertising, etc. Many of these applications focus on social media and treat this problem as a classification problem, which requires preparing training data. The typical method for annotating the training data by human experts is considered time consuming, labor intensive and sometimes prone to error. Moreover, such an approach is not easily extensible to new domains/languages since such extensions require annotating new training data. In this study, we propose a distant supervised learning approach where the training sentences are automatically annotated based on the emojis they have. Such training data would be very cheap to produce compared with the manually created training data, thus, much larger training data can be easily obtained. On the other hand, this training data would naturally have lower quality as it may contain some errors in the annotation. Nonetheless, we experimentally show that training classifiers on cheap, large and possibly erroneous data annotated using this approach leads to more accurate results compared with training the same classifiers on the more expensive, much smaller and error-free manually annotated training data. Our experiments are conducted on an in-house dataset of emotional Arabic tweets and the classifiers we consider are: Support Vector Machine (SVM), Multinomial Naive Bayes (MNB) and Random Forest (RF). In addition to experimenting with single classifiers, we also consider using an ensemble of classifiers. The results show that using an automatically annotated training data that is only one order of magnitude larger the manually annotated one gives better results in almost all settings considered.
http://arxiv.org/abs/1902.08906
Efficient utilization of cooperating robots in the assembly of aircraft structures relies on balancing the workload of the robots and ensuring collision-free scheduling. We cast this problem as that of allocating a large number of repetitive assembly tasks, such as drilling holes and installing fasteners, among multiple robots. Such task allocation is often formulated as a Traveling Salesman Problem (TSP), which is NP-hard, implying that computing an exactly optimal solution is computationally prohibitive for real-world applications. The problem complexity is further exacerbated by intermittent robot failures necessitating real-time task reallocation. In this letter, we present an efficient method that exploits workpart geometry and problem structure to initially generate balanced and conflict-free robot schedules under nominal conditions. Subsequently, we deal with the failures by allowing the robots to first complete their nominal schedules and then employing a market-based optimizer to allocate the leftover tasks. Results show an improvement of 13% in schedule efficiencies as compared to a greedy multi-agent scheduler on a four robot system, which is especially promising for aircraft assembly processes that take many hours to complete. Moreover, the computation times are similar and small, typically less than one second.
http://arxiv.org/abs/1902.08905
We present a method for fine-grained face manipulation. Given a face image with an arbitrary expression, our method can synthesize another arbitrary expression by the same person. This is achieved by first fitting a 3D face model and then disentangling the face into a texture and a shape. We then learn different networks in these two spaces. In the texture space, we use a conditional generative network to change the appearance, and carefully design input formats and loss functions to achieve the best results. In the shape space, we use a fully connected network to predict the accurate shapes and use the available depth data for supervision. Both networks are conditioned on expression coefficients rather than discrete labels, allowing us to generate an unlimited amount of expressions. We show the superiority of this disentangling approach through both quantitative and qualitative studies. In a user study, our method is preferred in 85% of cases when compared to the most recent work. When compared to the ground truth, annotators cannot reliably distinguish between our synthesized images and real images, preferring our method in 53% of the cases.
http://arxiv.org/abs/1902.08900
With the advent of powerful, low-cost IoT systems, processing data closer to where the data originates, known as edge computing, has become an increasingly viable option. In addition to lowering the cost of networking infrastructures, edge computing reduces edge-cloud delay, which is essential for mission-critical applications. In this paper, we show the feasibility and study the performance of image classification using IoT devices. Specifically, we explore the relationships between various factors of image classification algorithms that may affect energy consumption such as dataset size, image resolution, algorithm type, algorithm phase, and device hardware. Our experiments show a strong, positive linear relationship between three predictor variables, namely model complexity, image resolution, and dataset size, with respect to energy consumption. In addition, in order to provide a means of predicting the energy consumption of an edge device performing image classification, we investigate the usage of three machine learning algorithms using the data generated from our experiments. The performance as well as the trade offs for using linear regression, Gaussian process, and random forests are discussed and validated. Our results indicate that the random forest model outperforms the two former algorithms, with an R-squared value of 0.95 and 0.79 for two different validation datasets.
http://arxiv.org/abs/1902.11119
This paper describes the ARIEL-CMU submissions to the Low Resource Human Language Technologies (LoReHLT) 2018 evaluations for the tasks Machine Translation (MT), Entity Discovery and Linking (EDL), and detection of Situation Frames in Text and Speech (SF Text and Speech).
