aid stringlengths 9 15 | mid stringlengths 7 10 | abstract stringlengths 78 2.56k | related_work stringlengths 92 1.77k | ref_abstract dict |
|---|---|---|---|---|
1901.01535 | 2951394060 | In this paper, we consider the problem of reconstructing a dense 3D model using images captured from different views. Recent methods based on convolutional neural networks (CNN) allow learning the entire task from data. However, they do not incorporate the physics of image formation such as perspective geometry and occ... | A major limitation of all aforementioned approaches is that they require full 3D supervision for training, which is quite restrictive. Tulsiani al @cite_8 relax these assumptions by formulating a differentiable view consistency loss that measures the inconsistency between the predicted 3D shape and its observation. Sim... | {
"cite_N": [
"@cite_18",
"@cite_17",
"@cite_3",
"@cite_8"
],
"mid": [
"2551540143",
"2963730200",
"2619556892",
"2609026071"
],
"abstract": [
"Understanding the 3D world is a fundamental problem in computer vision. However, learning a good representation of 3D objects is s... |
1907.04449 | 2956978206 | Although deep neural networks have been widely applied in many application domains, they are found to be vulnerable to adversarial attacks. A recent promising set of attacking techniques have been proposed, which mainly focus on generating adversarial examples under digital-world settings. Such strategies are unfortuna... | A very recent set of works took the first step in studying physical-world attacking of static physical objects @cite_2 @cite_27 , human objects @cite_41 @cite_33 , stop sign @cite_18 @cite_47 , and roadside sign @cite_12 . Although these works prove to be effective under the targeted scenarios and certain assumptions, ... | {
"cite_N": [
"@cite_18",
"@cite_33",
"@cite_41",
"@cite_27",
"@cite_2",
"@cite_47",
"@cite_16",
"@cite_12"
],
"mid": [
"2125085157",
"2804342109",
"2535873859",
"2963118571",
"2736899637",
"2126628495",
"2798302089",
"2906946247"
],
"abstract": ... |
1907.04449 | 2956978206 | Although deep neural networks have been widely applied in many application domains, they are found to be vulnerable to adversarial attacks. A recent promising set of attacking techniques have been proposed, which mainly focus on generating adversarial examples under digital-world settings. Such strategies are unfortuna... | GAN was first introduced in @cite_20 , implemented by a system of two neural networks contesting with each other in a zero-sum game framework. GAN is proved to be able to achieve visually appealing results in both face generation @cite_4 and manipulation @cite_37 . In order to further improve the quality of synthesis i... | {
"cite_N": [
"@cite_37",
"@cite_4",
"@cite_28",
"@cite_6",
"@cite_40",
"@cite_20"
],
"mid": [
"2519536754",
"2618574778",
"2963073614",
"2783555701",
"2962793481",
"2099471712"
],
"abstract": [
"Realistic image manipulation is challenging because it require... |
1907.04360 | 2961140575 | Behaviour cloning is a commonly used strategy for imitation learning and can be extremely effective in constrained domains. However, in cases where the dynamics of an environment may be state dependent and varying, behaviour cloning places a burden on model capacity and the number of demonstrations required. This paper... | Numerous models and approaches @cite_18 have been developed to address the learning problem formulated above. Gaussian mixture models fit using expectation maximisation @cite_31 are widely used for clustering, while their switching state space analog, Gaussian emission hidden Markov models have a long history of applic... | {
"cite_N": [
"@cite_18",
"@cite_31",
"@cite_32",
"@cite_22"
],
"mid": [
"1530730921",
"2049633694",
"2105594594",
"2102716594"
],
"abstract": [
"1. Basic Principles: The Operating Regime Approach 2. Modelling: Fuzzy Set Methods for Local Modelling Identification 3. Modelli... |
1907.04360 | 2961140575 | Behaviour cloning is a commonly used strategy for imitation learning and can be extremely effective in constrained domains. However, in cases where the dynamics of an environment may be state dependent and varying, behaviour cloning places a burden on model capacity and the number of demonstrations required. This paper... | Learning for switching state space models can also be considered from a changepoint detection perspective, and a range of numerical inference techniques have been used to detect changepoints in sequential data @cite_13 . More recently, variational and gradient-based inference strategies for Bayesian learning have prove... | {
"cite_N": [
"@cite_27",
"@cite_29",
"@cite_26",
"@cite_13"
],
"mid": [
"2547875792",
"2108424780",
"",
