In this note we deal with the detection of oil spills in open sea via self similar, long range dependence random fields and wavelet filters. We show some preliminary experimental results of our technique with Sentinel 1 SAR images.
http://arxiv.org/abs/1903.03221
We present a real-time algorithm for emotion-aware navigation of a robot among pedestrians. Our approach estimates time-varying emotional behaviors of pedestrians from their faces and trajectories using a combination of Bayesian-inference, CNN-based learning, and the PAD (Pleasure-Arousal-Dominance) model from psychology. These PAD characteristics are used for long-term path prediction and generating proxemic constraints for each pedestrian. We use a multi-channel model to classify pedestrian characteristics into four emotion categories (happy, sad, angry, neutral). In our validation results, we observe an emotion detection accuracy of 85.33%. We formulate emotion-based proxemic constraints to perform socially-aware robot navigation in low- to medium-density environments. We demonstrate the benefits of our algorithm in simulated environments with tens of pedestrians as well as in a real-world setting with Pepper, a social humanoid robot.
http://arxiv.org/abs/1903.03217
Heterogeneous knowledge naturally arises among different agents in cooperative multiagent reinforcement learning. As such, learning can be greatly improved if agents can effectively pass their knowledge on to other agents. Existing work has demonstrated that peer-to-peer knowledge transfer, a process referred to as action advising, improves team-wide learning. In contrast to previous frameworks that advise at the level of primitive actions, we aim to learn high-level teaching policies that decide when and what high-level action (e.g., sub-goal) to advise a teammate. We introduce a new learning to teach framework, called hierarchical multiagent teaching (HMAT). The proposed framework solves difficulties faced by prior work on multiagent teaching when operating in domains with long horizons, delayed rewards, and continuous states/actions by leveraging temporal abstraction and deep function approximation. Our empirical evaluations show that HMAT accelerates team-wide learning progress in difficult environments that are more complex than those explored in previous work. HMAT also learns teaching policies that can be transferred to different teammates/tasks and can even teach teammates with heterogeneous action spaces.
http://arxiv.org/abs/1903.03216
A classifier trained on a dataset seldom works on other datasets obtained under different conditions due to domain shift. This problem is commonly addressed by domain adaptation methods. In this work we introduce a novel deep learning framework which unifies different paradigms in unsupervised domain adaptation. Specifically, we propose domain alignment layers which implement feature whitening for the purpose of matching source and target feature distributions. Additionally, we leverage the unlabeled target data by proposing the Min-Entropy Consensus loss, which regularizes training while avoiding the adoption of many user-defined hyper-parameters. We report results on publicly available datasets, considering both digit classification and object recognition tasks. We show that, in most of our experiments, our approach improves upon previous methods, setting new state-of-the-art performances.
http://arxiv.org/abs/1903.03215
This paper proposes a bandwidth tunable technique for real-time probabilistic scene modeling and mapping to enable co-robotic exploration in communication constrained environments such as the deep sea. The parameters of the system enable the user to characterize the scene complexity represented by the map, which in turn determines the bandwidth requirements. The approach is demonstrated using an underwater robot that learns an unsupervised scene model of the environment and then uses this scene model to communicate the spatial distribution of various high-level semantic scene constructs to a human operator. Preliminary experiments in an artificially constructed tank environment as well as simulated missions over a 10m$\times$10m coral reef using real data show the tunability of the maps to different bandwidth constraints and science interests. To our knowledge this is the first paper to quantify how the free parameters of the unsupervised scene model impact both the scientific utility of and bandwidth required to communicate the resulting scene model.
http://arxiv.org/abs/1903.03214
In this paper, multiple instance learning (MIL) algorithms to automatically perform root detection and segmentation in minirhizotron imagery using only image-level labels are proposed. Root and soil characteristics vary from location to location, thus, supervised machine learning approaches that are trained with local data provide the best ability to identify and segment roots in minirhizotron imagery. However, labeling roots for training data (or otherwise) is an extremely tedious and time-consuming task. This paper aims to address this problem by labeling data at the image level (rather than the individual root or root pixel level) and train algorithms to perform individual root pixel level segmentation using MIL strategies. Three MIL methods (MI-ACE, miSVM, MIForests) were applied to root detection and compared to non-MIL approches. The results show that MIL methods improve root segmentation in challenging minirhizotron imagery and reduce the labeling burden. In our results, miSVM outperformed other methods. The MI-ACE algorithm was a close second with an added advantage that it learned an interpretable root signature which identified the traits used to distinguish roots from soil and did not require parameter selection.
http://arxiv.org/abs/1903.03207
Three-way decision theory, which trisects the universe with less risks or costs, is considered as a powerful mathematical tool for handling uncertainty in incomplete and imprecise information tables, and provides an effective tool for conflict analysis decision making in real-time situations. In this paper, we propose the concepts of the agreement, disagreement and neutral subsets of a strategy with two evaluation functions, which establish the three-way decisions-based conflict analysis models(TWDCAMs) for trisecting the universe of agents, and employ a pair of two-way decisions models to interpret the mechanism of the three-way decision rules for an agent. Subsequently, we develop the concepts of the agreement, disagreement and neutral strategies of an agent group with two evaluation functions, which build the TWDCAMs for trisecting the universe of issues, and take a couple of two-way decisions models to explain the mechanism of the three-way decision rules for an issue. Finally, we reconstruct Fan, Qi and Wei’s conflict analysis models(FQWCAMs) and Sun, Ma and Zhao’s conflict analysis models(SMZCAMs) with two evaluation functions, and interpret FQWCAMs and SMZCAMs with a pair of two-day decisions models, which illustrates that FQWCAMs and SMZCAMs are special cases of TWDCAMs.
