The majority of approaches for acquiring dense 3D environment maps with RGB-D cameras assumes static environments or rejects moving objects as outliers. The representation and tracking of moving objects, however, has significant potential for applications in robotics or augmented reality. In this paper, we propose a novel approach to dynamic SLAM with dense object-level representations. We represent rigid objects in local volumetric signed distance function (SDF) maps, and formulate multi-object tracking as direct alignment of RGB-D images with the SDF representations. Our main novelty is a probabilistic formulation which naturally leads to strategies for data association and occlusion handling. We analyze our approach in experiments and demonstrate that our approach compares favorably with the state-of-the-art methods in terms of robustness and accuracy.
http://arxiv.org/abs/1904.11781
Eye tracking (ET) is a research method that receives growing interest in mathematics education research (MER). This paper aims to give a literature overview, specifically focusing on the evolution of interest in this technology, ET equipment, and analysis methods used in mathematics education. To capture the current state, we focus on papers published in the proceedings of PME, one of the primary conferences dedicated to MER, of the last ten years. We identify trends in interest, methodology, and methods of analysis that are used in the community, and discuss possible future developments.
http://arxiv.org/abs/1904.12581
Scarce data is a major challenge to scaling robot learning to truly complex tasks, as we need to generalize locally learned policies over different task contexts. Contextual policy search offers data-efficient learning and generalization by explicitly conditioning the policy on a parametric context space. In this paper, we further structure the contextual policy representation. We propose to factor contexts into two components: target contexts that describe the task objectives, e.g. target position for throwing a ball; and environment contexts that characterize the environment, e.g. initial position or mass of the ball. Our key observation is that experience can be directly generalized over target contexts. We show that this can be easily exploited in contextual policy search algorithms. In particular, we apply factorization to a Bayesian optimization approach to contextual policy search both in sampling-based and active learning settings. Our simulation results show faster learning and better generalization in various robotic domains. See our supplementary video: https://youtu.be/MNTbBAOufDY.
http://arxiv.org/abs/1904.11761
This work analyses the potential of restarts for probSAT, a quite successful algorithm for k-SAT, by estimating its runtime distributions on random 3-SAT instances that are close to the phase transition. We estimate an optimal restart time from empirical data, reaching a potential speedup factor of 1.39. Calculating restart times from fitted probability distributions reduces this factor to a maximum of 1.30. A spin-off result is that the Weibull distribution approximates the runtime distribution for over 93% of the used instances well. A machine learning pipeline is presented to compute a restart time for a fixed-cutoff strategy to exploit this potential. The main components of the pipeline are a random forest for determining the distribution type and a neural network for the distribution’s parameters. ProbSAT performs statistically significantly better than Luby’s restart strategy and the policy without restarts when using the presented approach. The structure is particularly advantageous on hard problems.
http://arxiv.org/abs/1904.11757
As one type of machine-learning model, a “decision-tree ensemble model” (DTEM) is represented by a set of decision trees. A DTEM is mainly known to be valid for structured data; however, like other machine-learning models, it is difficult to train so that it returns the correct output value for any input value. Accordingly, when a DTEM is used in regard to a system that requires reliability, it is important to comprehensively detect input values that lead to malfunctions of a system (failures) during development and take appropriate measures. One conceivable solution is to install an input filter that controls the input to the DTEM, and to use separate software to process input values that may lead to failures. To develop the input filter, it is necessary to specify the filtering condition of the input value that leads to the malfunction of the system. Given that necessity, in this paper, we propose a method for formally verifying a DTEM and, according to the result of the verification, if an input value leading to a failure is found, extracting the range in which such an input value exists. The proposed method can comprehensively extract the range in which the input value leading to the failure exists; therefore, by creating an input filter based on that range, it is possible to prevent the failure occurring in the system. In this paper, the algorithm of the proposed method is described, and the results of a case study using a dataset of house prices are presented. On the basis of those results, the feasibility of the proposed method is demonstrated, and its scalability is evaluated.
http://arxiv.org/abs/1904.11753
Deep Learning has enabled remarkable progress over the last years on a variety of tasks, such as image recognition, speech recognition, and machine translation. One crucial aspect for this progress are novel neural architectures. Currently employed architectures have mostly been developed manually by human experts, which is a time-consuming and error-prone process. Because of this, there is growing interest in automated neural architecture search methods. We provide an overview of existing work in this field of research and categorize them according to three dimensions: search space, search strategy, and performance estimation strategy.
https://arxiv.org/abs/1808.05377
Transfer learning is widely used in deep neural network models when there are few labeled examples available. The common approach is to take a pre-trained network in a similar task and finetune the model parameters. This is usually done blindly without a pre-selection from a set of pre-trained models, or by finetuning a set of models trained on different tasks and selecting the best performing one by cross-validation. We address this problem by proposing an approach to assess the relationship between visual tasks and their task-specific models. Our method uses Representation Similarity Analysis (RSA), which is commonly used to find a correlation between neuronal responses from brain data and models. With RSA we obtain a similarity score among tasks by computing correlations between models trained on different tasks. Our method is efficient as it requires only pre-trained models, and a few images with no further training. We demonstrate the effectiveness and efficiency of our method for generating task taxonomy on Taskonomy dataset. We next evaluate the relationship of RSA with the transfer learning performance on Taskonomy tasks and a new task: Pascal VOC semantic segmentation. Our results reveal that models trained on tasks with higher similarity score show higher transfer learning performance. Surprisingly, the best transfer learning result for Pascal VOC semantic segmentation is not obtained from the pre-trained model on semantic segmentation, probably due to the domain differences, and our method successfully selects the high performing models.