"2127074553"
],
"abstract": [
"Categorical variables are a natural choice for representing discrete structure in the world. However, stochastic neural networks rarely us... |
1907.04360 | 2961140575 | Behaviour cloning is a commonly used strategy for imitation learning and can be extremely effective in constrained domains. However, in cases where the dynamics of an environment may be state dependent and varying, behaviour cloning places a burden on model capacity and the number of demonstrations required. This paper... | Hierarchical modelling is an effective means of incorporating structure into a learning problem, so as to avoid sample inefficient learning and improve generalisation through abstraction. Work on options learning @cite_16 @cite_11 and skill identification @cite_6 @cite_0 has paid significant attention to hierarchical l... | {
"cite_N": [
"@cite_0",
"@cite_16",
"@cite_6",
"@cite_11"
],
"mid": [
"2211996086",
"2109910161",
"2217025414",
""
],
"abstract": [
"We present a method for segmenting a set of unstructured demonstration trajectories to discover reusable skills using inverse reinforcement ... |
1907.04360 | 2961140575 | Behaviour cloning is a commonly used strategy for imitation learning and can be extremely effective in constrained domains. However, in cases where the dynamics of an environment may be state dependent and varying, behaviour cloning places a burden on model capacity and the number of demonstrations required. This paper... | Our work is inspired by sequential composition theories in robotics @cite_23 , where tasks are solved by moving between sub-controllers lying within the domains of one another. Here, we seek to identify the sub-controllers required for a given task in an end-to-end fashion, from demonstration sequences. Learning from d... | {
"cite_N": [
"@cite_7",
"@cite_8",
"@cite_21",
"@cite_1",
"@cite_19",
"@cite_23",
"@cite_15",
"@cite_25",
"@cite_20"
],
"mid": [
"1540685400",
"2161395589",
"2121103318",
"1986014385",
"2105401337",
"2132714442",
"2148718436",
"2056884876",
... |
1907.04360 | 2961140575 | Behaviour cloning is a commonly used strategy for imitation learning and can be extremely effective in constrained domains. However, in cases where the dynamics of an environment may be state dependent and varying, behaviour cloning places a burden on model capacity and the number of demonstrations required. This paper... | More recently, trajectory optimisation approaches have been extended to incorporate end-to-end learning, demonstrating robust task level visuomotor control @cite_30 through guided policy search. End-to-end learning has allowed for the use of domain transfer to facilitate one-shot learning @cite_14 from human video demo... | {
"cite_N": [
"@cite_30",
"@cite_14",
"@cite_33",
"@cite_5",
"@cite_12"
],
"mid": [
"2964161785",
"2963703448",
"2963713397",
"2964231903",
"2757631751"
],
"abstract": [
"Policy search methods can allow robots to learn control policies for a wide range of tasks, but... |
1907.04360 | 2961140575 | Behaviour cloning is a commonly used strategy for imitation learning and can be extremely effective in constrained domains. However, in cases where the dynamics of an environment may be state dependent and varying, behaviour cloning places a burden on model capacity and the number of demonstrations required. This paper... | In computer vision, spatial transformers @cite_9 and capsule networks @cite_17 embed learnable structured transformations in an attempt to better capture the relational properties of image attributes in convolutional neural networks. Without this structure, convolution neural networks can learn jumbled image representa... | {
"cite_N": [
"@cite_9",
"@cite_4",
"@cite_17"
],
"mid": [
"603908379",
"",
"2963703618"
],
"abstract": [
"Convolutional Neural Networks define an exceptionally powerful class of models, but are still limited by the lack of ability to be spatially invariant to the input data in a c... |
1907.04360 | 2961140575 | Behaviour cloning is a commonly used strategy for imitation learning and can be extremely effective in constrained domains. However, in cases where the dynamics of an environment may be state dependent and varying, behaviour cloning places a burden on model capacity and the number of demonstrations required. This paper... | Switching density networks are closely related to mixture density networks @cite_10 , a family of neural network constructed using @math output distributions. In the Gaussian mixture case, MDNs fit a weighted combination of Gaussian distributions, using mean @math , variance @math and normalised weight parameters @math... | {
"cite_N": [
"@cite_10"
],
"mid": [
"1579853615"
],
"abstract": [
