May 6, 2019

2681 words 13 mins read

Paper Group ANR 377

Improving sentence compression by learning to predict gaze. Noise-Tolerant Life-Long Matrix Completion via Adaptive Sampling. Global Optimality of Local Search for Low Rank Matrix Recovery. Text-mining the NeuroSynth corpus using Deep Boltzmann Machines. Regularized Pel-Recursive Motion Estimation Using Generalized Cross-Validation and Spatial Adap …

Improving sentence compression by learning to predict gaze


Title	Improving sentence compression by learning to predict gaze
Authors	Sigrid Klerke, Yoav Goldberg, Anders Søgaard
Abstract	We show how eye-tracking corpora can be used to improve sentence compression models, presenting a novel multi-task learning algorithm based on multi-layer LSTMs. We obtain performance competitive with or better than state-of-the-art approaches.
Tasks	Eye Tracking, Multi-Task Learning, Sentence Compression
Published	2016-04-12
URL	http://arxiv.org/abs/1604.03357v1
PDF	http://arxiv.org/pdf/1604.03357v1.pdf
PWC	https://paperswithcode.com/paper/improving-sentence-compression-by-learning-to
Repo
Framework

Noise-Tolerant Life-Long Matrix Completion via Adaptive Sampling


Title	Noise-Tolerant Life-Long Matrix Completion via Adaptive Sampling
Authors	Maria-Florina Balcan, Hongyang Zhang
Abstract	We study the problem of recovering an incomplete $m\times n$ matrix of rank $r$ with columns arriving online over time. This is known as the problem of life-long matrix completion, and is widely applied to recommendation system, computer vision, system identification, etc. The challenge is to design provable algorithms tolerant to a large amount of noises, with small sample complexity. In this work, we give algorithms achieving strong guarantee under two realistic noise models. In bounded deterministic noise, an adversary can add any bounded yet unstructured noise to each column. For this problem, we present an algorithm that returns a matrix of a small error, with sample complexity almost as small as the best prior results in the noiseless case. For sparse random noise, where the corrupted columns are sparse and drawn randomly, we give an algorithm that exactly recovers an $\mu_0$-incoherent matrix by probability at least $1-\delta$ with sample complexity as small as $O\left(\mu_0rn\log (r/\delta)\right)$. This result advances the state-of-the-art work and matches the lower bound in a worst case. We also study the scenario where the hidden matrix lies on a mixture of subspaces and show that the sample complexity can be even smaller. Our proposed algorithms perform well experimentally in both synthetic and real-world datasets.
Tasks	Matrix Completion
Published	2016-12-01
URL	http://arxiv.org/abs/1612.00100v1
PDF	http://arxiv.org/pdf/1612.00100v1.pdf
PWC	https://paperswithcode.com/paper/noise-tolerant-life-long-matrix-completion
Repo
Framework

Global Optimality of Local Search for Low Rank Matrix Recovery


Title	Global Optimality of Local Search for Low Rank Matrix Recovery
Authors	Srinadh Bhojanapalli, Behnam Neyshabur, Nathan Srebro
Abstract	We show that there are no spurious local minima in the non-convex factorized parametrization of low-rank matrix recovery from incoherent linear measurements. With noisy measurements we show all local minima are very close to a global optimum. Together with a curvature bound at saddle points, this yields a polynomial time global convergence guarantee for stochastic gradient descent {\em from random initialization}.
Tasks
Published	2016-05-23
URL	http://arxiv.org/abs/1605.07221v2
PDF	http://arxiv.org/pdf/1605.07221v2.pdf
PWC	https://paperswithcode.com/paper/global-optimality-of-local-search-for-low
Repo
Framework

Text-mining the NeuroSynth corpus using Deep Boltzmann Machines


Title	Text-mining the NeuroSynth corpus using Deep Boltzmann Machines
Authors	Ricardo Pio Monti, Romy Lorenz, Robert Leech, Christoforos Anagnostopoulos, Giovanni Montana
Abstract	Large-scale automated meta-analysis of neuroimaging data has recently established itself as an important tool in advancing our understanding of human brain function. This research has been pioneered by NeuroSynth, a database collecting both brain activation coordinates and associated text across a large cohort of neuroimaging research papers. One of the fundamental aspects of such meta-analysis is text-mining. To date, word counts and more sophisticated methods such as Latent Dirichlet Allocation have been proposed. In this work we present an unsupervised study of the NeuroSynth text corpus using Deep Boltzmann Machines (DBMs). The use of DBMs yields several advantages over the aforementioned methods, principal among which is the fact that it yields both word and document embeddings in a high-dimensional vector space. Such embeddings serve to facilitate the use of traditional machine learning techniques on the text corpus. The proposed DBM model is shown to learn embeddings with a clear semantic structure.
Tasks
Published	2016-05-01
URL	http://arxiv.org/abs/1605.00223v1
PDF	http://arxiv.org/pdf/1605.00223v1.pdf
PWC	https://paperswithcode.com/paper/text-mining-the-neurosynth-corpus-using-deep
Repo
Framework