http://arxiv.org/abs/1902.08899
Tuberculosis is a deadly infectious disease prevalent around the world. Due to the lack of proper technology in place, the early detection of this disease is unattainable. Also, the available methods to detect Tuberculosis is not up-to a commendable standards due to their dependency on unnecessary features, this make such technology obsolete for a reliable health-care technology. In this paper, I propose a deep-learning based system which diagnoses tuberculosis based on the important features in Chest X-rays along with original chest X-rays. Employing our system will accelerate the process of tuberculosis diagnosis by overcoming the need to perform the time-consuming sputum-based testing method (Diagnostic Microbiology). In contrast to the previous methods \cite{kant2018towards, melendez2016automated}, our work utilizes the state-of-the-art ResNet \cite{he2016deep} with proper data augmentation using traditional robust features like Haar \cite{viola2005detecting,viola2001rapid} and LBP \cite{ojala1994performance,ojala1996comparative}. I observed that such a procedure enhances the rate of tuberculosis detection to a highly satisfactory level. Our work uses the publicly available pulmonary chest X-ray dataset to train our network \cite{jaeger2014two}. Nevertheless, the publicly available dataset is very small and is inadequate to achieve the best accuracy. To overcome this issue I have devised an intuitive feature based data augmentation pipeline. Our approach shall help the deep neural network \cite{lecun2015deep,he2016deep,krizhevsky2012imagenet} to focus its training on tuberculosis affected regions making it more robust and accurate, when compared to other conventional methods that use procedures like mirroring and rotation. By using our simple yet powerful techniques, I observed a 10\% boost in performance accuracy.
http://arxiv.org/abs/1902.08897
Data is one of the essential ingredients to power deep learning research. Small datasets, especially specific to medical institutes, bring challenges to deep learning training stage. This work aims to develop a practical deep multimodal that can classify patients into abnormal and normal categories accurately as well as assist radiologists to detect visual and textual anomalies by locating areas of interest. The detection of the anomalies is achieved through a novel technique which extends the integrated gradients methodology with an unsupervised clustering algorithm. This technique also introduces a tuning parameter which trades off true positive signals to denoise false positive signals in the detection process. To overcome the challenges of the small training dataset which only has 3K frontal X-ray images and medical reports in pairs, we have adopted transfer learning for the multimodal which concatenates the layers of image and text submodels. The image submodel was trained on the vast ChestX-ray14 dataset, while the text submodel transferred a pertained word embedding layer from a hospital-specific corpus. Experimental results show that our multimodal improves the accuracy of the classification by 4% and 7% on average of 50 epochs, compared to the individual text and image model, respectively.
http://arxiv.org/abs/1902.08888
We introduce a novel method that enables parameter-efficient transfer and multi-task learning with deep neural networks. The basic approach is to learn a model patch - a small set of parameters - that will specialize to each task, instead of fine-tuning the last layer or the entire network. For instance, we show that learning a set of scales and biases is sufficient to convert a pretrained network to perform well on qualitatively different problems (e.g. converting a Single Shot MultiBox Detection (SSD) model into a 1000-class image classification model while reusing 98% of parameters of the SSD feature extractor). Similarly, we show that re-learning existing low-parameter layers (such as depth-wise convolutions) while keeping the rest of the network frozen also improves transfer-learning accuracy significantly. Our approach allows both simultaneous (multi-task) as well as sequential transfer learning. In several multi-task learning problems, despite using much fewer parameters than traditional logits-only fine-tuning, we match single-task performance.
http://arxiv.org/abs/1810.10703
Semantic inpainting is the task of inferring missing pixels in an image given surrounding pixels and high level image semantics. Most semantic inpainting algorithms are deterministic: given an image with missing regions, a single inpainted image is generated. However, there are often several plausible inpaintings for a given missing region. In this paper, we propose a method to perform probabilistic semantic inpainting by building a model, based on PixelCNNs, that learns a distribution of images conditioned on a subset of visible pixels. Experiments on the MNIST and CelebA datasets show that our method produces diverse and realistic inpaintings.
http://arxiv.org/abs/1810.03728
Defining action spaces for conversational agents and optimizing their decision-making process with reinforcement learning is an enduring challenge. Common practice has been to use handcrafted dialog acts, or the output vocabulary, e.g. in neural encoder decoders, as the action spaces. Both have their own limitations. This paper proposes a novel latent action framework that treats the action spaces of an end-to-end dialog agent as latent variables and develops unsupervised methods in order to induce its own action space from the data. Comprehensive experiments are conducted examining both continuous and discrete action types and two different optimization methods based on stochastic variational inference. Results show that the proposed latent actions achieve superior empirical performance improvement over previous word-level policy gradient methods on both DealOrNoDeal and MultiWoz dialogs. Our detailed analysis also provides insights about various latent variable approaches for policy learning and can serve as a foundation for developing better latent actions in future research.
http://arxiv.org/abs/1902.08858
We present our system for the CLIN29 shared task on cross-genre gender detection for Dutch. We experimented with a multitude of neural models (CNN, RNN, LSTM, etc.), more “traditional” models (SVM, RF, LogReg, etc.), different feature sets as well as data pre-processing. The final results suggested that using tokenized, non-lowercased data works best for most of the neural models, while a combination of word clusters, character trigrams and word lists showed to be most beneficial for the majority of the more “traditional” (that is, non-neural) models, beating features used in previous tasks such as n-grams, character n-grams, part-of-speech tags and combinations thereof. In contradiction with the results described in previous comparable shared tasks, our neural models performed better than our best traditional approaches with our best feature set-up. Our final model consisted of a weighted ensemble model combining the top 25 models. Our final model won both the in-domain gender prediction task and the cross-genre challenge, achieving an average accuracy of 64.93% on the in-domain gender prediction task, and 56.26% on cross-genre gender prediction.