http://arxiv.org/abs/1903.03205
Noise pollution is one of the topmost quality of life issues for urban residents in the United States. Continued exposure to high levels of noise has proven effects on health, including acute effects such as sleep disruption, and long-term effects such as hypertension, heart disease, and hearing loss. To investigate and ultimately aid in the mitigation of urban noise, a network of 55 sensor nodes has been deployed across New York City for over two years, collecting sound pressure level (SPL) and audio data. This network has cumulatively amassed over 75 years of calibrated, high-resolution SPL measurements and 35 years of audio data. In addition, high frequency telemetry data has been collected that provides an indication of a sensors’ health. This telemetry data was analyzed over an 18 month period across 31 of the sensors. It has been used to develop a prototype model for pre-failure detection which has the ability to identify sensors in a prefail state 69.1% of the time. The entire network infrastructure is outlined, including the operation of the sensors, followed by an analysis of its data yield and the development of the fault detection approach and the future system integration plans for this.
http://arxiv.org/abs/1903.03195
This paper develops a path planner that minimizes risk (e.g. motion execution) while maximizing accumulated reward (e.g., quality of sensor viewpoint) motivated by visual assistance or tracking scenarios in unstructured or confined environments. In these scenarios, the robot should maintain the best viewpoint as it moves to the goal. However, in unstructured or confined environments, some paths may increase the risk of collision; therefore there is a tradeoff between risk and reward. Conventional state-dependent risk or probabilistic uncertainty modeling do not consider path-level risk or is difficult to acquire. This risk-reward planner explicitly represents risk as a function of motion plans, i.e., paths. Without manual assignment of the negative impact to the planner caused by risk, this planner takes in a pre-established viewpoint quality map and plans target location and path leading to it simultaneously, in order to maximize overall reward along the entire path while minimizing risk. Exact and approximate algorithms are presented, whose solution is further demonstrated on a physical tethered aerial vehicle. Other than the visual assistance problem, the proposed framework also provides a new planning paradigm to address minimum-risk planning under dynamical risk and absence of substructure optimality and to balance the trade-off between reward and risk.
http://arxiv.org/abs/1903.03187
We describe an efficient implementation of clause guidance in saturation-based automated theorem provers extending the ENIGMA approach. Unlike in the first ENIGMA implementation where fast linear classifier is trained and used together with manually engineered features, we have started to experiment with more sophisticated state-of-the-art machine learning methods such as gradient boosted trees and recursive neural networks. In particular the latter approach poses challenges in terms of efficiency of clause evaluation, however, we show that deep integration of the neural evaluation with the ATP data-structures can largely amortize this cost and lead to competitive real-time results. Both methods are evaluated on a large dataset of theorem proving problems and compared with the previous approaches. The resulting methods improve on the manually designed clause guidance, providing the first practically convincing application of gradient-boosted and neural clause guidance in saturation-style automated theorem provers.
http://arxiv.org/abs/1903.03182
Seam carving is a state-of-the-art content-aware image resizing technique that effectively preserves the salient areas of an image. However, when applied to video retargeting, not only is it time intensive, but it also creates highly visible frame-wise discontinuities. In this paper, we propose a novel video retargeting method based on seam carving. First, for a single frame, we locate and remove several seams instead of one seam at once. Second, we use a dynamic spatiotemporal buffer of energy maps and a standard deviation operator to carve out the same seams in a temporal cube of frames with low variation in energy. Last but not least, an improved energy function that considers motions detected through difference method is employed. During testing, these enhancements result in a 93 percent reduction in processing time and a higher frame-wise consistency, thus showing superior performance compared to existing video retargeting methods.
http://arxiv.org/abs/1903.03180
The Arcade Learning Environment (ALE) is a popular platform for evaluating reinforcement learning agents. Much of the appeal comes from the fact that Atari games are varied, showcase aspects of competency we expect from an intelligent agent, and are not biased towards any particular solution approach. The challenge of the ALE includes 1) the representation learning problem of extracting pertinent information from the raw pixels, and 2) the behavioural learning problem of leveraging complex, delayed associations between actions and rewards. Often, in reinforcement learning research, we care more about the latter, but the representation learning problem adds significant computational expense. In response, we introduce MinAtar, short for miniature Atari, a new evaluation platform that captures the general mechanics of specific Atari games, while simplifying certain aspects. In particular, we reduce the representational complexity to focus more on behavioural challenges. MinAtar consists of analogues to five Atari games which play out on a 10x10 grid. MinAtar provides a 10x10xn state representation. The n channels correspond to game-specific objects, such as ball, paddle and brick in the game Breakout. While significantly simplified, these domains are still rich enough to allow for interesting behaviours. To demonstrate the challenges posed by these domains, we evaluated a smaller version of the DQN architecture. We also tried variants of DQN without experience replay, and without a target network, to assess the impact of those two prominent components in the MinAtar environments. In addition, we evaluated a simpler agent that used actor-critic with eligibility traces, online updating, and no experience replay. We hope that by introducing a set of simplified, Atari-like games we can allow researchers to more efficiently investigate the unique behavioural challenges provided by the ALE.