Regularized Pel-Recursive Motion Estimation Using Generalized Cross-Validation and Spatial Adaptation


Title	Regularized Pel-Recursive Motion Estimation Using Generalized Cross-Validation and Spatial Adaptation
Authors	Vania V. Estrela, Luis A. Rivera, Paulo C. Beggio, Ricardo T. Lopes
Abstract	The computation of 2-D optical flow by means of regularized pel-recursive algorithms raises a host of issues, which include the treatment of outliers, motion discontinuities and occlusion among other problems. We propose a new approach which allows us to deal with these issues within a common framework. Our approach is based on the use of a technique called Generalized Cross-Validation to estimate the best regularization scheme for a given pixel. In our model, the regularization parameter is a matrix whose entries can account for diverse sources of error. The estimation of the motion vectors takes into consideration local properties of the image following a spatially adaptive approach where each moving pixel is supposed to have its own regularization matrix. Preliminary experiments indicate that this approach provides robust estimates of the optical flow.
Tasks	Motion Estimation, Optical Flow Estimation
Published	2016-11-04
URL	http://arxiv.org/abs/1611.01298v1
PDF	http://arxiv.org/pdf/1611.01298v1.pdf
PWC	https://paperswithcode.com/paper/regularized-pel-recursive-motion-estimation
Repo
Framework

Joint Large-Scale Motion Estimation and Image Reconstruction


Title	Joint Large-Scale Motion Estimation and Image Reconstruction
Authors	Hendrik Dirks
Abstract	This article describes the implementation of the joint motion estimation and image reconstruction framework presented by Burger, Dirks and Sch"onlieb and extends this framework to large-scale motion between consecutive image frames. The variational framework uses displacements between consecutive frames based on the optical flow approach to improve the image reconstruction quality on the one hand and the motion estimation quality on the other. The energy functional consists of a data-fidelity term with a general operator that connects the input sequence to the solution, it has a total variation term for the image sequence and is connected to the underlying flow using an optical flow term. Additional spatial regularity for the flow is modeled by a total variation regularizer for both components of the flow. The numerical minimization is performed in an alternating manner using primal-dual techniques. The resulting schemes are presented as pseudo-code together with a short numerical evaluation.
Tasks	Image Reconstruction, Motion Estimation, Optical Flow Estimation
Published	2016-10-31
URL	http://arxiv.org/abs/1610.09908v1
PDF	http://arxiv.org/pdf/1610.09908v1.pdf
PWC	https://paperswithcode.com/paper/joint-large-scale-motion-estimation-and-image
Repo
Framework

A Hierarchical Reinforcement Learning Method for Persistent Time-Sensitive Tasks


Title	A Hierarchical Reinforcement Learning Method for Persistent Time-Sensitive Tasks
Authors	Xiao Li, Calin Belta
Abstract	Reinforcement learning has been applied to many interesting problems such as the famous TD-gammon and the inverted helicopter flight. However, little effort has been put into developing methods to learn policies for complex persistent tasks and tasks that are time-sensitive. In this paper, we take a step towards solving this problem by using signal temporal logic (STL) as task specification, and taking advantage of the temporal abstraction feature that the options framework provide. We show via simulation that a relatively easy to implement algorithm that combines STL and options can learn a satisfactory policy with a small number of training cases
Tasks	Hierarchical Reinforcement Learning
Published	2016-06-20
URL	http://arxiv.org/abs/1606.06355v1
PDF	http://arxiv.org/pdf/1606.06355v1.pdf
PWC	https://paperswithcode.com/paper/a-hierarchical-reinforcement-learning-method
Repo
Framework

Exploring the potential of combining time of flight and thermal infrared cameras for person detection


Title	Exploring the potential of combining time of flight and thermal infrared cameras for person detection
Authors	Wim Abbeloos, Toon Goedemé
Abstract	Combining new, low-cost thermal infrared and time-of-flight range sensors provides new opportunities. In this position paper we explore the possibilities of combining these sensors and using their fused data for person detection. The proposed calibration approach for this sensor combination differs from the traditional stereo camera calibration in two fundamental ways. A first distinction is that the spectral sensitivity of the two sensors differs significantly. In fact, there is no sensitivity range overlap at all. A second distinction is that their resolution is typically very low, which requires special attention. We assume a situation in which the sensors’ relative position is known, but their orientation is unknown. In addition, some of the typical measurement errors are discussed, and methods to compensate for them are proposed. We discuss how the fused data could allow increased accuracy and robustness without the need for complex algorithms requiring large amounts of computational power and training data.
Tasks	Calibration, Human Detection
Published	2016-12-07
URL	http://arxiv.org/abs/1612.02223v1
PDF	http://arxiv.org/pdf/1612.02223v1.pdf
PWC	https://paperswithcode.com/paper/exploring-the-potential-of-combining-time-of
Repo
Framework

