May 6, 2019

3180 words 15 mins read

Paper Group ANR 266

Adapting Sampling Interval of Sensor Networks Using On-Line Reinforcement Learning. Towards a Mathematical Understanding of the Difficulty in Learning with Feedforward Neural Networks. Near-Optimal Active Learning of Halfspaces via Query Synthesis in the Noisy Setting. A Deep Neural Network for Chinese Zero Pronoun Resolution. Articulated Clinician …

Adapting Sampling Interval of Sensor Networks Using On-Line Reinforcement Learning


Title	Adapting Sampling Interval of Sensor Networks Using On-Line Reinforcement Learning
Authors	Gabriel Martins Dias, Maddalena Nurchis, Boris Bellalta
Abstract	Monitoring Wireless Sensor Networks (WSNs) are composed of sensor nodes that report temperature, relative humidity, and other environmental parameters. The time between two successive measurements is a critical parameter to set during the WSN configuration because it can impact the WSN’s lifetime, the wireless medium contention and the quality of the reported data. As trends in monitored parameters can significantly vary between scenarios and within time, identifying a sampling interval suitable for several cases is also challenging. In this work, we propose a dynamic sampling rate adaptation scheme based on reinforcement learning, able to tune sensors’ sampling interval on-the-fly, according to environmental conditions and application requirements. The primary goal is to set the sampling interval to the best value possible so as to avoid oversampling and save energy, while not missing environmental changes that can be relevant for the application. In simulations, our mechanism could reduce up to 73% the total number of transmissions compared to a fixed strategy and, simultaneously, keep the average quality of information provided by the WSN. The inherent flexibility of the reinforcement learning algorithm facilitates its use in several scenarios, so as to exploit the broad scope of the Internet of Things.
Tasks
Published	2016-06-07
URL	http://arxiv.org/abs/1606.02193v2
PDF	http://arxiv.org/pdf/1606.02193v2.pdf
PWC	https://paperswithcode.com/paper/adapting-sampling-interval-of-sensor-networks
Repo
Framework

Towards a Mathematical Understanding of the Difficulty in Learning with Feedforward Neural Networks


Title	Towards a Mathematical Understanding of the Difficulty in Learning with Feedforward Neural Networks
Authors	Hao Shen
Abstract	Training deep neural networks for solving machine learning problems is one great challenge in the field, mainly due to its associated optimisation problem being highly non-convex. Recent developments have suggested that many training algorithms do not suffer from undesired local minima under certain scenario, and consequently led to great efforts in pursuing mathematical explanations for such observations. This work provides an alternative mathematical understanding of the challenge from a smooth optimisation perspective. By assuming exact learning of finite samples, sufficient conditions are identified via a critical point analysis to ensure any local minimum to be globally minimal as well. Furthermore, a state of the art algorithm, known as the Generalised Gauss-Newton (GGN) algorithm, is rigorously revisited as an approximate Newton’s algorithm, which shares the property of being locally quadratically convergent to a global minimum under the condition of exact learning.
Tasks
Published	2016-11-17
URL	http://arxiv.org/abs/1611.05827v3
PDF	http://arxiv.org/pdf/1611.05827v3.pdf
PWC	https://paperswithcode.com/paper/towards-a-mathematical-understanding-of-the
Repo
Framework

Near-Optimal Active Learning of Halfspaces via Query Synthesis in the Noisy Setting


Title	Near-Optimal Active Learning of Halfspaces via Query Synthesis in the Noisy Setting
Authors	Lin Chen, Hamed Hassani, Amin Karbasi
Abstract	In this paper, we consider the problem of actively learning a linear classifier through query synthesis where the learner can construct artificial queries in order to estimate the true decision boundaries. This problem has recently gained a lot of interest in automated science and adversarial reverse engineering for which only heuristic algorithms are known. In such applications, queries can be constructed de novo to elicit information (e.g., automated science) or to evade detection with minimal cost (e.g., adversarial reverse engineering). We develop a general framework, called dimension coupling (DC), that 1) reduces a d-dimensional learning problem to d-1 low dimensional sub-problems, 2) solves each sub-problem efficiently, 3) appropriately aggregates the results and outputs a linear classifier, and 4) provides a theoretical guarantee for all possible schemes of aggregation. The proposed method is proved resilient to noise. We show that the DC framework avoids the curse of dimensionality: its computational complexity scales linearly with the dimension. Moreover, we show that the query complexity of DC is near optimal (within a constant factor of the optimum algorithm). To further support our theoretical analysis, we compare the performance of DC with the existing work. We observe that DC consistently outperforms the prior arts in terms of query complexity while often running orders of magnitude faster.
Tasks	Active Learning
Published	2016-03-11
URL	http://arxiv.org/abs/1603.03515v2
PDF	http://arxiv.org/pdf/1603.03515v2.pdf
PWC	https://paperswithcode.com/paper/near-optimal-active-learning-of-halfspaces
Repo
Framework