http://arxiv.org/abs/1903.03176
Visual Dialog is a multimodal task of answering a sequence of questions grounded in an image, using the conversation history as context. It entails challenges in vision, language, reasoning, and grounding. However, studying these subtasks in isolation on large, real datasets is infeasible as it requires prohibitively-expensive complete annotation of the ‘state’ of all images and dialogs. We develop CLEVR-Dialog, a large diagnostic dataset for studying multi-round reasoning in visual dialog. Specifically, we construct a dialog grammar that is grounded in the scene graphs of the images from the CLEVR dataset. This combination results in a dataset where all aspects of the visual dialog are fully annotated. In total, CLEVR-Dialog contains 5 instances of 10-round dialogs for about 85k CLEVR images, totaling to 4.25M question-answer pairs. We use CLEVR-Dialog to benchmark performance of standard visual dialog models; in particular, on visual coreference resolution (as a function of the coreference distance). This is the first analysis of its kind for visual dialog models that was not possible without this dataset. We hope the findings from CLEVR-Dialog will help inform the development of future models for visual dialog. Our dataset and code will be made public.
http://arxiv.org/abs/1903.03166
We propose a novel deep convolutional neural network (CNN) based multi-task learning approach for open-set visual recognition. We combine a classifier network and a decoder network with a shared feature extractor network within a multi-task learning framework. We show that this approach results in better open-set recognition accuracy. In our approach, reconstruction errors from the decoder network are utilized for open-set rejection. In addition, we model the tail of the reconstruction error distribution from the known classes using the statistical Extreme Value Theory to improve the overall performance. Experiments on multiple image classification datasets are performed and it is shown that this method can perform significantly better than many competitive open set recognition algorithms available in the literature. The code will be made available at: github.com/otkupjnoz/mlosr.
http://arxiv.org/abs/1903.03161
Hand-held haptic devices can allow for greater freedom of motion and larger workspaces than traditional grounded haptic devices. They can also provide more compelling haptic sensations to the users’ fingertips than many wearable haptic devices because reaction forces can be distributed over a larger area of skin far away from the stimulation site. This paper presents a hand-held kinesthetic gripper that provides guidance cues in four degrees of freedom (DOF). 2-DOF tangential forces on the thumb and index finger combine to create cues to translate or rotate the hand. We demonstrate the device’s capabilities in a three-part user study. First, users moved their hands in response to haptic cues before receiving instruction or training. Then, they trained on cues in eight directions in a forced-choice task. Finally, they repeated the first part, now knowing what each cue intended to convey. Users were able to discriminate each cue over 90% of the time. Users moved correctly in response to the guidance cues both before and after the training and indicated that the cues were easy to follow. The results show promise for holdable kinesthetic devices in haptic feedback and guidance for applications such as virtual reality, medical training, and teleoperation.
http://arxiv.org/abs/1903.03150
We consider the problem of segmenting a biomedical image into anatomical regions of interest. We specifically address the frequent scenario where we have no paired training data that contains images and their manual segmentations. Instead, we employ unpaired segmentation images to build an anatomical prior. Critically these segmentations can be derived from imaging data from a different dataset and imaging modality than the current task. We introduce a generative probabilistic model that employs the learned prior through a convolutional neural network to compute segmentations in an unsupervised setting. We conducted an empirical analysis of the proposed approach in the context of structural brain MRI segmentation, using a multi-study dataset of more than 14,000 scans. Our results show that an anatomical prior can enable fast unsupervised segmentation which is typically not possible using standard convolutional networks. The integration of anatomical priors can facilitate CNN-based anatomical segmentation in a range of novel clinical problems, where few or no annotations are available and thus standard networks are not trainable. The code is freely available at this http URL
http://arxiv.org/abs/1903.03148
k-Nearest Neighbors is one of the most fundamental but effective classification models. In this paper, we propose two families of models built on a sequence to sequence model and a memory network model to mimic the k-Nearest Neighbors model, which generate a sequence of labels, a sequence of out-of-sample feature vectors and a final label for classification, and thus they could also function as oversamplers. We also propose ‘out-of-core’ versions of our models which assume that only a small portion of data can be loaded into memory. Computational experiments show that our models outperform k-Nearest Neighbors, a feed-forward neural network and a memory network, due to the fact that our models must produce additional output and not just the label. As an oversample on imbalanced datasets, the sequence to sequence kNN model often outperforms Synthetic Minority Over-sampling Technique and Adaptive Synthetic Sampling.
https://arxiv.org/abs/1804.11214
In this paper we explore tying together the ideas from Scattering Transforms and Convolutional Neural Networks (CNN) for Image Analysis by proposing a learnable ScatterNet. Previous attempts at tying them together in hybrid networks have tended to keep the two parts separate, with the ScatterNet forming a fixed front end and a CNN forming a learned backend. We instead look at adding learning between scattering orders, as well as adding learned layers before the ScatterNet. We do this by breaking down the scattering orders into single convolutional-like layers we call ‘locally invariant’ layers, and adding a learned mixing term to this layer. Our experiments show that these locally invariant layers can improve accuracy when added to either a CNN or a ScatterNet. We also discover some surprising results in that the ScatterNet may be best positioned after one or more layers of learning rather than at the front of a neural network.
http://arxiv.org/abs/1903.03137
In this paper, we explore the use of motion for robot-to-human communication on three robotic platforms: the 5 degrees-of-freedom (DOF) Aqua autonomous underwater vehicle (AUV), a 3-DOF camera gimbal mounted on a Matrice 100 drone, and a 3-DOF Turtlebot2 terrestrial robot. While we previously explored the use of body language-like motion (called kinemes) versus other methods of communication for the Aqua AUV, we now extend those concepts to robots in two new and different domains. We evaluate all three platforms using a small interaction study where participants use gestures to communicate with the robot, receive information from the robot via kinemes, and then take actions based on the information. To compare the three domains we consider the accuracy of these interactions, the time it takes to complete them, and how confident users feel in the success of their interactions. The kineme systems perform with reasonable accuracy for all robots and experience gained in this study is used to form a set of prescriptions for further development of kineme systems.