http://arxiv.org/abs/1904.11740
The task of recognizing goals and plans from missing and full observations can be done efficiently by using automated planning techniques. In many applications, it is important to recognize goals and plans not only accurately, but also quickly. To address this challenge, we develop novel goal recognition approaches based on planning techniques that rely on planning landmarks. In automated planning, landmarks are properties (or actions) that cannot be avoided to achieve a goal. We show the applicability of a number of planning techniques with an emphasis on landmarks for goal and plan recognition tasks in two settings: (1) we use the concept of landmarks to develop goal recognition heuristics; and (2) we develop a landmark-based filtering method to refine existing planning-based goal and plan recognition approaches. These recognition approaches are empirically evaluated in experiments over several classical planning domains. We show that our goal recognition approaches yield not only accuracy comparable to (and often higher than) other state-of-the-art techniques, but also substantially faster recognition time over such techniques.
http://arxiv.org/abs/1904.11739
In this paper, we propose to learn a deep fitting degree scoring network for monocular 3D object detection, which aims to score fitting degree between proposals and object conclusively. Different from most existing monocular frameworks which use tight constraint to get 3D location, our approach achieves high-precision localization through measuring the visual fitting degree between the projected 3D proposals and the object. We first regress the dimension and orientation of the object using an anchor-based method so that a suitable 3D proposal can be constructed. We propose FQNet, which can infer the 3D IoU between the 3D proposals and the object solely based on 2D cues. Therefore, during the detection process, we sample a large number of candidates in the 3D space and project these 3D bounding boxes on 2D image individually. The best candidate can be picked out by simply exploring the spatial overlap between proposals and the object, in the form of the output 3D IoU score of FQNet. Experiments on the KITTI dataset demonstrate the effectiveness of our framework.
http://arxiv.org/abs/1904.12681
Deep learning based knowledge tracing model has been shown to outperform traditional knowledge tracing model without the need for human-engineered features, yet its parameters and representations have long been criticized for not being explainable. In this paper, we propose Deep-IRT which is a synthesis of the item response theory (IRT) model and a knowledge tracing model that is based on the deep neural network architecture called dynamic key-value memory network (DKVMN) to make deep learning based knowledge tracing explainable. Specifically, we use the DKVMN model to process the student’s learning trajectory and estimate the student ability level and the item difficulty level over time. Then, we use the IRT model to estimate the probability that a student will answer an item correctly using the estimated student ability and the item difficulty. Experiments show that the Deep-IRT model retains the performance of the DKVMN model, while it provides a direct psychological interpretation of both students and items.
http://arxiv.org/abs/1904.11738
We introduce PlatonicGAN to discover the 3D structure of an object class from an unstructured collection of 2D images, i.e. neither any relation between the images is available nor additional information about the images is known. The key idea is to train a deep neural network to generate 3D shapes which rendered to images are indistinguishable from ground truth images (for a discriminator) under various camera models (i.e. rendering layers) and camera poses. Discriminating 2D images instead of 3D shapes allows tapping into unstructured 2D photo collections instead of relying on curated (e.g., aligned, annotated, etc.) 3D data sets. To establish constraints between 2D image observation and their 3D interpretation, we suggest a family of rendering layers that are effectively differentiable. This family includes visual hull, absorption-only (akin to x-ray), and emission-absorption. We can successfully reconstruct 3D shapes from unstructured 2D images and extensively evaluate PlatonicGAN on a range of synthetic and real data sets achieving consistent improvements over baseline methods. We can also show that our method with additional 3D supervision further improves result quality and even surpasses the performance of 3D supervised methods.
http://arxiv.org/abs/1811.11606
Assessing whether an agent has abandoned a goal or is actively pursuing it is important when multiple agents are trying to achieve joint goals, or when agents commit to achieving goals for each other. Making such a determination for a single goal by observing only plan traces is not trivial as agents often deviate from optimal plans for various reasons, including the pursuit of multiple goals or the inability to act optimally. In this article, we develop an approach based on domain independent heuristics from automated planning, landmarks, and fact partitions to identify sub-optimal action steps - with respect to a plan - within a plan execution trace. Such capability is very important in domains where multiple agents cooperate and delegate tasks among themselves, e.g. through social commitments, and need to ensure that a delegating agent can infer whether or not another agent is actually progressing towards a delegated task. We demonstrate how an agent can use our technique to determine - by observing a trace - whether an agent is honouring a commitment. We empirically show, for a number of representative domains, that our approach infers sub-optimal action steps with very high accuracy and detects commitment abandonment in nearly all cases.
http://arxiv.org/abs/1904.11737
The fifth-generation cellular networks (5G) has boosted the unprecedented convergence between the information world and physical world. On the other hand, empowered with the enormous amount of data and information, artificial intelligence (AI) has been universally applied and pervasive AI is believed to be an integral part of the future cellular networks (e.g., beyond 5G, B5G). Consequently, benefiting from the advancement in communication technology and AI, we boldly argue that the conditions for collective intelligence (CI) will be mature in the B5G era and CI will emerge among the widely connected beings and things. Afterwards, we introduce a regular language (i.e., the information economy metalanguage) supporting the future communications among agents and augment human intelligence. Meanwhile, we demonstrate the achievement of agents in a simulated scenario where the agents collectively work together to form a pattern through simple indirect communications. Finally, we discuss an anytime universal intelligence test model to evaluate the intelligence level of collective agents.