A Deep Neural Network for Chinese Zero Pronoun Resolution


Title	A Deep Neural Network for Chinese Zero Pronoun Resolution
Authors	Qingyu Yin, Weinan Zhang, Yu Zhang, Ting Liu
Abstract	Existing approaches for Chinese zero pronoun resolution overlook semantic information. This is because zero pronouns have no descriptive information, which results in difficulty in explicitly capturing their semantic similarities with antecedents. Moreover, when dealing with candidate antecedents, traditional systems simply take advantage of the local information of a single candidate antecedent while failing to consider the underlying information provided by the other candidates from a global perspective. To address these weaknesses, we propose a novel zero pronoun-specific neural network, which is capable of representing zero pronouns by utilizing the contextual information at the semantic level. In addition, when dealing with candidate antecedents, a two-level candidate encoder is employed to explicitly capture both the local and global information of candidate antecedents. We conduct experiments on the Chinese portion of the OntoNotes 5.0 corpus. Experimental results show that our approach substantially outperforms the state-of-the-art method in various experimental settings.
Tasks	Chinese Zero Pronoun Resolution
Published	2016-04-20
URL	http://arxiv.org/abs/1604.05800v3
PDF	http://arxiv.org/pdf/1604.05800v3.pdf
PWC	https://paperswithcode.com/paper/a-deep-neural-network-for-chinese-zero
Repo
Framework

Articulated Clinician Detection Using 3D Pictorial Structures on RGB-D Data


Title	Articulated Clinician Detection Using 3D Pictorial Structures on RGB-D Data
Authors	Abdolrahim Kadkhodamohammadi, Afshin Gangi, Michel de Mathelin, Nicolas Padoy
Abstract	Reliable human pose estimation (HPE) is essential to many clinical applications, such as surgical workflow analysis, radiation safety monitoring and human-robot cooperation. Proposed methods for the operating room (OR) rely either on foreground estimation using a multi-camera system, which is a challenge in real ORs due to color similarities and frequent illumination changes, or on wearable sensors or markers, which are invasive and therefore difficult to introduce in the room. Instead, we propose a novel approach based on Pictorial Structures (PS) and on RGB-D data, which can be easily deployed in real ORs. We extend the PS framework in two ways. First, we build robust and discriminative part detectors using both color and depth images. We also present a novel descriptor for depth images, called histogram of depth differences (HDD). Second, we extend PS to 3D by proposing 3D pairwise constraints and a new method that makes exact inference tractable. Our approach is evaluated for pose estimation and clinician detection on a challenging RGB-D dataset recorded in a busy operating room during live surgeries. We conduct series of experiments to study the different part detectors in conjunction with the various 2D or 3D pairwise constraints. Our comparisons demonstrate that 3D PS with RGB-D part detectors significantly improves the results in a visually challenging operating environment.
Tasks	Pose Estimation
Published	2016-02-10
URL	http://arxiv.org/abs/1602.03468v4
PDF	http://arxiv.org/pdf/1602.03468v4.pdf
PWC	https://paperswithcode.com/paper/articulated-clinician-detection-using-3d
Repo
Framework