"Minimization of a sum-of-squares or cross-entropy error function leads to network outputs which approximate the conditional averages of the target data, conditioned on the input vector. For classifications problems, with a suitabl... |
1907.04404 | 2957889783 | In order to facilitate further research in stereo reconstruction with multi-date satellite images, the goal of this paper is to provide a set of stereo-rectified images and the associated groundtruthed disparities for 10 AOIs (Area of Interest) drawn from two sources: 8 AOIs from IARPA's MVS Challenge dataset and 2 AOI... | 3D reconstruction is a popular area of research in the computer vision community and there exist a number of groundtruthed datasets for benchmarking stereo matching algorithms. Although synthetic datasets created using rendered scenes such as the MPI Sintel stereo dataset @cite_15 might prove useful for certain tasks, ... | {
"cite_N": [
"@cite_15",
"@cite_8"
],
"mid": [
"1513100184",
"2108688738"
],
"abstract": [
"Ground truth optical flow is difficult to measure in real scenes with natural motion. As a result, optical flow data sets are restricted in terms of size, complexity, and diversity, making optical ... |
1907.04404 | 2957889783 | In order to facilitate further research in stereo reconstruction with multi-date satellite images, the goal of this paper is to provide a set of stereo-rectified images and the associated groundtruthed disparities for 10 AOIs (Area of Interest) drawn from two sources: 8 AOIs from IARPA's MVS Challenge dataset and 2 AOI... | The Middlebury datasets include the Middlebury2001 @cite_14 , Middlebury2003 @cite_22 , Middlebury2005 and Middlebury2006 datasets @cite_17 and more recently the high resolution Middlebury 2014 dataset @cite_2 . The last dataset was created using a stereo rig with cameras and structured light projectors and claims subp... | {
"cite_N": [
"@cite_14",
"@cite_2",
"@cite_22",
"@cite_17"
],
"mid": [
"2104974755",
"63091017",
"2155479981",
"2133255058"
],
"abstract": [
"Stereo matching is one of the most active research areas in computer vision. While a large number of algorithms for stereo correspo... |
1907.04404 | 2957889783 | In order to facilitate further research in stereo reconstruction with multi-date satellite images, the goal of this paper is to provide a set of stereo-rectified images and the associated groundtruthed disparities for 10 AOIs (Area of Interest) drawn from two sources: 8 AOIs from IARPA's MVS Challenge dataset and 2 AOI... | With a focus on autonomous driving, the KITTI2012 @cite_12 and KITTI2015 @cite_6 datasets were created to capture outdoor scenes. While the former pays attention to static environments, the latter is concerned with moving objects captured by a stereo camera. For generating groundtruth, scans were captured using a laser... | {
"cite_N": [
"@cite_6",
"@cite_12"
],
"mid": [
"1921093919",
"2150066425"
],
"abstract": [
"This paper proposes a novel model and dataset for 3D scene flow estimation with an application to autonomous driving. Taking advantage of the fact that outdoor scenes often decompose into a small n... |
1907.04404 | 2957889783 | In order to facilitate further research in stereo reconstruction with multi-date satellite images, the goal of this paper is to provide a set of stereo-rectified images and the associated groundtruthed disparities for 10 AOIs (Area of Interest) drawn from two sources: 8 AOIs from IARPA's MVS Challenge dataset and 2 AOI... | The datasets described thus far consist of images taken with projective cameras that are either handheld or mounted on stereo rigs. Recently, there was an announcement of a stereo dataset for satellite images @cite_26 that also provides groundtruthed disparities. That dataset however does not provide estimates of the e... | {
"cite_N": [
"@cite_26"
],
"mid": [
"2962880841"
],
"abstract": [
"The increasingly common use of incidental satellite images for stereo reconstruction versus rigidly tasked binocular or trinocular coincident collection is helping to enable timely global-scale 3D mapping; however, reliable stereo... |
1907.04423 | 2961688412 | The spectrum scarcity at sub-6 GHz spectrum has made millimeter-wave (mmWave) frequency band a key component of the next-generation wireless networks. While mmWave spectrum offers extremely large transmission bandwidths to accommodate ever-increasing data rates, unique characteristics of this new spectrum need special ... | The spatial covariance exploits the relatively stationary long-term statistics of the propagation channel, and it can be leveraged for precoder design in mmWave networks @cite_37 @cite_17 @cite_2 @cite_25 @cite_33 . The rationale behind the use of spatial covariance matrix are two-fold. Firstly, in many cases, the angu... | {