http://arxiv.org/abs/1905.00719
Aiming to help researchers capture early-stage Product Development (PD) activity, this article presents a new method for digitally capturing prototypes. The motivation for this work is to understand prototyping in the early stages of PD projects, and this article investigates if and how digital capture of physical prototypes can be used for this purpose. In PD case studies, such early-stage prototypes are usually rough and of low-fidelity and are thus often discarded or substantially modified through the projects. Hence, retrospective access to prototypes is a challenge when trying to gather accurate empirical data. To capture the prototypes developed through the early stages of a project, a new method has been developed for digitally capturing physical prototypes through multi-view images, along with metadata describing by who, when and where the prototypes were captured. In this article, one project is shown in detail to demonstrate how this capturing system can gather empirical data for enriching PD case studies on early-stage projects that focus on prototyping for concept generation. The first approach is to use the multi-view images for a qualitative assessment of the projects, which can provide new insights and understanding on various aspects like design decisions, trade-offs and specifications. The second approach is to analyse the metadata provided by the system to give understanding into prototyping patterns in the projects. The analysis of metadata provides insight into prototyping progression, including the frequency of prototyping, which days the project participants are most active, and how the prototyping changes over time.
http://arxiv.org/abs/1905.01950
Nonlinear interactions in the dendritic tree play a key role in neural computation. Nevertheless, modeling frameworks aimed at the construction of large-scale, functional spiking neural networks tend to assume linear, current-based superposition of post-synaptic currents. We extend the theory underlying the Neural Engineering Framework to systematically exploit nonlinear interactions between the local membrane potential and conductance-based synaptic channels as a computational resource. In particular, we demonstrate that even a single passive distal dendritic compartment with AMPA and GABA-A synapses connected to a leaky integrate-and-fire neuron supports the computation of a wide variety of multivariate, bandlimited functions, including the Euclidean norm, controlled shunting, and non-negative multiplication. Our results demonstrate that, for certain operations, the accuracy of dendritic computation is on a par with or even surpasses the accuracy of an additional layer of neurons in the network. These findings allow modelers to construct large-scale models of neurobiological systems that closer approximate network topologies and computational resources available in biology. Our results may inform neuromorphic hardware design and could lead to a better utilization of resources on existing neuromorphic hardware platforms.
http://arxiv.org/abs/1904.11713
Simultaneous localization and mapping (SLAM) has been richly researched in past years particularly with regard to range-based or visual-based sensors. Instead of deploying dedicated devices that use visual features, it is more pragmatic to exploit the radio features to achieve this task, due to their ubiquitous nature and the wide deployment of Wifi wireless network. In this paper, we present a novel approach for crowd-sensing simultaneous localization and radio fingerprint mapping (C-SLAM-RF) in large unknown indoor environments. The proposed system makes use of the received signal strength (RSS) from surrounding Wifi access points (AP) and the motion tracking data from a smart phone (Tango as an example). These measurements are captured duration the walking of multiple users in unknown environments without map information and location of the AP. The experiments were done in a university building with dynamic environment and the results show that the proposed system is capable of estimating the tracks of a group of users with an accuracy of 1.74 meters when compared to the ground truth acquired from a point cloud-based SLAM.
http://arxiv.org/abs/1904.11712
Existing methods for CWS usually rely on a large number of labeled sentences to train word segmentation models, which are expensive and time-consuming to annotate. Luckily, the unlabeled data is usually easy to collect and many high-quality Chinese lexicons are off-the-shelf, both of which can provide useful information for CWS. In this paper, we propose a neural approach for Chinese word segmentation which can exploit both lexicon and unlabeled data. Our approach is based on a variant of posterior regularization algorithm, and the unlabeled data and lexicon are incorporated into model training as indirect supervision by regularizing the prediction space of CWS models. Extensive experiments on multiple benchmark datasets in both in-domain and cross-domain scenarios validate the effectiveness of our approach.
http://arxiv.org/abs/1905.01963
Chinese named entity recognition (CNER) is an important task in Chinese natural language processing field. However, CNER is very challenging since Chinese entity names are highly context-dependent. In addition, Chinese texts lack delimiters to separate words, making it difficult to identify the boundary of entities. Besides, the training data for CNER in many domains is usually insufficient, and annotating enough training data for CNER is very expensive and time-consuming. In this paper, we propose a neural approach for CNER. First, we introduce a CNN-LSTM-CRF neural architecture to capture both local and long-distance contexts for CNER. Second, we propose a unified framework to jointly train CNER and word segmentation models in order to enhance the ability of CNER model in identifying entity boundaries. Third, we introduce an automatic method to generate pseudo labeled samples from existing labeled data which can enrich the training data. Experiments on two benchmark datasets show that our approach can effectively improve the performance of Chinese named entity recognition, especially when training data is insufficient.
http://arxiv.org/abs/1905.01964
High resolution computed tomography (HRCT) is the most important imaging modality for interstitial lung diseases, where the radiologists are interested in identifying certain patterns, and their volumetric and regional distribution. The use of machine learning can assist the radiologists with both these tasks by performing semantic segmentation. In this paper, we propose an interactive annotation-tool for semantic segmentation that assists the radiologist in labeling CT scans. The annotation tool is evaluated by six radiologists and radiology residents classifying healthy lung and reticular pattern i HRCT images. The usability of the system is evaluated with a System Usability Score (SUS) and interaction information from the readers that used the tool for annotating the CT volumes. It was discovered that the experienced usability and how the users interactied with the system differed between the users. A higher SUS-score was given by users that prioritized learning speed over model accuracy and spent less time with manual labeling and instead utilized the suggestions provided by the GUI. An analysis of the annotation variations between the readers show substantial agreement (Cohen’s kappa=0.69) for classification of healthy and affected lung parenchyma in pulmonary fibrosis. The inter-reader variation is a challenge for the definition of ground truth.