An Adaptive Learning Mechanism for Selection of Increasingly More Complex Systems


Title	An Adaptive Learning Mechanism for Selection of Increasingly More Complex Systems
Authors	Fouad Khan
Abstract	Recently it has been demonstrated that causal entropic forces can lead to the emergence of complex phenomena associated with human cognitive niche such as tool use and social cooperation. Here I show that even more fundamental traits associated with human cognition such as ‘self-awareness’ can easily be demonstrated to be arising out of merely a selection for ‘better regulators’; i.e. systems which respond comparatively better to threats to their existence which are internal to themselves. A simple model demonstrates how indeed the average self-awareness for a universe of systems continues to rise as less self-aware systems are eliminated. The model also demonstrates however that the maximum attainable self-awareness for any system is limited by the plasticity and energy availability for that typology of systems. I argue that this rise in self-awareness may be the reason why systems tend towards greater complexity.
Tasks
Published	2016-04-19
URL	http://arxiv.org/abs/1604.05393v1
PDF	http://arxiv.org/pdf/1604.05393v1.pdf
PWC	https://paperswithcode.com/paper/an-adaptive-learning-mechanism-for-selection
Repo
Framework

LOTS about Attacking Deep Features


Title	LOTS about Attacking Deep Features
Authors	Andras Rozsa, Manuel Günther, Terrance E. Boult
Abstract	Deep neural networks provide state-of-the-art performance on various tasks and are, therefore, widely used in real world applications. DNNs are becoming frequently utilized in biometrics for extracting deep features, which can be used in recognition systems for enrolling and recognizing new individuals. It was revealed that deep neural networks suffer from a fundamental problem, namely, they can unexpectedly misclassify examples formed by slightly perturbing correctly recognized inputs. Various approaches have been developed for generating these so-called adversarial examples, but they aim at attacking end-to-end networks. For biometrics, it is natural to ask whether systems using deep features are immune to or, at least, more resilient to attacks than end-to-end networks. In this paper, we introduce a general technique called the layerwise origin-target synthesis (LOTS) that can be efficiently used to form adversarial examples that mimic the deep features of the target. We analyze and compare the adversarial robustness of the end-to-end VGG Face network with systems that use Euclidean or cosine distance between gallery templates and extracted deep features. We demonstrate that iterative LOTS is very effective and show that systems utilizing deep features are easier to attack than the end-to-end network.
Tasks
Published	2016-11-18
URL	http://arxiv.org/abs/1611.06179v5
PDF	http://arxiv.org/pdf/1611.06179v5.pdf
PWC	https://paperswithcode.com/paper/lots-about-attacking-deep-features
Repo
Framework

Supervised Anomaly Detection in Uncertain Pseudoperiodic Data Streams


Title	Supervised Anomaly Detection in Uncertain Pseudoperiodic Data Streams
Authors	Jiangang Ma, Le Sun, Hua Wang, Yanchun Zhang, Uwe Aickelin
Abstract	Uncertain data streams have been widely generated in many Web applications. The uncertainty in data streams makes anomaly detection from sensor data streams far more challenging. In this paper, we present a novel framework that supports anomaly detection in uncertain data streams. The proposed framework adopts an efficient uncertainty pre-processing procedure to identify and eliminate uncertainties in data streams. Based on the corrected data streams, we develop effective period pattern recognition and feature extraction techniques to improve the computational efficiency. We use classification methods for anomaly detection in the corrected data stream. We also empirically show that the proposed approach shows a high accuracy of anomaly detection on a number of real datasets.
Tasks	Anomaly Detection
Published	2016-07-20
URL	http://arxiv.org/abs/1607.05909v1
PDF	http://arxiv.org/pdf/1607.05909v1.pdf
PWC	https://paperswithcode.com/paper/supervised-anomaly-detection-in-uncertain
Repo
Framework

Deep learning based fence segmentation and removal from an image using a video sequence