http://arxiv.org/abs/1902.08856
Recently remarkable success has been achieved in machine reading comprehension (MRC). However, it is still difficult to interpret the predictions of existing MRC models. In this paper, we focus on: extracting evidence sentences that can explain/support answer predictions for multiple-choice MRC tasks, where the majority of answer options cannot be directly extracted from reference documents; studying the impacts of using the extracted sentences as the input of MRC models. Due to the lack of ground truth evidence sentence labels in most cases, we apply distant supervision to generate imperfect labels and then use them to train a neural evidence extractor. To denoise the noisy labels, we treat labels as latent variables and define priors over these latent variables by incorporating rich linguistic knowledge under a recently proposed deep probabilistic logic learning framework. We feed the extracted evidence sentences into existing MRC models and evaluate the end-to-end performance on three challenging multiple-choice MRC datasets: MultiRC, DREAM, and RACE, achieving comparable or better performance than the same models that take the full context as input. Our evidence extractor also outperforms a state-of-the-art sentence selector by a large margin on two open-domain question answering datasets: Quasar-T and SearchQA. To the best of our knowledge, this is the first work addressing evidence sentence extraction for multiple-choice MRC.
http://arxiv.org/abs/1902.08852
In this paper, we propose a novel representation for text documents based on aggregating word embedding vectors into document embeddings. Our approach is inspired by the Vector of Locally-Aggregated Descriptors used for image representation, and it works as follows. First, the word embeddings gathered from a collection of documents are clustered by k-means in order to learn a codebook of semnatically-related word embeddings. Each word embedding is then associated to its nearest cluster centroid (codeword). The Vector of Locally-Aggregated Word Embeddings (VLAWE) representation of a document is then computed by accumulating the differences between each codeword vector and each word vector (from the document) associated to the respective codeword. We plug the VLAWE representation, which is learned in an unsupervised manner, into a classifier and show that it is useful for a diverse set of text classification tasks. We compare our approach with a broad range of recent state-of-the-art methods, demonstrating the effectiveness of our approach. Furthermore, we obtain a considerable improvement on the Movie Review data set, reporting an accuracy of 93.3%, which represents an absolute gain of 10% over the state-of-the-art approach.
http://arxiv.org/abs/1902.08850
We study the relationship between geometry and capacity measures for deep neural networks from an invariance viewpoint. We introduce a new notion of capacity — the Fisher-Rao norm — that possesses desirable invariance properties and is motivated by Information Geometry. We discover an analytical characterization of the new capacity measure, through which we establish norm-comparison inequalities and further show that the new measure serves as an umbrella for several existing norm-based complexity measures. We discuss upper bounds on the generalization error induced by the proposed measure. Extensive numerical experiments on CIFAR-10 support our theoretical findings. Our theoretical analysis rests on a key structural lemma about partial derivatives of multi-layer rectifier networks.
http://arxiv.org/abs/1711.01530
Automatically evaluating the quality of dialogue responses for unstructured domains is a challenging problem. ADEM(Lowe et al. 2017) formulated the automatic evaluation of dialogue systems as a learning problem and showed that such a model was able to predict responses which correlate significantly with human judgements, both at utterance and system level. Their system was shown to have beaten word-overlap metrics such as BLEU with large margins. We start with the question of whether an adversary can game the ADEM model. We design a battery of targeted attacks at the neural network based ADEM evaluation system and show that automatic evaluation of dialogue systems still has a long way to go. ADEM can get confused with a variation as simple as reversing the word order in the text! We report experiments on several such adversarial scenarios that draw out counterintuitive scores on the dialogue responses. We take a systematic look at the scoring function proposed by ADEM and connect it to linear system theory to predict the shortcomings evident in the system. We also devise an attack that can fool such a system to rate a response generation system as favorable. Finally, we allude to future research directions of using the adversarial attacks to design a truly automated dialogue evaluation system.
http://arxiv.org/abs/1902.08832
Categories such as animal or furniture are acquired at an early age and play an important role in processing, organizing, and communicating world knowledge. Categories exist across cultures: they allow to efficiently represent the complexity of the world, and members of a community strongly agree on their nature, revealing a shared mental representation. Models of category learning and representation, however, are typically tested on data from small-scale experiments involving small sets of concepts with artificially restricted features; and experiments predominantly involve participants of selected cultural and socio-economical groups (very often involving western native speakers of English such as U.S. college students) . This work investigates whether models of categorization generalize (a) to rich and noisy data approximating the environment humans live in; and (b) across languages and cultures. We present a Bayesian cognitive model designed to jointly learn categories and their structured representation from natural language text which allows us to (a) evaluate performance on a large scale, and (b) apply our model to a diverse set of languages. We show that meaningful categories comprising hundreds of concepts and richly structured featural representations emerge across languages. Our work illustrates the potential of recent advances in computational modeling and large scale naturalistic datasets for cognitive science research.