http://arxiv.org/abs/1903.03134
We discuss the geometry of rational maps from a projective space of an arbitrary dimension to the product of projective spaces of lower dimensions induced by linear projections. In particular, we give a purely algebro-geometric proof of the projective reconstruction theorem by Hartley and Schaffalitzky [HS09].
http://arxiv.org/abs/1710.06205
We study lifted weight learning of Markov logic networks. We show that there is an algorithm for maximum-likelihood learning of 2-variable Markov logic networks which runs in time polynomial in the domain size. Our results are based on existing lifted-inference algorithms and recent algorithmic results on computing maximum entropy distributions.
http://arxiv.org/abs/1903.03099
We introduce a large scale crowdsourced text adventure game as a research platform for studying grounded dialogue. In it, agents can perceive, emote, and act whilst conducting dialogue with other agents. Models and humans can both act as characters within the game. We describe the results of training state-of-the-art generative and retrieval models in this setting. We show that in addition to using past dialogue, these models are able to effectively use the state of the underlying world to condition their predictions. In particular, we show that grounding on the details of the local environment, including location descriptions, and the objects (and their affordances) and characters (and their previous actions) present within it allows better predictions of agent behavior and dialogue. We analyze the ingredients necessary for successful grounding in this setting, and how each of these factors relate to agents that can talk and act successfully.
http://arxiv.org/abs/1903.03094
Recently, label consistent k-svd(LC-KSVD) algorithm has been successfully applied in image classification. The objective function of LC-KSVD is consisted of reconstruction error, classification error and discriminative sparse codes error with l0-norm sparse regularization term. The l0-norm, however, leads to NP-hard issue. Despite some methods such as orthogonal matching pursuit can help solve this problem to some extent, it is quite difficult to find the optimum sparse solution. To overcome this limitation, we propose a label embedded dictionary learning(LEDL) method to utilise the $\ell_1$-norm as the sparse regularization term so that we can avoid the hard-to-optimize problem by solving the convex optimization problem. Alternating direction method of multipliers and blockwise coordinate descent algorithm are then used to optimize the corresponding objective function. Extensive experimental results on six benchmark datasets illustrate that the proposed algorithm has achieved superior performance compared to some conventional classification algorithms.
http://arxiv.org/abs/1903.03087
Maritime monitoring systems support safe shipping as they allow for the real-time detection of dangerous, suspicious and illegal vessel activities. We present such a system using the Run-Time Event Calculus, a composite event recognition system with formal, declarative semantics. For effective recognition, we developed a library of maritime patterns in close collaboration with domain experts. We present a thorough evaluation of the system and the patterns both in terms of predictive accuracy and computational efficiency, using real-world datasets of vessel position streams and static geographical information.
http://arxiv.org/abs/1903.03078
Characterizing the tissue morphology and anatomy of seagrasses is essential to predicting their acoustic behavior. In this pilot study, we use histology techniques and whole slide imaging (WSI) to describe the composition and topology of the aerenchyma of an entire leaf blade in an automatic way combining the advantages of X-ray microtomography and optical microscopy. Paraffin blocks are prepared in such a way that microtome slices contain an arbitrarily large number of cross sections distributed along the full length of a blade. The sample organization in the paraffin block coupled with whole slide image analysis allows high throughput data extraction and an exhaustive characterization along the whole blade length. The core of the work are image processing algorithms that can identify cells and air lacunae (or void) from fiber strand, epidermis, mesophyll and vascular system. A set of specific features is developed to adequately describe the convexity of cells and voids where standard descriptors fail. The features scrutinize the local curvature of the object borders to allow an accurate discrimination between void and cell through machine learning. The algorithm allows to reconstruct the cells and cell membrane features that are relevant to tissue density, compressibility and rigidity. Size distribution of the different cell types and gas spaces, total biomass and total void volume fraction are then extracted from the high resolution slices to provide a complete characterization of the tissue along the leave from its base to the apex.
http://arxiv.org/abs/1903.03044
Multiple-choice reading comprehension (MCRC) is the task of selecting the correct answer from multiple options given a question and an article. Existing MCRC models typically either read each option independently or compute a fixed-length representation for each option before comparing them. However, humans typically compare the options at multiple-granularity level before reading the article in detail to make reasoning more efficient. Mimicking humans, we propose an option comparison network (OCN) for MCRC which compares options at word-level to better identify their correlations to help reasoning. Specially, each option is encoded into a vector sequence using a skimmer to retain fine-grained information as much as possible. An attention mechanism is leveraged to compare these sequences vector-by-vector to identify more subtle correlations between options, which is potentially valuable for reasoning. Experimental results on the human English exam MCRC dataset RACE show that our model outperforms existing methods significantly. Moreover, it is also the first model that surpasses Amazon Mechanical Turker performance on the whole dataset.
http://arxiv.org/abs/1903.03033
Adversarial images are samples that are intentionally modified to deceive machine learning systems. They are widely used in applications such as CAPTHAs to help distinguish legitimate human users from bots. However, the noise introduced during the adversarial image generation process degrades the perceptual quality and introduces artificial colours; making it also difficult for humans to classify images and recognise objects. In this letter, we propose a method to enhance the perceptual quality of these adversarial images. The proposed method is attack type agnostic and could be used in association with the existing attacks in the literature. Our experiments show that the generated adversarial images have lower Euclidean distance values while maintaining the same adversarial attack performance. Distances are reduced by 5.88% to 41.27% with an average reduction of 22% over the different attack and network types.