Deep Online Convex Optimization with Gated Games


Title	Deep Online Convex Optimization with Gated Games
Authors	David Balduzzi
Abstract	Methods from convex optimization are widely used as building blocks for deep learning algorithms. However, the reasons for their empirical success are unclear, since modern convolutional networks (convnets), incorporating rectifier units and max-pooling, are neither smooth nor convex. Standard guarantees therefore do not apply. This paper provides the first convergence rates for gradient descent on rectifier convnets. The proof utilizes the particular structure of rectifier networks which consists in binary active/inactive gates applied on top of an underlying linear network. The approach generalizes to max-pooling, dropout and maxout. In other words, to precisely the neural networks that perform best empirically. The key step is to introduce gated games, an extension of convex games with similar convergence properties that capture the gating function of rectifiers. The main result is that rectifier convnets converge to a critical point at a rate controlled by the gated-regret of the units in the network. Corollaries of the main result include: (i) a game-theoretic description of the representations learned by a neural network; (ii) a logarithmic-regret algorithm for training neural nets; and (iii) a formal setting for analyzing conditional computation in neural nets that can be applied to recently developed models of attention.
Tasks
Published	2016-04-07
URL	http://arxiv.org/abs/1604.01952v1
PDF	http://arxiv.org/pdf/1604.01952v1.pdf
PWC	https://paperswithcode.com/paper/deep-online-convex-optimization-with-gated
Repo
Framework

Generating Videos with Scene Dynamics


Title	Generating Videos with Scene Dynamics
Authors	Carl Vondrick, Hamed Pirsiavash, Antonio Torralba
Abstract	We capitalize on large amounts of unlabeled video in order to learn a model of scene dynamics for both video recognition tasks (e.g. action classification) and video generation tasks (e.g. future prediction). We propose a generative adversarial network for video with a spatio-temporal convolutional architecture that untangles the scene’s foreground from the background. Experiments suggest this model can generate tiny videos up to a second at full frame rate better than simple baselines, and we show its utility at predicting plausible futures of static images. Moreover, experiments and visualizations show the model internally learns useful features for recognizing actions with minimal supervision, suggesting scene dynamics are a promising signal for representation learning. We believe generative video models can impact many applications in video understanding and simulation.
Tasks	Action Classification, Future prediction, Representation Learning, Video Generation, Video Recognition, Video Understanding
Published	2016-09-08
URL	http://arxiv.org/abs/1609.02612v3
PDF	http://arxiv.org/pdf/1609.02612v3.pdf
PWC	https://paperswithcode.com/paper/generating-videos-with-scene-dynamics
Repo
Framework

Rain structure transfer using an exemplar rain image for synthetic rain image generation


Title	Rain structure transfer using an exemplar rain image for synthetic rain image generation
Authors	Chang-Hwan Son, Xiao-Ping Zhang
Abstract	This letter proposes a simple method of transferring rain structures of a given exemplar rain image into a target image. Given the exemplar rain image and its corresponding masked rain image, rain patches including rain structures are extracted randomly, and then residual rain patches are obtained by subtracting those rain patches from their mean patches. Next, residual rain patches are selected randomly, and then added to the given target image along a raster scanning direction. To decrease boundary artifacts around the added patches on the target image, minimum error boundary cuts are found using dynamic programming, and then blending is conducted between overlapping patches. Our experiment shows that the proposed method can generate realistic rain images that have similar rain structures in the exemplar images. Moreover, it is expected that the proposed method can be used for rain removal. More specifically, natural images and synthetic rain images generated via the proposed method can be used to learn classifiers, for example, deep neural networks, in a supervised manner.
Tasks	Image Generation, Rain Removal
Published	2016-10-03
URL	http://arxiv.org/abs/1610.00427v1
PDF	http://arxiv.org/pdf/1610.00427v1.pdf
PWC	https://paperswithcode.com/paper/rain-structure-transfer-using-an-exemplar
Repo
Framework

Learning to Rank Personalized Search Results in Professional Networks


Title	Learning to Rank Personalized Search Results in Professional Networks
Authors	Viet Ha-Thuc, Shakti Sinha
Abstract	LinkedIn search is deeply personalized - for the same queries, different searchers expect completely different results. This paper presents our approach to achieving this by mining various data sources available in LinkedIn to infer searchers’ intents (such as hiring, job seeking, etc.), as well as extending the concept of homophily to capture the searcher-result similarities on many aspects. Then, learning-to-rank (LTR) is applied to combine these signals with standard search features.
Tasks	Learning-To-Rank
Published	2016-05-16
URL	http://arxiv.org/abs/1605.04624v1
PDF	http://arxiv.org/pdf/1605.04624v1.pdf
PWC	https://paperswithcode.com/paper/learning-to-rank-personalized-search-results
Repo
Framework

Kernel-based Sensor Fusion with Application to Audio-Visual Voice Activity Detection