http://arxiv.org/abs/1902.08830
In this paper, we present the mechanics and algorithms to compute the set of feasible motions of an object pushed in a plane. This set is known as the motion cone and was previously described for non-prehensile manipulation tasks in the horizontal plane. We generalize its geometric construction to a broader set of planar tasks, where external forces such as gravity influence the dynamics of pushing, and prehensile tasks, where there are complex interactions between the gripper, object, and pusher. We show that the motion cone is defined by a set of low-curvature surfaces and provide a polyhedral cone approximation to it. We verify its validity with 2000 pushing experiments recorded with motion tracking system. Motion cones abstract the algebra involved in simulating frictional pushing by providing bounds on the set of feasible motions and by characterizing which pushes will stick or slip. We demonstrate their use for the dynamic propagation step in a sampling-based planning algorithm for in-hand manipulation. The planner generates trajectories that involve sequences of continuous pushes with 5-1000x speed improvements to equivalent algorithms. Video Summary – https://youtu.be/tVDO8QMuYhc
http://arxiv.org/abs/1810.00219
While neural networks have been used extensively to make substantial progress in the machine translation task, they are known for being heavily dependent on the availability of large amounts of training data. Recent efforts have tried to alleviate the data sparsity problem by augmenting the training data using different strategies, such as back-translation. Along with the data scarcity, the out-of-vocabulary words, mostly entities and terminological expressions, pose a difficult challenge to Neural Machine Translation systems. In this paper, we hypothesize that knowledge graphs enhance the semantic feature extraction of neural models, thus optimizing the translation of entities and terminological expressions in texts and consequently leading to a better translation quality. We hence investigate two different strategies for incorporating knowledge graphs into neural models without modifying the neural network architectures. We also examine the effectiveness of our augmentation method to recurrent and non-recurrent (self-attentional) neural architectures. Our knowledge graph augmented neural translation model, dubbed KG-NMT, achieves significant and consistent improvements of +3 BLEU, METEOR and chrF3 on average on the newstest datasets between 2014 and 2018 for WMT English-German translation task.
http://arxiv.org/abs/1902.08816
Social media brings about new ways of communication among people and is influencing trading strategies in the market. The popularity of social networks produces a large collection of unstructured data such as text and image in a variety of disciplines like business and health. The main element of social media arises as text which provokes a set of challenges for traditional information retrieval and natural language processing tools. Informal language, spelling errors, abbreviations, and special characters are typical in social media posts. These features lead to a prohibitively large vocabulary size for text mining methods. Another problem with traditional social text mining techniques is that they fail to take semantic relations into account, which is essential in a domain of applications such as event detection, opinion mining, and news recommendation. This paper set out to employ a network-based viewpoint on text documents and investigate the usefulness of graph representation to exploit word relations and semantics of the textual data. Moreover, the proposed approach makes use of a random walker to extract deep features of a graph to facilitate the task of document classification. The experimental results indicate that the proposed approach defeats the earlier sentiment analysis methods based on several benchmark datasets, and it generalizes well on different datasets without dependency for pre-trained word embeddings.
http://arxiv.org/abs/1902.10247
Face recognition in real life situations like low illumination condition is still an open challenge in biometric security. It is well established that the state-of-the-art methods in face recognition provide low accuracy in the case of poor illumination. In this work, we propose an algorithm for a more robust illumination invariant face recognition using a multi-modal approach. We propose a new dataset consisting of aligned faces of thermal and visual images of a hundred subjects. We then apply face detection on thermal images using the biggest blob extraction method and apply them for fusing images of different modalities for the purpose of face recognition. An algorithm is proposed to implement fusion of thermal and visual images. We reason for why relying on only one modality can give erroneous results. We use a lighter and faster CNN model called MobileNet for the purpose of face recognition with faster inferencing and to be able to be use it in real time biometric systems. We test our proposed method on our own created dataset to show that real-time face recognition on fused images shows far better results than using visual or thermal images separately.
http://arxiv.org/abs/1902.08802
Chinese is a logographic writing system, and the shape of Chinese characters contain rich syntactic and semantic information. In this paper, we propose a model to learn Chinese word embeddings via two-level composition: (1) a convolutional neural network to extract the intra-character compositionality from the visual shape of a character; (2) a recurrent neural network with self-attention to compose character representation into word embeddings. The word embeddings along with the network parameters are learned in the Skip-Gram framework. Evaluations demonstrate the superior performance of our model on four tasks: word similarity, sentiment analysis, named entity recognition and part-of-speech tagging.
http://arxiv.org/abs/1902.08795
Background: Building visual encoding models to accurately predict visual responses is a central challenge for current vision-based brain-machine interface techniques. To achieve high prediction accuracy on neural signals, visual encoding models should include precise visual features and appropriate prediction algorithms. Most existing visual encoding models employ hand-craft visual features (e.g., Gabor wavelets or semantic labels) or data-driven features (e.g., features extracted from deep neural networks (DNN)). They also assume a linear mapping between feature representation to brain activity. However, it remains unknown whether such linear mapping is sufficient for maximizing prediction accuracy. New Method: We construct a new visual encoding framework to predict cortical responses in a benchmark functional magnetic resonance imaging (fMRI) dataset. In this framework, we employ the transfer learning technique to incorporate a pre-trained DNN (i.e., AlexNet) and train a nonlinear mapping from visual features to brain activity. This nonlinear mapping replaces the conventional linear mapping and is supposed to improve prediction accuracy on brain activity. Results: The proposed framework can significantly predict responses of over 20% voxels in early visual areas (i.e., V1-lateral occipital region, LO) and achieve unprecedented prediction accuracy. Comparison with Existing Methods: Comparing to two conventional visual encoding models, we find that the proposed encoding model shows consistent higher prediction accuracy in all early visual areas, especially in relatively anterior visual areas (i.e., V4 and LO). Conclusions: Our work proposes a new framework to utilize pre-trained visual features and train non-linear mappings from visual features to brain activity.