"cite_N": [
"@cite_37",
"@cite_18",
"@cite_33",
"@cite_36",
"@cite_27",
"@cite_2",
"@cite_25",
"@cite_17"
],
"mid": [
"",
"2962819920",
"2611749634",
"2962952148",
"2968634732",
"",
"2592085864",
"2241455052"
],
"abstract": [
"",
"Milli... |
1907.04423 | 2961688412 | The spectrum scarcity at sub-6 GHz spectrum has made millimeter-wave (mmWave) frequency band a key component of the next-generation wireless networks. While mmWave spectrum offers extremely large transmission bandwidths to accommodate ever-increasing data rates, unique characteristics of this new spectrum need special ... | Estimating the covariance is complicated due to the fact that only the signals pre-combined by the analog precombiner are available at the baseband. Based on the way the covariance matrix is estimated, it can be broadly categorized into two methods: 1) which we will refer as the indirect method, and 2) which we will re... | {
"cite_N": [
"@cite_35",
"@cite_4",
"@cite_9",
"@cite_6",
"@cite_27"
],
"mid": [
"2848225903",
"2597836615",
"2921341714",
"2963304087",
"2968634732"
],
"abstract": [
"The millimeter-wave (mmWave) communications is a promising technology for next-generation wireles... |
1907.04423 | 2961688412 | The spectrum scarcity at sub-6 GHz spectrum has made millimeter-wave (mmWave) frequency band a key component of the next-generation wireless networks. While mmWave spectrum offers extremely large transmission bandwidths to accommodate ever-increasing data rates, unique characteristics of this new spectrum need special ... | In the literature, several CS approaches have been utilized to estimate the channel and the spatial covariance. For the indirect approach, the channel estimates can be obtained using the SMV CS techniques such as @cite_9 @cite_35 . However, these SMV techniques fail to exploit the common support of the channel estimate... | {
"cite_N": [
"@cite_35",
"@cite_33",
"@cite_8",
"@cite_9",
"@cite_21",
"@cite_6",
"@cite_43",
"@cite_27",
"@cite_12"
],
"mid": [
"2848225903",
"2611749634",
"2109958193",
"2921341714",
"2919210464",
"2963304087",
"1575224478",
"2968634732",
... |
1907.04423 | 2961688412 | The spectrum scarcity at sub-6 GHz spectrum has made millimeter-wave (mmWave) frequency band a key component of the next-generation wireless networks. While mmWave spectrum offers extremely large transmission bandwidths to accommodate ever-increasing data rates, unique characteristics of this new spectrum need special ... | The CS-based methods discussed in are based on the concept of @cite_32 , which provide a virtual angular representation of MIMO channels employing a discretization procedure. The discretization procedure results in an exact sparse representation of the virtual channel model only when the true AoA and AoD lies on one of... | {
"cite_N": [
"@cite_14",
"@cite_41",
"@cite_32",
"@cite_0",
"@cite_11"
],
"mid": [
"2067878805",
"2565293665",
"2128865660",
"2123457453",
"2957141087"
],
"abstract": [
"Compressive Sensing theory details how a sparsely represented signal in a known basis can be re... |
1907.04423 | 2961688412 | The spectrum scarcity at sub-6 GHz spectrum has made millimeter-wave (mmWave) frequency band a key component of the next-generation wireless networks. While mmWave spectrum offers extremely large transmission bandwidths to accommodate ever-increasing data rates, unique characteristics of this new spectrum need special ... | A natural approach to the problem of off-grid basis mismatch is to increase the number of grid points corresponding to decrease in grid sizes. However, this is an inefficient approach due to the following two main problems: Firstly, it increases the mutual coherence of the dictionary, violating the restricted isometric... | {
"cite_N": [
"@cite_30",
"@cite_14",
"@cite_41",
"@cite_0",
"@cite_5"
],
"mid": [
"2114221796",
"2067878805",
"2565293665",
"2123457453",
"2015418199"
],
"abstract": [
"Pulse-Doppler radar has been successfully applied to surveillance and tracking of both moving an... |
1907.04423 | 2961688412 | The spectrum scarcity at sub-6 GHz spectrum has made millimeter-wave (mmWave) frequency band a key component of the next-generation wireless networks. While mmWave spectrum offers extremely large transmission bandwidths to accommodate ever-increasing data rates, unique characteristics of this new spectrum need special ... | An alternative is to tackle the off-grid effects upfront without increasing the grid size. For example, in the context of channel estimation, @cite_22 provide improved off-grid sparse Bayesian algorithm for the channel estimation framework. A grid-less CS technique is developed via atomic norm minimization in the form ... | {