http://arxiv.org/abs/1904.11701
We propose the Neural Logic Machine (NLM), a neural-symbolic architecture for both inductive learning and logic reasoning. NLMs exploit the power of both neural networks—as function approximators, and logic programming—as a symbolic processor for objects with properties, relations, logic connectives, and quantifiers. After being trained on small-scale tasks (such as sorting short arrays), NLMs can recover lifted rules, and generalize to large-scale tasks (such as sorting longer arrays). In our experiments, NLMs achieve perfect generalization in a number of tasks, from relational reasoning tasks on the family tree and general graphs, to decision making tasks including sorting arrays, finding shortest paths, and playing the blocks world. Most of these tasks are hard to accomplish for neural networks or inductive logic programming alone.
http://arxiv.org/abs/1904.11694
We propose the Neuro-Symbolic Concept Learner (NS-CL), a model that learns visual concepts, words, and semantic parsing of sentences without explicit supervision on any of them; instead, our model learns by simply looking at images and reading paired questions and answers. Our model builds an object-based scene representation and translates sentences into executable, symbolic programs. To bridge the learning of two modules, we use a neuro-symbolic reasoning module that executes these programs on the latent scene representation. Analogical to human concept learning, the perception module learns visual concepts based on the language description of the object being referred to. Meanwhile, the learned visual concepts facilitate learning new words and parsing new sentences. We use curriculum learning to guide the searching over the large compositional space of images and language. Extensive experiments demonstrate the accuracy and efficiency of our model on learning visual concepts, word representations, and semantic parsing of sentences. Further, our method allows easy generalization to new object attributes, compositions, language concepts, scenes and questions, and even new program domains. It also empowers applications including visual question answering and bidirectional image-text retrieval.
http://arxiv.org/abs/1904.12584
Semantic segmentation has achieved huge progress via adopting deep Fully Convolutional Networks (FCN). However, the performance of FCN based models severely rely on the amounts of pixel-level annotations which are expensive and time-consuming. To address this problem, it is a good choice to learn to segment with weak supervision from bounding boxes. How to make full use of the class-level and region-level supervisions from bounding boxes is the critical challenge for the weakly supervised learning task. In this paper, we first introduce a box-driven class-wise masking model (BCM) to remove irrelevant regions of each class. Moreover, based on the pixel-level segment proposal generated from the bounding box supervision, we could calculate the mean filling rates of each class to serve as an important prior cue, then we propose a filling rate guided adaptive loss (FR-Loss) to help the model ignore the wrongly labeled pixels in proposals. Unlike previous methods directly training models with the fixed individual segment proposals, our method can adjust the model learning with global statistical information. Thus it can help reduce the negative impacts from wrongly labeled proposals. We evaluate the proposed method on the challenging PASCAL VOC 2012 benchmark and compare with other methods. Extensive experimental results show that the proposed method is effective and achieves the state-of-the-art results.
http://arxiv.org/abs/1904.11693
Most deep learning approaches for text-to-SQL generation are limited to the WikiSQL dataset, which only supports very simple queries. Recently, template-based and sequence-to-sequence approaches were proposed to support complex queries, which contain join queries, nested queries, and other types. However, Finegan-Dollak et al. (2018) demonstrated that both the approaches lack the ability to generate SQL of unseen templates. In this paper, we propose a template-based one-shot learning model for the text-to-SQL generation so that the model can generate SQL of an untrained template based on a single example. First, we classify the SQL template using the Matching Network that is augmented by our novel architecture Candidate Search Network. Then, we fill the variable slots in the predicted template using the Pointer Network. We show that our model outperforms state-of-the-art approaches for various text-to-SQL datasets in two aspects: 1) the SQL generation accuracy for the trained templates, and 2) the adaptability to the unseen SQL templates based on a single example without any additional training.
https://arxiv.org/abs/1905.11499
The quality and size of training set have great impact on the results of deep learning-based face related tasks. However, collecting and labeling adequate samples with high quality and balanced distributions still remains a laborious and expensive work, and various data augmentation techniques have thus been widely used to enrich the training dataset. In this paper, we systematically review the existing works of face data augmentation from the perspectives of the transformation types and methods, with the state-of-the-art approaches involved. Among all these approaches, we put the emphasis on the deep learning-based works, especially the generative adversarial networks which have been recognized as more powerful and effective tools in recent years. We present their principles, discuss the results and show their applications as well as limitations. Different evaluation metrics for evaluating these approaches are also introduced. We point out the challenges and opportunities in the field of face data augmentation, and provide brief yet insightful discussions.
http://arxiv.org/abs/1904.11685
Compressing large Neural Networks (NN) by quantizing the parameters, while maintaining the performance is highly desirable due to reduced memory and time complexity. In this work, we cast NN quantization as a discrete labelling problem and leverage results from the extensively studied MRF optimization literature. Specifically, we examine relaxations to the discrete labelling problem, leading to an efficient iterative optimization procedure that involves stochastic gradient descent followed by a projection. We prove that our simple projected gradient descent approach is, in fact, equivalent to a proximal version of the well-known mean-field method. These findings allow the decades-old and theoretically grounded research on MRF optimization to be used to design better network quantization schemes. Our experiments on standard classification datasets (MNIST, CIFAR10/100, TinyImageNet) with convolutional and residual architectures evidence that our algorithm obtains fully-quantized networks with accuracies very close to the floating-point reference networks.