http://arxiv.org/abs/1902.08793
The existing core methods, which are employed by the popular CTL model checking tools, are facing the famous state explode problem. In our previous study, a method based on the Machine Learning (ML) algorithms was proposed to address this problem. However, the accuracy is not satisfactory. First, we conduct a comprehensive experiment on Graph Lab to seek the optimal accuracy using the five machine learning algorithms. Second, given the optimal accuracy, the average time is seeked. The results show that the Logistic Regressive (LR)-based approach can simulate CTL model checking with the accuracy of 98.8%, and its average efficiency is 459 times higher than that of the existing method, as well as the Boosted Tree (BT)-based approach can simulate CTL model checking with the accuracy of 98.7%, and its average efficiency is 639 times higher than that of the existing method.
http://arxiv.org/abs/1902.08789
Deep learning based facial expression recognition (FER) has received a lot of attention in the past few years. Most of the existing deep learning based FER methods do not consider domain knowledge well, which thereby fail to extract representative features. In this work, we propose a novel FER framework, named Facial Motion Prior Networks (FMPN). Particularly, we introduce an addition branch to generate a facial mask so as to focus on facial muscle moving regions. To guide the facial mask learning, we propose to incorporate prior domain knowledge by using the average differences between neutral faces and the corresponding expressive faces as the guidance. Extensive experiments on four facial expression benchmark datasets demonstrate the effectiveness of the proposed method, compared with the state-of-the-art approaches.
http://arxiv.org/abs/1902.08788
Deep neural networks (DNN) have been a de facto standard for nowadays biometric recognition solutions. A serious, but still overlooked problem in these DNN-based recognition systems is their vulnerability against adversarial attacks. Adversarial attacks can easily cause the output of a DNN system to greatly distort with only tiny changes in its input. Such distortions can potentially lead to an unexpected match between a valid biometric and a synthetic one constructed by a strategic attacker, raising security issue. In this work, we show how this issue can be resolved by learning robust biometric features through a deep, information-theoretic framework, which builds upon the recent deep variational information bottleneck method but is carefully adapted to biometric recognition tasks. Empirical evaluation demonstrates that our method not only offers stronger robustness against adversarial attacks but also provides better recognition performance over state-of-the-art approaches.
http://arxiv.org/abs/1902.08785
Blooms Taxonomy (BT) have been used to classify the objectives of learning outcome by dividing the learning into three different domains; the cognitive domain, the effective domain and the psychomotor domain. In this paper, we are introducing a new approach to classify the questions and learning outcome statements (LOS) into Blooms taxonomy (BT) and to verify BT verb lists, which are being cited and used by academicians to write questions and (LOS). An experiment was designed to investigate the semantic relationship between the action verbs used in both questions and LOS to obtain more accurate classification of the levels of BT. A sample of 775 different action verbs collected from different universities allows us to measure an accurate and clear-cut cognitive level for the action verb. It is worth mentioning that natural language processing techniques were used to develop our rules as to induce the questions into chunks in order to extract the action verbs. Our proposed solution was able to classify the action verb into a precise level of the cognitive domain. We, on our side, have tested and evaluated our proposed solution using confusion matrix. The results of evaluation tests yielded 97% for the macro average of precision and 90% for F1. Thus, the outcome of the research suggests that it is crucial to analyse and verify the action verbs cited and used by academicians to write LOS and classify their questions based on blooms taxonomy in order to obtain a definite and more accurate classification.
http://arxiv.org/abs/1706.03191
Pruning large neural networks while maintaining their performance is often desirable due to the reduced space and time complexity. In existing methods, pruning is done within an iterative optimization procedure with either heuristically designed pruning schedules or additional hyperparameters, undermining their utility. In this work, we present a new approach that prunes a given network once at initialization prior to training. To achieve this, we introduce a saliency criterion based on connection sensitivity that identifies structurally important connections in the network for the given task. This eliminates the need for both pretraining and the complex pruning schedule while making it robust to architecture variations. After pruning, the sparse network is trained in the standard way. Our method obtains extremely sparse networks with virtually the same accuracy as the reference network on the MNIST, CIFAR-10, and Tiny-ImageNet classification tasks and is broadly applicable to various architectures including convolutional, residual and recurrent networks. Unlike existing methods, our approach enables us to demonstrate that the retained connections are indeed relevant to the given task.
http://arxiv.org/abs/1810.02340
Verification of neural networks enables us to gauge their robustness against adversarial attacks. Verification algorithms fall into two categories: exact verifiers that run in exponential time and relaxed verifiers that are efficient but incomplete. In this paper, we unify all existing LP-relaxed verifiers, to the best of our knowledge, under a general convex relaxation framework. This framework works for neural networks with diverse architectures and nonlinearities and covers both primal and dual views of robust verification. We further prove strong duality between the primal and dual problems under very mild conditions. Next, we perform large-scale experiments, amounting to more than 22 CPU-years, to obtain exact solution to the convex-relaxed problem that is optimal within our framework for ReLU networks. We find the exact solution does not significantly improve upon the gap between PGD and existing relaxed verifiers for various networks trained normally or robustly on MNIST and CIFAR datasets. Our results suggest there is an inherent barrier to tight verification for the large class of methods captured by our framework. We discuss possible causes of this barrier and potential future directions for bypassing it.