"cite_N": [
"@cite_35",
"@cite_1",
"@cite_22"
],
"mid": [
"2848225903",
"2739833327",
"2810297269"
],
"abstract": [
"The millimeter-wave (mmWave) communications is a promising technology for next-generation wireless networks with its available broad spectrum. Along with massive n... |
1907.04428 | 2959421486 | The proliferation of ubiquitous computing requires energy-efficient as well as secure operation of modern processors. Side channel attacks are becoming a critical threat to security and privacy of devices embedded in modern computing infrastructures. Unintended information leakage via physical signatures such as power ... | Many of the past efforts have focused on EM based side-channel analysis. Nazari et. al. and Callan et. al. presented EM-based acquisition and analysis flow to detect code change injected by an adversary @cite_7 @cite_4 . ML-models were used to learn features extracted from power side channel emissions to create a model... | {
"cite_N": [
"@cite_10",
"@cite_4",
"@cite_7",
"@cite_11"
],
"mid": [
"2124631616",
"2625110865",
"2117552728",
"2751592946"
],
"abstract": [
"Medical devices based on embedded systems are ubiquitous in clinical settings. Increasingly, they connect to networks and run off-... |
1907.04428 | 2959421486 | The proliferation of ubiquitous computing requires energy-efficient as well as secure operation of modern processors. Side channel attacks are becoming a critical threat to security and privacy of devices embedded in modern computing infrastructures. Unintended information leakage via physical signatures such as power ... | DVFS based power management has been explored extensively across all platforms but only recently researchers have started exploring the interactions of DVFS and security @cite_1 @cite_8 @cite_5 @cite_13 . Yang et. al. has demonstrated use of DVFS as a countermeasure to power side channel attack on encryption engines @c... | {
"cite_N": [
"@cite_13",
"@cite_5",
"@cite_1",
"@cite_8"
],
"mid": [
"2750990141",
"2900861686",
"2096490242",
"2798745612"
],
"abstract": [
"",
"This paper demonstrates the improved power and electromagnetic (EM) side-channel attack (SCA) resistance of 128-bit Advance... |
1907.04428 | 2959421486 | The proliferation of ubiquitous computing requires energy-efficient as well as secure operation of modern processors. Side channel attacks are becoming a critical threat to security and privacy of devices embedded in modern computing infrastructures. Unintended information leakage via physical signatures such as power ... | Both profiling based and ML based techniques have been utilized for application inferencing. For instance, to protect the devices against malwares, authors have demonstrated malware detection based on HPCs by selectively choosing application specific hardware and software events @cite_15 . Similar approach, is shown by... | {
"cite_N": [
"@cite_15",
"@cite_12",
"@cite_2"
],
"mid": [
"2319159802",
"2144219822",
"2163643194"
],
"abstract": [
"Hardware Performance Counter-based (HPC) runtime checking is an effective way to identify malicious behaviors of malware and detect malicious modifications to a le... |
1907.04269 | 2962406362 | We present a scheme for sequential decision making with a risk-sensitive objective and constraints in a dynamic environment. A neural network is trained as an approximator of the mapping from parameter space to space of risk and policy with risk-sensitive constraints. For a given risk-sensitive problem, in which the ob... | The SDM problems considering risk, dynamic environment, and constraint are usually studied separately. Besides the works reviewed in , Shen @cite_13 generalized risk measures to the valuation functions. The author applied a set of valuation functions, derived some model-free risk-sensitive reinforcement learning algori... | {
"cite_N": [
"@cite_1",
"@cite_31",
"@cite_13"
],
"mid": [
"2339343364",
"2285087685",
"2181918151"
],
"abstract": [
"In sequential decision-making problems under uncertainty, an agent makes decisions, one after another, considering the current state of the environment where she e... |
1907.04427 | 2957141087 | In this paper, we tackle channel estimation in millimeter-wave hybrid multiple-input multiple-output systems by considering off-grid effects. In particular, we assume that spatial parameters can take any value in the angular domain, and need not fall on predefined discretized angles. Instead of increasing the number of... | One of the most promising features of next-generation wireless systems is to use high-frequency high-bandwidth signals in millimeter-wave (mmWave) frequency bands. These mmWave bands combined with multiple-input multiple-output (MIMO) technology have great potential in delivering higher data rates, higher spectral effi... | {