http://arxiv.org/abs/1903.03029
Deep learning based speech enhancement and source separation systems have recently reached unprecedented levels of quality, to the point that performance is reaching a new ceiling. Most systems rely on estimating the magnitude of a target source by estimating a real-valued mask to be applied to a time-frequency representation of the mixture signal. A limiting factor in such approaches is a lack of phase estimation: the phase of the mixture is most often used when reconstructing the estimated time-domain signal. Here, we propose “magbook”, “phasebook”, and “combook”, three new types of layers based on discrete representations that can be used to estimate complex time-frequency masks. Magbook layers extend classical sigmoidal units and a recently introduced convex softmax activation for mask-based magnitude estimation. Phasebook layers use a similar structure to give an estimate of the phase mask without suffering from phase wrapping issues. Combook layers are an alternative to the magbook-phasebook combination that directly estimate complex masks. We present various training and inference schemes involving these representations, and explain in particular how to include them in an end-to-end learning framework. We also present an oracle study to assess upper bounds on performance for various types of masks using discrete phase representations. We evaluate the proposed methods on the wsj0-2mix dataset, a well-studied corpus for single-channel speaker-independent speaker separation, matching the performance of state-of-the-art mask-based approaches without requiring additional phase reconstruction steps.
http://arxiv.org/abs/1810.01395
In materials science and particularly electron microscopy, Electron Back-scatter Diffraction (EBSD) is a common and powerful mapping technique for collecting local crystallographic data at the sub-micron scale. The quality of the reconstruction of the maps is critical to study the spatial distribution of phases and crystallographic orientation relationships between phases, a key interest in materials science. However, EBSD data is known to suffer from distortions that arise from several instrument and detector artifacts. In this paper, we present an unsupervised method that corrects those distortions, and enables or enhances phase differentiation in EBSD data. The method uses a segmented electron image of the phases of interest (laths, precipitates, voids, inclusions) gathered using detectors that generate less distorted data, of the same area than the EBSD map, and then searches for the best transformation to correct the distortions of the initial EBSD data. To do so, the Covariance Matrix Adaptation Evolution Strategy (CMA-ES) is implemented to distort the EBSD until it matches the reference electron image. Fast and versatile, this method does not require any human annotation and can be applied to large datasets and wide areas, where the distortions are important. Besides, this method requires very little assumption concerning the shape of the distortion function. Some application examples in multiphase materials with feature sizes down to 1 $\mu$m are presented, including a Titanium alloy and a Nickel-base superalloy.
http://arxiv.org/abs/1903.02982
Current advances in machine translation increase the need for translators to switch from traditional translation to post-editing of machine-translated text, a process that saves time and improves quality. Human and artificial intelligence need to be integrated in an efficient way to leverage the advantages of both for the translation task. This paper outlines approaches at this boundary of AI and HCI and discusses open research questions to further advance the field.
http://arxiv.org/abs/1903.02978
Image representations are commonly learned from class labels, which are a simplistic approximation of human image understanding. In this paper we demonstrate that transferable representations of images can be learned without manual annotations by modeling human visual attention. The basis of our analyses is a unique gaze tracking dataset of sonographers performing routine clinical fetal anomaly screenings. Models of sonographer visual attention are learned by training a convolutional neural network (CNN) to predict gaze on ultrasound video frames through visual saliency prediction or gaze-point regression. We evaluate the transferability of the learned representations to the task of ultrasound standard plane detection in two contexts. Firstly, we perform transfer learning by fine-tuning the CNN with a limited number of labeled standard plane images. We find that fine-tuning the saliency predictor is superior to training from random initialization, with an average F1-score improvement of 9.6% overall and 15.3% for the cardiac planes. Secondly, we train a simple softmax regression on the feature activations of each CNN layer in order to evaluate the representations independently of transfer learning hyper-parameters. We find that the attention models derive strong representations, approaching the precision of a fully-supervised baseline model for all but the last layer.
http://arxiv.org/abs/1903.02974
Multimodal language models attempt to incorporate non-linguistic features for the language modeling task. In this work, we extend a standard recurrent neural network (RNN) language model with features derived from videos. We train our models on data that is two orders-of-magnitude bigger than datasets used in prior work. We perform a thorough exploration of model architectures for combining visual and text features. Our experiments on two corpora (YouCookII and 20bn-something-something-v2) show that the best performing architecture consists of middle fusion of visual and text features, yielding over 25% relative improvement in perplexity. We report analysis that provides insights into why our multimodal language model improves upon a standard RNN language model.
http://arxiv.org/abs/1903.02930
Transparent surfaces, such as glass, transmit most of the visible light that falls on them, making accurate pose estimation challenging. We propose a method to locate glass objects to millimetre accuracy using a simple Laser Range Finder (LRF) attached to the robot end-effector. The method, derived from a physical understanding of laser-glass interactions, consists of (i) sampling points on the glass border by looking at the glass surface from an angle of approximately 45 degrees, and (ii) performing Iterative Closest Point registration on the sampled points. We verify experimentally that the proposed method can locate a transparent, non-planar, side car glass to millimetre accuracy.