http://arxiv.org/abs/1812.04353
We address the multi-focus image fusion problem, where multiple images captured with different focal settings are to be fused into an all-in-focus image of higher quality. Algorithms for this problem necessarily admit the source image characteristics along with focused and blurred features. However, most sparsity-based approaches use a single dictionary in focused feature space to describe multi-focus images, and ignore the representations in blurred feature space. We propose a multi-focus image fusion approach based on sparse representation using a coupled dictionary. It exploits the observations that the patches from a given training set can be sparsely represented by a couple of overcomplete dictionaries related to the focused and blurred categories of images and that a sparse approximation based on such coupled dictionary leads to a more flexible and therefore better fusion strategy than the one based on just selecting the sparsest representation in the original image estimate. In addition, to improve the fusion performance, we employ a coupled dictionary learning approach that enforces pairwise correlation between atoms of dictionaries learned to represent the focused and blurred feature spaces. We also discuss the advantages of the fusion approach based on coupled dictionary learning, and present efficient algorithms for fusion based on coupled dictionary learning. Extensive experimental comparisons with state-of-the-art multi-focus image fusion algorithms validate the effectiveness of the proposed approach.
http://arxiv.org/abs/1705.10574
The explosive growth of fake news and its erosion of democracy, justice, and public trust has significantly increased the demand for accurate fake news detection. Recent advancements in this area have proposed novel techniques that aim to detect fake news by exploring how it propagates on social networks. However, to achieve fake news early detection, one is only provided with limited to no information on news propagation; hence, motivating the need to develop approaches that can detect fake news by focusing mainly on news content. In this paper, a theory-driven model is proposed for fake news detection. The method investigates news content at various levels: lexicon-level, syntax-level, semantic-level and discourse-level. We represent news at each level, relying on well-established theories in social and forensic psychology. Fake news detection is then conducted within a supervised machine learning framework. As an interdisciplinary research, our work explores potential fake news patterns, enhances the interpretability in fake news feature engineering, and studies the relationships among fake news, deception/disinformation, and clickbaits. Experiments conducted on two real-world datasets indicate that the proposed method can outperform the state-of-the-art and enable fake news early detection, even when there is limited content information.
http://arxiv.org/abs/1904.11679
Intrinsically, driving is a Markov Decision Process which suits well the reinforcement learning paradigm. In this paper, we propose a novel agent which learns to drive a vehicle without any human assistance. We use the concept of reinforcement learning and evolutionary strategies to train our agent in a 2D simulation environment. Our model’s architecture goes beyond the World Model’s by introducing difference images in the auto encoder. This novel involvement of difference images in the auto-encoder gives better representation of the latent space with respect to the motion of vehicle and helps an autonomous agent to learn more efficiently how to drive a vehicle. Results show that our method requires fewer (96% less) total agents, (87.5% less) agents per generations, (70% less) generations and (90% less) rollouts than the original architecture while achieving the same accuracy of the original.
http://arxiv.org/abs/1904.12738
We consider thyroid-malignancy prediction from ultra-high-resolution whole-slide cytopathology images. We propose a deep-learning-based algorithm that is inspired by the way a cytopathologist diagnoses the slides. The algorithm identifies diagnostically relevant image regions and assigns them local malignancy scores, that in turn are incorporated into a global malignancy prediction. We discuss the relation of our deep-learning-based approach to multiple-instance learning (MIL) and describe how it deviates from classical MIL methods by the use of a supervised procedure to extract relevant regions from the whole-slide. The analysis of our algorithm further reveals a close relation to hypothesis testing, which, along with unique characteristics of thyroid cytopathology, allows us to devise an improved training strategy. We further propose an ordinal regression framework for the simultaneous prediction of thyroid malignancy and an ordered diagnostic score acting as a regularizer, which further improves the predictions of the network. Experimental results demonstrate that the proposed algorithm outperforms several competing methods, achieving performance comparable to human experts.
http://arxiv.org/abs/1904.12739
The recent success of transformer networks for neural machine translation and other NLP tasks has led to a surge in research work trying to apply it for speech recognition. Recent efforts studied key research questions around ways of combining positional embedding with speech features, and stability of optimization for large scale learning of transformer networks. In this paper, we propose replacing the sinusoidal positional embedding for transformers with convolutionally learned input representations. These contextual representations provide subsequent transformer blocks with relative positional information needed for discovering long-range relationships between local concepts. The proposed system has favorable optimization characteristics where our reported results are produced with fixed learning rate of 1.0 and no warmup steps. The proposed model reduces the word error rate (WER) by 12% and 16% relative to previously published work on Librispeech “dev other” and “test other” subsets respectively, when no extra LM text is provided. Full code to reproduce our results will be available online at the time of publication.