Title	Kernel-based Sensor Fusion with Application to Audio-Visual Voice Activity Detection
Authors	David Dov, Ronen Talmon, Israel Cohen
Abstract	In this paper, we address the problem of multiple view data fusion in the presence of noise and interferences. Recent studies have approached this problem using kernel methods, by relying particularly on a product of kernels constructed separately for each view. From a graph theory point of view, we analyze this fusion approach in a discrete setting. More specifically, based on a statistical model for the connectivity between data points, we propose an algorithm for the selection of the kernel bandwidth, a parameter, which, as we show, has important implications on the robustness of this fusion approach to interferences. Then, we consider the fusion of audio-visual speech signals measured by a single microphone and by a video camera pointed to the face of the speaker. Specifically, we address the task of voice activity detection, i.e., the detection of speech and non-speech segments, in the presence of structured interferences such as keyboard taps and office noise. We propose an algorithm for voice activity detection based on the audio-visual signal. Simulation results show that the proposed algorithm outperforms competing fusion and voice activity detection approaches. In addition, we demonstrate that a proper selection of the kernel bandwidth indeed leads to improved performance.
Tasks	Action Detection, Activity Detection, Sensor Fusion
Published	2016-04-11
URL	http://arxiv.org/abs/1604.02946v1
PDF	http://arxiv.org/pdf/1604.02946v1.pdf
PWC	https://paperswithcode.com/paper/kernel-based-sensor-fusion-with-application
Repo
Framework

A Hybrid, PDE-ODE Control Strategy for Intercepting an Intelligent, well-informed Target in a Stationary, Cluttered Environment


Title	A Hybrid, PDE-ODE Control Strategy for Intercepting an Intelligent, well-informed Target in a Stationary, Cluttered Environment
Authors	Ahmad A. Masoud
Abstract	In [1,2] a new class of intelligent controllers that can semantically embed an agent in a spatial context constraining its behavior in a goal-oriented manner was suggested. A controller of such a class can guide an agent in a stationary unknown environment to a fixed target zone along an obstacle-free trajectory. Here, an extension is suggested that would enable the interception of an intelligent target that is maneuvering to evade capture amidst stationary clutter (i.e. the target zone is moving). This is achieved by forcing the differential properties of the potential field used to induce the control action to satisfy the wave equation. Background of the problem, theoretical developments, as well as, proofs of the ability of the modified control to intercept the target along an obstacle-free trajectory are supplied. Simulation results are also provided.
Tasks
Published	2016-08-20
URL	http://arxiv.org/abs/1608.05864v1
PDF	http://arxiv.org/pdf/1608.05864v1.pdf
PWC	https://paperswithcode.com/paper/a-hybrid-pde-ode-control-strategy-for
Repo
Framework

Using a Distributional Semantic Vector Space with a Knowledge Base for Reasoning in Uncertain Conditions


Title	Using a Distributional Semantic Vector Space with a Knowledge Base for Reasoning in Uncertain Conditions
Authors	Douglas Summers-Stay, Clare Voss, Taylor Cassidy
Abstract	The inherent inflexibility and incompleteness of commonsense knowledge bases (KB) has limited their usefulness. We describe a system called Displacer for performing KB queries extended with the analogical capabilities of the word2vec distributional semantic vector space (DSVS). This allows the system to answer queries with information which was not contained in the original KB in any form. By performing analogous queries on semantically related terms and mapping their answers back into the context of the original query using displacement vectors, we are able to give approximate answers to many questions which, if posed to the KB alone, would return no results. We also show how the hand-curated knowledge in a KB can be used to increase the accuracy of a DSVS in solving analogy problems. In these ways, a KB and a DSVS can make up for each other’s weaknesses.
Tasks
Published	2016-06-13
URL	http://arxiv.org/abs/1606.04000v1
PDF	http://arxiv.org/pdf/1606.04000v1.pdf
PWC	https://paperswithcode.com/paper/using-a-distributional-semantic-vector-space
Repo
Framework

Structured Prediction of 3D Human Pose with Deep Neural Networks


Title	Structured Prediction of 3D Human Pose with Deep Neural Networks
Authors	Bugra Tekin, Isinsu Katircioglu, Mathieu Salzmann, Vincent Lepetit, Pascal Fua
Abstract	Most recent approaches to monocular 3D pose estimation rely on Deep Learning. They either train a Convolutional Neural Network to directly regress from image to 3D pose, which ignores the dependencies between human joints, or model these dependencies via a max-margin structured learning framework, which involves a high computational cost at inference time. In this paper, we introduce a Deep Learning regression architecture for structured prediction of 3D human pose from monocular images that relies on an overcomplete auto-encoder to learn a high-dimensional latent pose representation and account for joint dependencies. We demonstrate that our approach outperforms state-of-the-art ones both in terms of structure preservation and prediction accuracy.
Tasks	3D Pose Estimation, Pose Estimation, Structured Prediction
Published	2016-05-17
URL	http://arxiv.org/abs/1605.05180v1
PDF	http://arxiv.org/pdf/1605.05180v1.pdf
PWC	https://paperswithcode.com/paper/structured-prediction-of-3d-human-pose-with
Repo
Framework