http://arxiv.org/abs/1903.02908
Federated Learning allows multiple parties to jointly train a deep learning model on their combined data, without any of the participants having to reveal their local data to a centralized server. This form of privacy-preserving collaborative learning however comes at the cost of a significant communication overhead during training. To address this problem, several compression methods have been proposed in the distributed training literature that can reduce the amount of required communication by up to three orders of magnitude. These existing methods however are only of limited utility in the Federated Learning setting, as they either only compress the upstream communication from the clients to the server (leaving the downstream communication uncompressed) or only perform well under idealized conditions such as iid distribution of the client data, which typically can not be found in Federated Learning. In this work, we propose Sparse Ternary Compression (STC), a new compression framework that is specifically designed to meet the requirements of the Federated Learning environment. Our experiments on four different learning tasks demonstrate that STC distinctively outperforms Federated Averaging in common Federated Learning scenarios where clients either a) hold non-iid data, b) use small batch sizes during training, or where c) the number of clients is large and the participation rate in every communication round is low. We furthermore show that even if the clients hold iid data and use medium sized batches for training, STC still behaves pareto-superior to Federated Averaging in the sense that it achieves fixed target accuracies on our benchmarks within both fewer training iterations and a smaller communication budget.
http://arxiv.org/abs/1903.02891
One of the fundamental challenges to realize massive Multiple-Input Multiple-Output (MIMO) communications is the accurate acquisition of channel state information for a plurality of users at the base station. This is usually accomplished in the UpLink (UL) direction profiting from the time division duplexing mode. In practical base station transceivers, there exist inevitably nonlinear hardware components, like signal amplifiers and various analog filters, which complicates the calibration task. To deal with this challenge, we design a deep neural network for channel calibration between the UL and DownLink (DL) directions. During the initial training phase, the deep neural network is trained from both UL and DL channel measurements. We then leverage the trained deep neural network with the instantaneously estimated UL channel to calibrate the DL one, which is not observable during the UL transmission phase. Our numerical results confirm the merits of the proposed approach, and show that it can achieve performance comparable to conventional approaches, like the Agros method and methods based on least squares, that however assume linear hardware behavior models. More importantly, considering generic nonlinear relationships between the UL and DL channels, it is demonstrated that our deep neural network approach exhibits robust performance, even when the number of training sequences is limited.
http://arxiv.org/abs/1903.02875
There are substantial instructional videos on the Internet, which enables us to acquire knowledge for completing various tasks. However, most existing datasets for instructional video analysis have the limitations in diversity and scale,which makes them far from many real-world applications where more diverse activities occur. Moreover, it still remains a great challenge to organize and harness such data. To address these problems, we introduce a large-scale dataset called “COIN” for COmprehensive INstructional video analysis. Organized with a hierarchical structure, the COIN dataset contains 11,827 videos of 180 tasks in 12 domains (e.g., vehicles, gadgets, etc.) related to our daily life. With a new developed toolbox, all the videos are annotated effectively with a series of step descriptions and the corresponding temporal boundaries. Furthermore, we propose a simple yet effective method to capture the dependencies among different steps, which can be easily plugged into conventional proposal-based action detection methods for localizing important steps in instructional videos. In order to provide a benchmark for instructional video analysis, we evaluate plenty of approaches on the COIN dataset under different evaluation criteria. We expect the introduction of the COIN dataset will promote the future in-depth research on instructional video analysis for the community.
http://arxiv.org/abs/1903.02874
We present an approach for fully automatic urinary bladder segmentation in CT images with artificial neural networks in this study. Automatic medical image analysis has become an invaluable tool in the different treatment stages of diseases. Especially medical image segmentation plays a vital role, since segmentation is often the initial step in an image analysis pipeline. Since deep neural networks have made a large impact on the field of image processing in the past years, we use two different deep learning architectures to segment the urinary bladder. Both of these architectures are based on pre-trained classification networks that are adapted to perform semantic segmentation. Since deep neural networks require a large amount of training data, specifically images and corresponding ground truth labels, we furthermore propose a method to generate such a suitable training data set from Positron Emission Tomography/Computed Tomography image data. This is done by applying thresholding to the Positron Emission Tomography data for obtaining a ground truth and by utilizing data augmentation to enlarge the dataset. In this study, we discuss the influence of data augmentation on the segmentation results, and compare and evaluate the proposed architectures in terms of qualitative and quantitative segmentation performance. The results presented in this study allow concluding that deep neural networks can be considered a promising approach to segment the urinary bladder in CT images.
http://arxiv.org/abs/1903.02871
In recent years, many methods have been developed to identify important portions of text documents. Summarization tools can utilize these methods to extract summaries from large volumes of textual information. However, to identify concepts representing central ideas within a text document and to extract the most informative sentences that best convey those concepts still remain two crucial tasks in summarization methods. In this paper, we introduce a graph-based method to address these two challenges in the context of biomedical text summarization. We show that how a summarizer can discover meaningful concepts within a biomedical text document using the Helmholtz principle. The summarizer considers the meaningful concepts as the main topics and constructs a graph based on the topics that the sentences share. The summarizer can produce an informative summary by extracting those sentences having higher values of the degree. We assess the performance of our method for summarization of biomedical articles using the Recall-Oriented Understudy for Gisting Evaluation (ROUGE) toolkit. The results show that the degree can be a useful centrality measure to identify important sentences in this type of graph-based modelling. Our method can improve the performance of biomedical text summarization compared to some state-of-the-art and publicly available summarizers. Combining a concept-based modelling strategy and a graph-based approach to sentence extraction, our summarizer can produce summaries with the highest scores of informativeness among the comparison methods. This research work can be regarded as a start point to the study of small-world networks in summarization of biomedical texts.