http://arxiv.org/abs/1904.11660
Action recognition with skeleton data has recently attracted much attention in computer vision. Previous studies are mostly based on fixed skeleton graphs, only capturing local physical dependencies among joints, which may miss implicit joint correlations. To capture richer dependencies, we introduce an encoder-decoder structure, called A-link inference module, to capture action-specific latent dependencies, i.e. actional links, directly from actions. We also extend the existing skeleton graphs to represent higher-order dependencies, i.e. structural links. Combing the two types of links into a generalized skeleton graph, we further propose the actional-structural graph convolution network (AS-GCN), which stacks actional-structural graph convolution and temporal convolution as a basic building block, to learn both spatial and temporal features for action recognition. A future pose prediction head is added in parallel to the recognition head to help capture more detailed action patterns through self-supervision. We validate AS-GCN in action recognition using two skeleton data sets, NTU-RGB+D and Kinetics. The proposed AS-GCN achieves consistently large improvement compared to the state-of-the-art methods. As a side product, AS-GCN also shows promising results for future pose prediction.
http://arxiv.org/abs/1904.12659
Automatic measuring of speaker sincerity degree is a novel research problem in computational paralinguistics. This paper proposes covariance-based feature vectors to model speech and ensembles of support vector regressors to estimate the degree of sincerity of a speaker. The elements of each covariance vector are pairwise statistics between the short-term feature components. These features are used alone as well as in combination with the ComParE acoustic feature set. The experimental results on the development set of the Sincerity Speech Corpus using a cross-validation procedure have shown an 8.1% relative improvement in the Spearman’s correlation coefficient over the baseline system.
http://arxiv.org/abs/1904.11641
We propose a method of using a Weighted second-order cone programming twin support vector machine (WSOCP-TWSVM) for imbalanced data classification. This method constructs a graph based under-sampling method which is utilized to remove outliers and reduce the dispensable majority samples. Then, appropriate weights are set in order to decrease the impact of samples of the majority class and increase the effect of the minority class in the optimization formula of the classifier. These weights are embedded in the optimization problem of the Second Order Cone Programming (SOCP) Twin Support Vector Machine formulations. This method is tested, and its performance is compared to previous methods on standard datasets. Results of experiments confirm the feasibility and efficiency of the proposed method.
http://arxiv.org/abs/1904.11634
Rule-based machine translation is more data efficient than the big data-based machine translation approaches, making it appropriate for languages with low bilingual corpus resources – i.e., minority languages. However, the rule-based approach has declined in popularity relative to its big data cousins primarily because of the extensive training and labour required to define the language rules. To address this, we present a semantic representation that 1) treats all bits of meaning as individual concepts that 2) modify or further specify one another to build a network that relates entities in space and time. Also, the representation can 3) encapsulate propositions and thereby define concepts in terms of other concepts, supporting the abstraction of underlying linguistic and ontological details. These features afford an exact, yet intuitive semantic representation aimed at handling the great variety in language and reducing labour and training time. The proposed natural language generation, parsing, and translation strategies are also amenable to probabilistic modeling and thus to learning the necessary rules from example data.
http://arxiv.org/abs/1807.02226
In this paper, we consider the problem of choosing the optimal scenario of the impact between nodes based on of the introduced criteria for the optimality of the impact. Two criteria for the optimality of the impact, which are called the force of impact and the speed of implementation of the scenario, are considered. To obtain a unique solution of the problem, a multi-criterial assessment of the received scenarios using the Pareto principle was applied. Based on the criteria of a force of impact and the speed of implementation of the scenario, the choice of the optimal scenario of impact was justified. The results and advantages of the proposed approach in comparison with the Kosko model are presented.
http://arxiv.org/abs/1904.13308
We propose a novel way of reducing the number of parameters in the storage-hungry fully connected layers of a neural network by using pre-defined sparsity, where the majority of connections are absent prior to starting training. Our results indicate that convolutional neural networks can operate without any loss of accuracy at less than half percent classification layer connection density, or less than 5 percent overall network connection density. We also investigate the effects of pre-defining the sparsity of networks with only fully connected layers. Based on our sparsifying technique, we introduce the `scatter’ metric to characterize the quality of a particular connection pattern. As proof of concept, we show results on CIFAR, MNIST and a new dataset on classifying Morse code symbols, which highlights some interesting trends and limits of sparse connection patterns.
http://arxiv.org/abs/1711.02131
Learning speaker-specific features is vital in many applications like speaker recognition, diarization and speech recognition. This paper provides a novel approach, we term Neural Predictive Coding (NPC), to learn speaker-specific characteristics in a completely unsupervised manner from large amounts of unlabeled training data that even contain many non-speech events and multi-speaker audio streams. The NPC framework exploits the proposed short-term active-speaker stationarity hypothesis which assumes two temporally-close short speech segments belong to the same speaker, and thus a common representation that can encode the commonalities of both the segments, should capture the vocal characteristics of that speaker. We train a convolutional deep siamese network to produce “speaker embeddings” by learning to separate same' vs different’ speaker pairs which are generated from an unlabeled data of audio streams. Two sets of experiments are done in different scenarios to evaluate the strength of NPC embeddings and compare with state-of-the-art in-domain supervised methods. First, two speaker identification experiments with different context lengths are performed in a scenario with comparatively limited within-speaker channel variability. NPC embeddings are found to perform the best at short duration experiment, and they provide complementary information to i-vectors for full utterance experiments. Second, a large scale speaker verification task having a wide range of within-speaker channel variability is adopted as an upper-bound experiment where comparisons are drawn with in-domain supervised methods.