http://arxiv.org/abs/1902.08722
Prognostic tumor growth modeling via medical imaging observations is a challenging yet important problem in precision and predictive medicine. Traditionally, this problem is tackled through mathematical modeling and evaluated using relatively small patient datasets. Recent advances of convolutional networks (ConvNets) have demonstrated their higher accuracy than mathematical models in predicting future tumor volumes. This indicates that deep learning may have great potentials on addressing such problem. The state-of-the-art work models the cell invasion and mass-effect of tumor growth by training separate ConvNets on 2D image patches. Nevertheless such a 2D modeling approach cannot make full use of the spatial-temporal imaging context of the tumor’s longitudinal 4D (3D + time) patient data. Moreover, previous methods are incapable to predict clinically-relevant tumor properties, other than the tumor volumes. In this paper, we exploit to formulate the tumor growth process through convolutional LSTMs (ConvLSTM) that extract tumor’s static imaging appearances and simultaneously capture its temporal dynamic changes within a single network. We extend ConvLSTM into the spatial-temporal domain (ST-ConvLSTM) by jointly learning the inter-slice 3D contexts and the longitudinal dynamics. Our approach can incorporate other non-imaging patient information in an end-to-end trainable manner. Experiments are conducted on the largest 4D longitudinal tumor dataset of 33 patients to date. Results validate that the proposed ST-ConvLSTM model produces a Dice score of 83.2%+-5.1% and a RVD of 11.2%+-10.8%, both statistically significantly outperforming (p<0.05) other compared methods of traditional linear model, ConvLSTM, and generative adversarial network (GAN) under the metric of predicting future tumor volumes. Last, our new method enables the prediction of both cell density and CT intensity numbers.
http://arxiv.org/abs/1902.08716
Neural architecture search (NAS) has a great impact by automatically designing effective neural network architectures. However, the prohibitive computational demand of conventional NAS algorithms (e.g. $10^4$ GPU hours) makes it difficult to \emph{directly} search the architectures on large-scale tasks (e.g. ImageNet). Differentiable NAS can reduce the cost of GPU hours via a continuous representation of network architecture but suffers from the high GPU memory consumption issue (grow linearly w.r.t. candidate set size). As a result, they need to utilize~\emph{proxy} tasks, such as training on a smaller dataset, or learning with only a few blocks, or training just for a few epochs. These architectures optimized on proxy tasks are not guaranteed to be optimal on the target task. In this paper, we present \emph{ProxylessNAS} that can \emph{directly} learn the architectures for large-scale target tasks and target hardware platforms. We address the high memory consumption issue of differentiable NAS and reduce the computational cost (GPU hours and GPU memory) to the same level of regular training while still allowing a large candidate set. Experiments on CIFAR-10 and ImageNet demonstrate the effectiveness of directness and specialization. On CIFAR-10, our model achieves 2.08\% test error with only 5.7M parameters, better than the previous state-of-the-art architecture AmoebaNet-B, while using 6$\times$ fewer parameters. On ImageNet, our model achieves 3.1\% better top-1 accuracy than MobileNetV2, while being 1.2$\times$ faster with measured GPU latency. We also apply ProxylessNAS to specialize neural architectures for hardware with direct hardware metrics (e.g. latency) and provide insights for efficient CNN architecture design.
https://arxiv.org/abs/1812.00332
Modern neural networks are often augmented with an attention mechanism, which tells the network where to focus within the input. We propose in this paper a new framework for sparse and structured attention, building upon a smoothed max operator. We show that the gradient of this operator defines a mapping from real values to probabilities, suitable as an attention mechanism. Our framework includes softmax and a slight generalization of the recently-proposed sparsemax as special cases. However, we also show how our framework can incorporate modern structured penalties, resulting in more interpretable attention mechanisms, that focus on entire segments or groups of an input. We derive efficient algorithms to compute the forward and backward passes of our attention mechanisms, enabling their use in a neural network trained with backpropagation. To showcase their potential as a drop-in replacement for existing ones, we evaluate our attention mechanisms on three large-scale tasks: textual entailment, machine translation, and sentence summarization. Our attention mechanisms improve interpretability without sacrificing performance; notably, on textual entailment and summarization, we outperform the standard attention mechanisms based on softmax and sparsemax.
http://arxiv.org/abs/1705.07704
Efficient audio synthesis is an inherently difficult machine learning task, as human perception is sensitive to both global structure and fine-scale waveform coherence. Autoregressive models, such as WaveNet, model local structure at the expense of global latent structure and slow iterative sampling, while Generative Adversarial Networks (GANs), have global latent conditioning and efficient parallel sampling, but struggle to generate locally-coherent audio waveforms. Herein, we demonstrate that GANs can in fact generate high-fidelity and locally-coherent audio by modeling log magnitudes and instantaneous frequencies with sufficient frequency resolution in the spectral domain. Through extensive empirical investigations on the NSynth dataset, we demonstrate that GANs are able to outperform strong WaveNet baselines on automated and human evaluation metrics, and efficiently generate audio several orders of magnitude faster than their autoregressive counterparts.