http://arxiv.org/abs/1903.02861
The goal of this paper is to simulate the benefits of jointly applying active learning (AL) and semi-supervised training (SST) in a new speech recognition application. Our data selection approach relies on confidence filtering, and its impact on both the acoustic and language models (AM and LM) is studied. While AL is known to be beneficial to AM training, we show that it also carries out substantial improvements to the LM when combined with SST. Sophisticated confidence models, on the other hand, did not prove to yield any data selection gain. Our results indicate that, while SST is crucial at the beginning of the labeling process, its gains degrade rapidly as AL is set in place. The final simulation reports that AL allows a transcription cost reduction of about 70% over random selection. Alternatively, for a fixed transcription budget, the proposed approach improves the word error rate by about 12.5% relative.
http://arxiv.org/abs/1903.02852
Voice Activity Detection (VAD) refers to the problem of distinguishing speech segments from background noise. Numerous approaches have been proposed for this purpose. Some are based on features derived from the power spectral density, others exploit the periodicity of the signal. The goal of this paper is to investigate the joint use of source and filter-based features. Interestingly, a mutual information-based assessment shows superior discrimination power for the source-related features, especially the proposed ones. The features are further the input of an artificial neural network-based classifier trained on a multi-condition database. Two strategies are proposed to merge source and filter information: feature and decision fusion. Our experiments indicate an absolute reduction of 3% of the equal error rate when using decision fusion. The final proposed system is compared to four state-of-the-art methods on 150 minutes of data recorded in real environments. Thanks to the robustness of its source-related features, its multi-condition training and its efficient information fusion, the proposed system yields over the best state-of-the-art VAD a substantial increase of accuracy across all conditions (24% absolute on average).
http://arxiv.org/abs/1903.02844
Sketch recognition allows natural and efficient interaction in pen-based interfaces. A key obstacle to building accurate sketch recognizers has been the difficulty of creating large amounts of annotated training data. Several authors have attempted to address this issue by creating synthetic data, and by building tools that support efficient annotation. Two prominent sets of approaches stand out from the rest of the crowd. They use interim classifiers trained with a small set of labeled data to aid the labeling of the remainder of the data. The first set of approaches uses a classifier trained with a partially labeled dataset to automatically label unlabeled instances. The others, based on active learning, save annotation effort by giving priority to labeling informative data instances. The former is sub-optimal since it doesn’t prioritize the order of labeling to favor informative instances, while the latter makes the strong assumption that unlabeled data comes in an already segmented form (i.e. the ink in the training data is already assembled into groups forming isolated object instances). In this paper, we propose an active learning framework that combines the strengths of these methods, while addressing their weaknesses. In particular, we propose two methods for deciding how batches of unsegmented sketch scenes should be labeled. The first method, scene-wise selection, assesses the informativeness of each drawing (sketch scene) as a whole, and asks the user to annotate all objects in the drawing. The latter, segment-wise selection, attempts more precise targeting to locate informative fragments of drawings for user labeling. We show that both selection schemes outperform random selection. Furthermore, we demonstrate that precise targeting yields superior performance. Overall, our approach allows reaching top accuracy figures with up to 30% savings in annotation cost.
http://arxiv.org/abs/1903.02832
We perform trend detection on two datasets of Arxiv papers, derived from its machine learning (cs.LG) and natural language processing (cs.CL) categories. Our approach is bottom-up: we first rank papers by their normalized citation counts, then group top-ranked papers into different categories based on the tasks that they pursue and the methods they use. We then analyze these resulting topics. We find that the dominating paradigm in cs.CL revolves around natural language generation problems and those in cs.LG revolve around reinforcement learning and adversarial principles. By extrapolation, we predict that these topics will remain lead problems/approaches in their fields in the short- and mid-term.
http://arxiv.org/abs/1903.02831
Given a training dataset composed of images and corresponding category labels, deep convolutional neural networks show a strong ability in mining discriminative parts for image classification. However, deep convolutional neural networks trained with image level labels only tend to focus on the most discriminative parts while missing other object parts, which could provide complementary information. In this paper, we approach this problem from a different perspective. We build complementary parts models in a weakly supervised manner to retrieve information suppressed by dominant object parts detected by convolutional neural networks. Given image level labels only, we first extract rough object instances by performing weakly supervised object detection and instance segmentation using Mask R-CNN and CRF-based segmentation. Then we estimate and search for the best parts model for each object instance under the principle of preserving as much diversity as possible. In the last stage, we build a bi-directional long short-term memory (LSTM) network to fuze and encode the partial information of these complementary parts into a comprehensive feature for image classification. Experimental results indicate that the proposed method not only achieves significant improvement over our baseline models, but also outperforms state-of-the-art algorithms by a large margin (6.7%, 2.8%, 5.2% respectively) on Stanford Dogs 120, Caltech-UCSD Birds 2011-200 and Caltech 256.
http://arxiv.org/abs/1903.02827
Most deep learning-based models for speech enhancement have mainly focused on estimating the magnitude of spectrogram while reusing the phase from noisy speech for reconstruction. This is due to the difficulty of estimating the phase of clean speech. To improve speech enhancement performance, we tackle the phase estimation problem in three ways. First, we propose Deep Complex U-Net, an advanced U-Net structured model incorporating well-defined complex-valued building blocks to deal with complex-valued spectrograms. Second, we propose a polar coordinate-wise complex-valued masking method to reflect the distribution of complex ideal ratio masks. Third, we define a novel loss function, weighted source-to-distortion ratio (wSDR) loss, which is designed to directly correlate with a quantitative evaluation measure. Our model was evaluated on a mixture of the Voice Bank corpus and DEMAND database, which has been widely used by many deep learning models for speech enhancement. Ablation experiments were conducted on the mixed dataset showing that all three proposed approaches are empirically valid. Experimental results show that the proposed method achieves state-of-the-art performance in all metrics, outperforming previous approaches by a large margin.