http://arxiv.org/abs/1802.07860
Visual knowledge bases such as Visual Genome power numerous applications in computer vision, including visual question answering and captioning, but suffer from sparse, incomplete relationships. All scene graph models to date are limited to training on a small set of visual relationships that have thousands of training labels each. Hiring human annotators is expensive, and using textual knowledge base completion methods are incompatible with visual data. In this paper, we introduce a semi-supervised method that assigns probabilistic relationship labels to a large number of unlabeled images using few labeled examples. We analyze visual relationships to suggest two types of image-agnostic features that are used to generate noisy heuristics, whose outputs are aggregated using a factor graph-based generative model. With as few as 10 labeled relationship examples, the generative model creates enough training data to train any existing state-of-the-art scene graph model. We demonstrate that our method for generating training data outperforms all baseline approaches by 5.16 recall@100. Since we only use a few labels, we define a complexity metric for relationships that serves as an indicator (R^2 = 0.778) for conditions under which our method succeeds over transfer learning, the de-facto approach for training with limited labels.
http://arxiv.org/abs/1904.11622
Training models to high-end performance requires availability of large labeled datasets, which are expensive to get. The goal of our work is to automatically synthesize labeled datasets that are relevant for a downstream task. We propose Meta-Sim, which learns a generative model of synthetic scenes, and obtain images as well as its corresponding ground-truth via a graphics engine. We parametrize our dataset generator with a neural network, which learns to modify attributes of scene graphs obtained from probabilistic scene grammars, so as to minimize the distribution gap between its rendered outputs and target data. If the real dataset comes with a small labeled validation set, we additionally aim to optimize a meta-objective, i.e. downstream task performance. Experiments show that the proposed method can greatly improve content generation quality over a human-engineered probabilistic scene grammar, both qualitatively and quantitatively as measured by performance on a downstream task.
http://arxiv.org/abs/1904.11621
Infrared (IR) images are essential to improve the visibility of dark or camouflaged objects. Object recognition and segmentation based on a neural network using IR images provide more accuracy and insight than color visible images. But the bottleneck is the amount of relevant IR images for training. It is difficult to collect real-world IR images for special purposes, including space exploration, military and fire-fighting applications. To solve this problem, we created color visible and IR images using a Unity-based 3D game editor. These synthetically generated color visible and IR images were used to train cycle consistent adversarial networks (CycleGAN) to convert visible images to IR images. CycleGAN has the advantage that it does not require precisely matching visible and IR pairs for transformation training. In this study, we discovered that additional synthetic data can help improve CycleGAN performance. Neural network training using real data (N = 20) performed more accurate transformations than training using real (N = 10) and synthetic (N = 10) data combinations. The result indicates that the synthetic data cannot exceed the quality of the real data. Neural network training using real (N = 10) and synthetic (N = 100) data combinations showed almost the same performance as training using real data (N = 20). At least 10 times more synthetic data than real data is required to achieve the same performance. In summary, CycleGAN is used with synthetic data to improve the IR image conversion performance of visible images.
http://arxiv.org/abs/1904.11620
It is important to find the target as soon as possible for search and rescue operations. Surveillance camera systems and unmanned aerial vehicles (UAVs) are used to support search and rescue. Automatic object detection is important because a person cannot monitor multiple surveillance screens simultaneously for 24 hours. Also, the object is often too small to be recognized by the human eye on the surveillance screen. This study used UAVs around the Port of Houston and fixed surveillance cameras to build an automatic target detection system that supports the US Coast Guard (USCG) to help find targets (e.g., person overboard). We combined image segmentation, enhancement, and convolution neural networks to reduce detection time to detect small targets. We compared the performance between the auto-detection system and the human eye. Our system detected the target within 8 seconds, but the human eye detected the target within 25 seconds. Our systems also used synthetic data generation and data augmentation techniques to improve target detection accuracy. This solution may help the search and rescue operations of the first responders in a timely manner.
http://arxiv.org/abs/1904.11619
Photorealistic style transfer aims to transfer the style of one image to another, but preserves the original structure and detail outline of the content image, which makes the content image still look like a real shot after the style transfer. Although some realistic image styling methods have been proposed, these methods are vulnerable to lose the details of the content image and produce some irregular distortion structures. In this paper, we use a high-resolution network as the image generation network. Compared to other methods, which reduce the resolution and then restore the high resolution, our generation network maintains high resolution throughout the process. By connecting high-resolution subnets to low-resolution subnets in parallel and repeatedly multi-scale fusion, high-resolution subnets can continuously receive information from low-resolution subnets. This allows our network to discard less information contained in the image, so the generated images may have a more elaborate structure and less distortion, which is crucial to the visual quality. We conducted extensive experiments and compared the results with existing methods. The experimental results show that our model is effective and produces better results than existing methods for photorealistic image stylization. Our source code with PyTorch framework will be publicly available at https://github.com/limingcv/Photorealistic-Style-Transfer
http://arxiv.org/abs/1904.11617
Deep neural networks, although shown to be a successful class of machine learning algorithms, are known to be extremely unstable to adversarial perturbations. Improving the robustness of neural networks against these attacks is important, especially for security-critical applications. To defend against such attacks, we propose dividing the input image into multiple patches, denoising each patch independently, and reconstructing the image, without losing significant image content. We call our method D3. This proposed defense mechanism is non-differentiable which makes it non-trivial for an adversary to apply gradient-based attacks. Moreover, we do not fine-tune the network with adversarial examples, making it more robust against unknown attacks. We present an analysis of the tradeoff between accuracy and robustness against adversarial attacks. We evaluate our method under black-box, grey-box, and white-box settings. On the ImageNet dataset, our method outperforms the state-of-the-art by 19.7% under grey-box setting, and performs comparably under black-box setting. For the white-box setting, the proposed method achieves 34.4% accuracy compared to the 0% reported in the recent works.