US-Cut: Interactive Algorithm for rapid Detection and Segmentation of Liver Tumors in Ultrasound Acquisitions


Title	US-Cut: Interactive Algorithm for rapid Detection and Segmentation of Liver Tumors in Ultrasound Acquisitions
Authors	Jan Egger, Philip Voglreiter, Mark Dokter, Michael Hofmann, Xiaojun Chen, Wolfram G. Zoller, Dieter Schmalstieg, Alexander Hann
Abstract	Ultrasound (US) is the most commonly used liver imaging modality worldwide. It plays an important role in follow-up of cancer patients with liver metastases. We present an interactive segmentation approach for liver tumors in US acquisitions. Due to the low image quality and the low contrast between the tumors and the surrounding tissue in US images, the segmentation is very challenging. Thus, the clinical practice still relies on manual measurement and outlining of the tumors in the US images. We target this problem by applying an interactive segmentation algorithm to the US data, allowing the user to get real-time feedback of the segmentation results. The algorithm has been developed and tested hand-in-hand by physicians and computer scientists to make sure a future practical usage in a clinical setting is feasible. To cover typical acquisitions from the clinical routine, the approach has been evaluated with dozens of datasets where the tumors are hyperechoic (brighter), hypoechoic (darker) or isoechoic (similar) in comparison to the surrounding liver tissue. Due to the interactive real-time behavior of the approach, it was possible even in difficult cases to find satisfying segmentations of the tumors within seconds and without parameter settings, and the average tumor deviation was only 1.4mm compared with manual measurements. However, the long term goal is to ease the volumetric acquisition of liver tumors in order to evaluate for treatment response. Additional aim is the registration of intraoperative US images via the interactive segmentations to the patient’s pre-interventional CT acquisitions.
Tasks	Interactive Segmentation
Published	2016-03-02
URL	http://arxiv.org/abs/1603.00546v1
PDF	http://arxiv.org/pdf/1603.00546v1.pdf
PWC	https://paperswithcode.com/paper/us-cut-interactive-algorithm-for-rapid
Repo
Framework

Neutrosophic Overset, Neutrosophic Underset, and Neutrosophic Offset. Similarly for Neutrosophic Over-/Under-/Off- Logic, Probability, and Statistics


Title	Neutrosophic Overset, Neutrosophic Underset, and Neutrosophic Offset. Similarly for Neutrosophic Over-/Under-/Off- Logic, Probability, and Statistics
Authors	Florentin Smarandache
Abstract	Neutrosophic Over-/Under-/Off-Set and -Logic were defined by the author in 1995 and published for the first time in 2007. We extended the neutrosophic set respectively to Neutrosophic Overset {when some neutrosophic component is over 1}, Neutrosophic Underset {when some neutrosophic component is below 0}, and to Neutrosophic Offset {when some neutrosophic components are off the interval [0, 1], i.e. some neutrosophic component over 1 and other neutrosophic component below 0}. This is no surprise with respect to the classical fuzzy set/logic, intuitionistic fuzzy set/logic, or classical/imprecise probability, where the values are not allowed outside the interval [0, 1], since our real-world has numerous examples and applications of over-/under-/off-neutrosophic components. For example, person working overtime deserves a membership degree over 1, while a person producing more damage than benefit to a company deserves a membership below 0. Then, similarly, the Neutrosophic Logic/Measure/Probability/Statistics etc. were extended to respectively Neutrosophic Over-/Under-/Off-Logic, -Measure, -Probability, -Statistics etc. [Smarandache, 2007].
Tasks
Published	2016-06-30
URL	http://arxiv.org/abs/1607.00234v1
PDF	http://arxiv.org/pdf/1607.00234v1.pdf
PWC	https://paperswithcode.com/paper/neutrosophic-overset-neutrosophic-underset
Repo
Framework