http://arxiv.org/abs/1903.03107
In this paper, we propose a novel method that exploits music listening log data for general-purpose music feature extraction. Despite the wealth of information available in the log data of user-item interactions, it has been mostly used for collaborative filtering to find similar items or users and was not fully investigated for content-based music applications. We resolve this problem by extending intra-domain knowledge distillation to cross-domain: i.e., by transferring knowledge obtained from the user-item domain to the music content domain. The proposed system first trains the model that estimates log information from the audio contents; then it uses the model to improve other task-specific models. The experiments on various music classification and regression tasks show that the proposed method successfully improves the performances of the task-specific models.
http://arxiv.org/abs/1903.02794
In crowd scenarios, reliable trajectory prediction of pedestrians requires insightful understanding of their social behaviors. These behaviors have been well investigated by plenty of studies, while it is hard to be fully expressed by hand-craft rules. Recent studies based on LSTM networks have shown great ability to learn social behaviors. However, many of these methods rely on previous neighboring hidden states but ignore the important current intention of the neighbors. In order to address this issue, we propose a data-driven state refinement module for LSTM network (SR-LSTM), which activates the utilization of the current intention of neighbors, and jointly and iteratively refines the current states of all participants in the crowd through a message passing mechanism. To effectively extract the social effect of neighbors, we further introduce a social-aware information selection mechanism consisting of an element-wise motion gate and a pedestrian-wise attention to select useful message from neighboring pedestrians. Experimental results on two public datasets, i.e. ETH and UCY, demonstrate the effectiveness of our proposed SR-LSTM and we achieves state-of-the-art results.
http://arxiv.org/abs/1903.02793
Arabic is recognised as the 4th most used language of the Internet. Arabic has three main varieties: (1) classical Arabic (CA), (2) Modern Standard Arabic (MSA), (3) Arabic Dialect (AD). MSA and AD could be written either in Arabic or in Roman script (Arabizi), which corresponds to Arabic written with Latin letters, numerals and punctuation. Due to the complexity of this language and the number of corresponding challenges for NLP, many surveys have been conducted, in order to synthesise the work done on Arabic. However these surveys principally focus on two varieties of Arabic (MSA and AD, written in Arabic letters only), they are slightly old (no such survey since 2015) and therefore do not cover recent resources and tools. To bridge the gap, we propose a survey focusing on 90 recent research papers (74% of which were published after 2015). Our study presents and classifies the work done on the three varieties of Arabic, by concentrating on both Arabic and Arabizi, and associates each work to its publicly available resources whenever available.
http://arxiv.org/abs/1903.02784
Robust segmentation of hair from portrait images remains challenging: hair does not conform to a uniform shape, style or even color; dark hair in particular lacks features. We present a novel computational imaging solution that tackles the problem from both input and processing fronts. We explore using Time-of-Flight (ToF) RGBD sensors on recent mobile devices. We first conduct a comprehensive analysis to show that scattering and inter-reflection cause different noise patterns on hair vs. non-hair regions on ToF images, by changing the light path and/or combining multiple paths. We then develop a deep network based approach that employs both ToF depth map and the RGB gradient maps to produce an initial hair segmentation with labeled hair components. We then refine the result by imposing ToF noise prior under the conditional random field. We collect the first ToF RGBD hair dataset with 20k+ head images captured on 30 human subjects with a variety of hairstyles at different view angles. Comprehensive experiments show that our approach outperforms the RGB based techniques in accuracy and robustness and can handle traditionally challenging cases such as dark hair, similar hair/background, similar hair/foreground, etc.
http://arxiv.org/abs/1903.02775
The detection of curved lanes is still challenging for autonomous driving systems. Although current cutting-edge approaches have performed well in real applications, most of them are based on strict model assumptions. Similar to other visual recognition tasks, lane detection can be formulated as a two-dimensional graph searching problem, which can be solved by finding several optimal paths along with line segments and boundaries. In this paper, we present a directed graph model, in which dynamic programming is used to deal with a specific shortest path problem. This model is particularly suitable to represent objects with long continuous shape structure, e.g., lanes and roads. We apply the designed model and proposed an algorithm for detecting lanes by formulating it as the shortest path problem. To evaluate the performance of our proposed algorithm, we tested five sequences (including 1573 frames) from the KITTI database. The results showed that our method achieves an average successful detection precision of 97.5%.
http://arxiv.org/abs/1903.02765
Humans can routinely follow a trajectory defined by a list of images/landmarks. However, traditional robot navigation methods require accurate mapping of the environment, localization, and planning. Moreover, these methods are sensitive to subtle changes in the environment. In this paper, we propose a Deep Visual MPC-policy learning method that can perform visual navigation while avoiding collisions with unseen objects on the navigation path. Our model PoliNet takes in as input a visual trajectory and the image of the robot’s current view and outputs velocity commands for a planning horizon of $N$ steps that optimally balance between trajectory following and obstacle avoidance. PoliNet is trained using a strong image predictive model and traversability estimation model in a MPC setup, with minimal human supervision. Different from prior work, PoliNet can be applied to new scenes without retraining. We show experimentally that the robot can follow a visual trajectory when varying start position and in the presence of previously unseen obstacles. We validated our algorithm with tests both in a realistic simulation environment and in the real world. We also show that we can generate visual trajectories in simulation and execute the corresponding path in the real environment. Our approach outperforms classical approaches as well as previous learning-based baselines in success rate of goal reaching, sub-goal coverage rate, and computational load.
http://arxiv.org/abs/1903.02749