Title	Deep learning based fence segmentation and removal from an image using a video sequence
Authors	Sankaraganesh Jonna, Krishna K. Nakka, Rajiv R. Sahay
Abstract	Conventional approaches to image de-fencing use multiple adjacent frames for segmentation of fences in the reference image and are limited to restoring images of static scenes only. In this paper, we propose a de-fencing algorithm for images of dynamic scenes using an occlusion-aware optical flow method. We divide the problem of image de-fencing into the tasks of automated fence segmentation from a single image, motion estimation under known occlusions and fusion of data from multiple frames of a captured video of the scene. Specifically, we use a pre-trained convolutional neural network to segment fence pixels from a single image. The knowledge of spatial locations of fences is used to subsequently estimate optical flow in the occluded frames of the video for the final data fusion step. We cast the fence removal problem in an optimization framework by modeling the formation of the degraded observations. The inverse problem is solved using fast iterative shrinkage thresholding algorithm (FISTA). Experimental results show the effectiveness of proposed algorithm.
Tasks	Motion Estimation, Optical Flow Estimation
Published	2016-09-25
URL	http://arxiv.org/abs/1609.07727v2
PDF	http://arxiv.org/pdf/1609.07727v2.pdf
PWC	https://paperswithcode.com/paper/deep-learning-based-fence-segmentation-and
Repo
Framework

Efficient Representation of Low-Dimensional Manifolds using Deep Networks


Title	Efficient Representation of Low-Dimensional Manifolds using Deep Networks
Authors	Ronen Basri, David Jacobs
Abstract	We consider the ability of deep neural networks to represent data that lies near a low-dimensional manifold in a high-dimensional space. We show that deep networks can efficiently extract the intrinsic, low-dimensional coordinates of such data. We first show that the first two layers of a deep network can exactly embed points lying on a monotonic chain, a special type of piecewise linear manifold, mapping them to a low-dimensional Euclidean space. Remarkably, the network can do this using an almost optimal number of parameters. We also show that this network projects nearby points onto the manifold and then embeds them with little error. We then extend these results to more general manifolds.
Tasks
Published	2016-02-15
URL	http://arxiv.org/abs/1602.04723v1
PDF	http://arxiv.org/pdf/1602.04723v1.pdf
PWC	https://paperswithcode.com/paper/efficient-representation-of-low-dimensional
Repo
Framework

A Note on Alternating Minimization Algorithm for the Matrix Completion Problem


Title	A Note on Alternating Minimization Algorithm for the Matrix Completion Problem
Authors	David Gamarnik, Sidhant Misra
Abstract	We consider the problem of reconstructing a low rank matrix from a subset of its entries and analyze two variants of the so-called Alternating Minimization algorithm, which has been proposed in the past. We establish that when the underlying matrix has rank $r=1$, has positive bounded entries, and the graph $\mathcal{G}$ underlying the revealed entries has bounded degree and diameter which is at most logarithmic in the size of the matrix, both algorithms succeed in reconstructing the matrix approximately in polynomial time starting from an arbitrary initialization. We further provide simulation results which suggest that the second algorithm which is based on the message passing type updates, performs significantly better.
Tasks	Matrix Completion
Published	2016-02-05
URL	http://arxiv.org/abs/1602.02164v1
PDF	http://arxiv.org/pdf/1602.02164v1.pdf
PWC	https://paperswithcode.com/paper/a-note-on-alternating-minimization-algorithm
Repo
Framework

Learning a Metric Embedding for Face Recognition using the Multibatch Method


Title	Learning a Metric Embedding for Face Recognition using the Multibatch Method
Authors	Oren Tadmor, Yonatan Wexler, Tal Rosenwein, Shai Shalev-Shwartz, Amnon Shashua
Abstract	This work is motivated by the engineering task of achieving a near state-of-the-art face recognition on a minimal computing budget running on an embedded system. Our main technical contribution centers around a novel training method, called Multibatch, for similarity learning, i.e., for the task of generating an invariant “face signature” through training pairs of “same” and “not-same” face images. The Multibatch method first generates signatures for a mini-batch of $k$ face images and then constructs an unbiased estimate of the full gradient by relying on all $k^2-k$ pairs from the mini-batch. We prove that the variance of the Multibatch estimator is bounded by $O(1/k^2)$, under some mild conditions. In contrast, the standard gradient estimator that relies on random $k/2$ pairs has a variance of order $1/k$. The smaller variance of the Multibatch estimator significantly speeds up the convergence rate of stochastic gradient descent. Using the Multibatch method we train a deep convolutional neural network that achieves an accuracy of $98.2%$ on the LFW benchmark, while its prediction runtime takes only $30$msec on a single ARM Cortex A9 core. Furthermore, the entire training process took only 12 hours on a single Titan X GPU.
Tasks	Face Recognition
Published	2016-05-24
URL	http://arxiv.org/abs/1605.07270v1
PDF	http://arxiv.org/pdf/1605.07270v1.pdf
PWC	https://paperswithcode.com/paper/learning-a-metric-embedding-for-face
Repo
Framework