http://arxiv.org/abs/1802.06806
We examine a large dialog corpus obtained from the conversation history of a single individual with 104 conversation partners. The corpus consists of half a million instant messages, across several messaging platforms. We focus our analyses on seven speaker attributes, each of which partitions the set of speakers, namely: gender; relative age; family member; romantic partner; classmate; co-worker; and native to the same country. In addition to the content of the messages, we examine conversational aspects such as the time messages are sent, messaging frequency, psycholinguistic word categories, linguistic mirroring, and graph-based features reflecting how people in the corpus mention each other. We present two sets of experiments predicting each attribute using (1) short context windows; and (2) a larger set of messages. We find that using all features leads to gains of 9-14% over using message text only.
http://arxiv.org/abs/1904.11610
We present DeepPerimeter, a deep learning based pipeline for inferring a full indoor perimeter (i.e. exterior boundary map) from a sequence of posed RGB images. Our method relies on robust deep methods for depth estimation and wall segmentation to generate an exterior boundary point cloud, and then uses deep unsupervised clustering to fit wall planes to obtain a final boundary map of the room. We demonstrate that DeepPerimeter results in excellent visual and quantitative performance on the popular ScanNet and FloorNet datasets and works for room shapes of various complexities as well as in multiroom scenarios. We also establish important baselines for future work on indoor perimeter estimation, topics which will become increasingly prevalent as application areas like augmented reality and robotics become more significant.
http://arxiv.org/abs/1904.11595
Optical flow techniques are becoming increasingly performant and robust when estimating motion in a scene, but their performance has yet to be proven in the area of facial expression recognition. In this work, a variety of optical flow approaches are evaluated across multiple facial expression datasets, so as to provide a consistent performance evaluation. Additionally, the strengths of multiple optical flow approaches are combined in a novel data augmentation scheme. Under this scheme, increases in average accuracy of up to 6% (depending on the choice of optical flow approaches and dataset) have been achieved.
http://arxiv.org/abs/1904.11592
Dark Channel Prior (DCP) is a widely recognized traditional dehazing algorithm. However, it may fail in bright region and the brightness of the restored image is darker than hazy image. In this paper, we propose an effective method to optimize DCP. We build a multiple linear regression haze-removal model based on DCP atmospheric scattering model and train this model with RESIDE dataset, which aims to reduce the unexpected errors caused by the rough estimations of transmission map t(x) and atmospheric light A. The RESIDE dataset provides enough synthetic hazy images and their corresponding groundtruth images to train and test. We compare the performances of different dehazing algorithms in terms of two important full-reference metrics, the peak-signal-to-noise ratio (PSNR) as well as the structural similarity index measure (SSIM). The experiment results show that our model gets highest SSIM value and its PSNR value is also higher than most of state-of-the-art dehazing algorithms. Our results also overcome the weakness of DCP on real-world hazy images
http://arxiv.org/abs/1904.11587
We present the task of Spatio-Temporal Video Question Answering, which requires intelligent systems to simultaneously retrieve relevant moments and detect referenced visual concepts (people and objects) to answer natural language questions about videos. We first augment the TVQA dataset with 310.8k bounding boxes, linking depicted objects to visual concepts in questions and answers. We name this augmented version as TVQA+. We then propose Spatio-Temporal Answerer with Grounded Evidence (STAGE), a unified framework that grounds evidence in both the spatial and temporal domains to answer questions about videos. Comprehensive experiments and analyses demonstrate the effectiveness of our framework and how the rich annotations in our TVQA+ dataset can contribute to the question answering task. As a side product, by performing this joint task, our model is able to produce more insightful intermediate results. Dataset and code are publicly available.
http://arxiv.org/abs/1904.11574
For many applications such as action detection or robotic interaction, segmenting all moving objects is a crucial first step. While this problem has been well-studied under the formulation of spatiotemporal video segmentation, virtually none of the prior works use learning-based approaches, despite significant advances in single-frame instance segmentation. We propose the first deep-learning based approach for spatio-temporal grouping. Our model extends the state-of-the-art Mask R-CNN architecture to the video domain. It takes a video frame together with its optical flow as input, and passes them through appearance and motion streams respectively. It then combines the motion cues, which provide a bottom-up signal for object detection, with appearance cues that allow capturing the full extent of the object via a joint RPN module. We show state-of-the-art results on the Freiburg Berkeley Motion Segmentation dataset by a wide margin. One potential worry with learning-based methods is that they might overfit to the particular type of objects that they have been trained on. While current recognition systems tend to be limited to a “closed world” of N objects on which they are trained, our model can segment almost anything that moves.
http://arxiv.org/abs/1902.03715
The tasks that an agent will need to solve often are not known during training. However, if the agent knows which properties of the environment are important then, after learning how its actions affect those properties, it may be able to use this knowledge to solve complex tasks without training specifically for them. Towards this end, we consider a setup in which an environment is augmented with a set of user defined attributes that parameterize the features of interest. We propose a method that learns a policy for transitioning between “nearby” sets of attributes, and maintains a graph of possible transitions. Given a task at test time that can be expressed in terms of a target set of attributes, and a current state, our model infers the attributes of the current state and searches over paths through attribute space to get a high level plan, and then uses its low level policy to execute the plan. We show in 3D block stacking, grid-world games, and StarCraft that our model is able to generalize to longer, more complex tasks at test time by composing simpler learned policies.
http://arxiv.org/abs/1803.00512