http://arxiv.org/abs/1902.08710
In this paper, we present our method of using fixed-size ordinally forgetting encoding (FOFE) to solve the word sense disambiguation (WSD) problem. FOFE enables us to encode variable-length sequence of words into a theoretically unique fixed-size representation that can be fed into a feed forward neural network (FFNN), while keeping the positional information between words. In our method, a FOFE-based FFNN is used to train a pseudo language model over unlabelled corpus, then the pre-trained language model is capable of abstracting the surrounding context of polyseme instances in labelled corpus into context embeddings. Next, we take advantage of these context embeddings towards WSD classification. We conducted experiments on several WSD data sets, which demonstrates that our proposed method can achieve comparable performance to that of the state-of-the-art approach at the expense of much lower computational cost.
http://arxiv.org/abs/1902.10246
Learning accurate dynamics models is necessary for optimal, compliant control of robotic systems. Current approaches to white-box modeling using analytic parameterizations, or black-box modeling using neural networks, can suffer from high bias or high variance. We address the need for a flexible, gray-box model of mechanical systems that can seamlessly incorporate prior knowledge where it is available, and train expressive function approximators where it is not. We propose to parameterize a mechanical system using neural networks to model its Lagrangian and the generalized forces that act on it. We test our method on a simulated, actuated double pendulum. We show that our method outperforms a naive, black-box model in terms of data-efficiency, as well as performance in model-based reinforcement learning. We also conduct a systematic study of our method’s ability to incorporate available prior knowledge about the system to improve data efficiency.
http://arxiv.org/abs/1902.08705
We introduce adaptive input representations for neural language modeling which extend the adaptive softmax of Grave et al. (2017) to input representations of variable capacity. There are several choices on how to factorize the input and output layers, and whether to model words, characters or sub-word units. We perform a systematic comparison of popular choices for a self-attentional architecture. Our experiments show that models equipped with adaptive embeddings are more than twice as fast to train than the popular character input CNN while having a lower number of parameters. On the WikiText-103 benchmark we achieve 18.7 perplexity, an improvement of 10.5 perplexity compared to the previously best published result and on the Billion Word benchmark, we achieve 23.02 perplexity.
http://arxiv.org/abs/1809.10853
Neural network-based representations (“embeddings”) have dramatically advanced natural language processing (NLP) tasks in the past few years. This certainly holds for clinical concept extraction, especially when combined with deep learning-based models. Recently, however, more advanced embedding methods and representations (e.g., ELMo, BERT) have further pushed the state-of-the-art in NLP. While these certainly improve clinical concept extraction as well, there are no commonly agreed upon best practices for how to integrate these representations for extracting concepts. The purpose of this study, then, is to explore the space of possible options in utilizing these new models, including comparing these to more traditional word embedding methods (word2vec, GloVe, fastText). We evaluate a battery of embedding methods on four clinical concept extraction corpora, explore effects of pre-training on extraction performance, and present an intuitive way to understand the semantic information encoded by advanced contextualized representations. Notably, we achieved new state-of-the-art performances across all four corpora.
http://arxiv.org/abs/1902.08691
Deep neural networks, in particular convolutional neural networks, have become highly effective tools for compressing images and solving inverse problems including denoising, inpainting, and reconstruction from few and noisy measurements. This success can be attributed in part to their ability to represent and generate natural images well. Contrary to classical tools such as wavelets, image-generating deep neural networks have a large number of parameters—typically a multiple of their output dimension—and need to be trained on large datasets. In this paper, we propose an untrained simple image model, called the deep decoder, which is a deep neural network that can generate natural images from very few weight parameters. The deep decoder has a simple architecture with no convolutions and fewer weight parameters than the output dimensionality. This underparameterization enables the deep decoder to compress images into a concise set of network weights, which we show is on par with wavelet-based thresholding. Further, underparameterization provides a barrier to overfitting, allowing the deep decoder to have state-of-the-art performance for denoising. The deep decoder is simple in the sense that each layer has an identical structure that consists of only one upsampling unit, pixel-wise linear combination of channels, ReLU activation, and channelwise normalization. This simplicity makes the network amenable to theoretical analysis, and it sheds light on the aspects of neural networks that enable them to form effective signal representations.
http://arxiv.org/abs/1810.03982
Wings of flying animals not only can generate lift and control torque but also can sense their surroundings. Such dual functions of sensing and actuation coupled in one element are particularly useful for small sized bio-inspired robotic flyers, whose weight, size, and power are under constraint. In this work, we present the first flapping-wing robot using its flapping wings for environmental perception and navigation in tight space, without the need for any visual feedback. Specifically, we introduce Purdue Hummingbird, a flapping-wing robot with 17cm wingspan and 12 grams weight, as our test platform. By interpreting the wing loading feedback and its variations, the vehicle can detect the presence of environmental changes such as grounds, walls, stairs, obstacles and wind gust. The instantaneous wing loading can be obtained through the measurements and interpretation of the current feedback by the motor that actuates the wing. The effectiveness of the proposed approach is experimentally demonstrated on several challenging flight tasks without vision: terrain following, wall following and going through a narrow corridor. To ensure flight stability, a robust controller was designed for handling unforeseen disturbances during the flight. Sensing and navigating one’s environment through actuator loading is a promising method for mobile robots and it can serve as an alternative or complementary method to visual perception.
http://arxiv.org/abs/1902.08688
Most of the prior work on multi-agent reinforcement learning (MARL) achieves optimal collaboration by directly controlling the agents to maximize a common reward. In this paper, we aim to address this from a different angle. In particular, we consider scenarios where there are self-interested agents (i.e., worker agents) which have their own minds (preferences, intentions, skills, etc.) and can not be dictated to perform tasks they do not wish to do. For achieving optimal coordination among these agents, we train a super agent (i.e., the manager) to manage them by first inferring their minds based on both current and past observations and then initiating contracts to assign suitable tasks to workers and promise to reward them with corresponding bonuses so that they will agree to work together. The objective of the manager is maximizing the overall productivity as well as minimizing payments made to the workers for ad-hoc worker teaming. To train the manager, we propose Mind-aware Multi-agent Management Reinforcement Learning (M^3RL), which consists of agent modeling and policy learning. We have evaluated our approach in two environments, Resource Collection and Crafting, to simulate multi-agent management problems with various task settings and multiple designs for the worker agents. The experimental results have validated the effectiveness of our approach in modeling worker agents’ minds online, and in achieving optimal ad-hoc teaming with good generalization and fast adaptation.
http://arxiv.org/abs/1810.00147
Using deep reinforcement learning, we train control policies for autonomous vehicles leading a platoon of vehicles onto a roundabout. Using Flow, a library for deep reinforcement learning in micro-simulators, we train two policies, one policy with noise injected into the state and action space and one without any injected noise. In simulation, the autonomous vehicle learns an emergent metering behavior for both policies in which it slows to allow for smoother merging. We then directly transfer this policy without any tuning to the University of Delaware Scaled Smart City (UDSSC), a 1:25 scale testbed for connected and automated vehicles. We characterize the performance of both policies on the scaled city. We show that the noise-free policy winds up crashing and only occasionally metering. However, the noise-injected policy consistently performs the metering behavior and remains collision-free, suggesting that the noise helps with the zero-shot policy transfer. Additionally, the transferred, noise-injected policy leads to a 5% reduction of average travel time and a reduction of 22% in maximum travel time in the UDSSC. Videos of the controllers can be found at https://sites.google.com/view/iccps-policy-transfer.
http://arxiv.org/abs/1812.06120
Accurate diagnosis of breast cancer in histopathology images is challenging due to the heterogeneity of cancer cell growth as well as of a variety of benign breast tissue proliferative lesions. In this paper, we propose a practical and self-interpretable invasive cancer diagnosis solution. With minimum annotation information, the proposed method mines contrast patterns between normal and malignant images in unsupervised manner and generates a probability map of abnormalities to verify its reasoning. Particularly, a fully convolutional autoencoder is used to learn the dominant structural patterns among normal image patches. Patches that do not share the characteristics of this normal population are detected and analyzed by one-class support vector machine and 1-layer neural network. We apply the proposed method to a public breast cancer image set. Our results, in consultation with a senior pathologist, demonstrate that the proposed method outperforms existing methods. The obtained probability map could benefit the pathology practice by providing visualized verification data and potentially leads to a better understanding of data-driven diagnosis solutions.
http://arxiv.org/abs/1902.08670
We introduce and study the problem of planning a trajectory for an agent to carry out a scouting mission while avoiding being detected by an adversarial guard. This introduces an adversarial version of classical visibility-based planning problems such as the Watchman Route Problem. The agent receives a positive reward for increasing its visibility and a negative penalty when it is detected by the guard. The objective is to find a finite-horizon path for the agent that balances the trade-off maximizing visibility and minimizing detectability. We model this problem as a sequential two-player zero-sum discrete game. A minimax tree search can give the optimal policy for the agent but requires an exponential-time computation and space. We propose several pruning techniques to reduce the computational cost while still preserving optimality guarantees. Simulation results show that the proposed strategy prunes approximately three orders of magnitude nodes as compared to the brute-force strategy.
http://arxiv.org/abs/1807.09437
A good conversation requires balance – between simplicity and detail; staying on topic and changing it; asking questions and answering them. Although dialogue agents are commonly evaluated via human judgments of overall quality, the relationship between quality and these individual factors is less well-studied. In this work, we examine two controllable neural text generation methods, conditional training and weighted decoding, in order to control four important attributes for chitchat dialogue: repetition, specificity, response-relatedness and question-asking. We conduct a large-scale human evaluation to measure the effect of these control parameters on multi-turn interactive conversations on the PersonaChat task. We provide a detailed analysis of their relationship to high-level aspects of conversation, and show that by controlling combinations of these variables our models obtain clear improvements in human quality judgments.
http://arxiv.org/abs/1902.08654