"cite_N": [
"@cite_16",
"@cite_1"
],
"mid": [
"2195693430",
"2195833401"
],
"abstract": [
"Communication at millimeter wave (mmWave) frequencies is defining a new era of wireless communication. The mmWave band offers higher bandwidth communication channels versus those presently used in ... |
1907.04427 | 2957141087 | In this paper, we tackle channel estimation in millimeter-wave hybrid multiple-input multiple-output systems by considering off-grid effects. In particular, we assume that spatial parameters can take any value in the angular domain, and need not fall on predefined discretized angles. Instead of increasing the number of... | The HADB architecture complicates the channel estimation process, because only the low dimensional signals pre-combined by the analog combiner are available at baseband, which severely degrades the channel estimation process. The accuracy with which the channel is estimated plays a critical role in physical layer perfo... | {
"cite_N": [
"@cite_4",
"@cite_9",
"@cite_16",
"@cite_10",
"@cite_12"
],
"mid": [
"2128865660",
"2921341714",
"2195693430",
"2510608419",
"2339667469"
],
"abstract": [
"Accurate and tractable channel modeling is critical to realizing the full potential of antenna a... |
1907.04427 | 2957141087 | In this paper, we tackle channel estimation in millimeter-wave hybrid multiple-input multiple-output systems by considering off-grid effects. In particular, we assume that spatial parameters can take any value in the angular domain, and need not fall on predefined discretized angles. Instead of increasing the number of... | The virtual channel model describes the channel with respect to (w.r.t.) fixed basis functions corresponding to spatial angles within a finite discrete dictionary. In other words, the continuous parameter space of spatial angular features is discretized into a finite set of pre-defined spatial angles, which emphasizes ... | {
"cite_N": [
"@cite_11"
],
"mid": [
"2565293665"
],
"abstract": [
"This paper investigates the problem of estimating the frequency components of a mixture of s complex sinusoids from a random subset of n regularly spaced samples. Unlike previous work in compressed sensing, the frequencies are not... |
1907.04427 | 2957141087 | In this paper, we tackle channel estimation in millimeter-wave hybrid multiple-input multiple-output systems by considering off-grid effects. In particular, we assume that spatial parameters can take any value in the angular domain, and need not fall on predefined discretized angles. Instead of increasing the number of... | A natural yet inefficient approach to reduce off-grid effects is to increase the number of discretized points, corresponding to increased grid resolution. This approach not only increases the mutual coherence of the dictionary matrix, leading to loss of the restricted isometric property, but also increases the problem ... | {
"cite_N": [
"@cite_15",
"@cite_3",
"@cite_6",
"@cite_8"
],
"mid": [
"2810297269",
"2739833327",
"2127271355",
"2848225903"
],
"abstract": [
"In this letter, an angle domain off-grid channel estimation algorithm for the uplink millimeter wave (mmWave) massive multiple-inpu... |
1907.04427 | 2957141087 | In this paper, we tackle channel estimation in millimeter-wave hybrid multiple-input multiple-output systems by considering off-grid effects. In particular, we assume that spatial parameters can take any value in the angular domain, and need not fall on predefined discretized angles. Instead of increasing the number of... | Interestingly, standard CS methods based on sparsity fail to leverage Dirichlet structure in the Fourier domain. We exploit this structure to improve the channel estimation process. In particular, we propose low-complexity algorithms based on OMP @cite_6 , owing to its computational tractability. Our numerical results ... | {
"cite_N": [
"@cite_6"
],
"mid": [
"2127271355"
],
"abstract": [
"This paper demonstrates theoretically and empirically that a greedy algorithm called orthogonal matching pursuit (OMP) can reliably recover a signal with m nonzero entries in dimension d given O(m ln d) random linear measurements o... |
1901.01028 | 2906683466 | This paper offers three new, open-source, deep learning-based iris segmentation methods, and the methodology how to use irregular segmentation masks in a conventional Gabor-wavelet-based iris recognition. To train and validate the methods, we used a wide spectrum of iris images acquired by different teams and different... | The dominant approach to iris segmentation is certainly the one based on circular approximations of the inner and outer iris boundaries @cite_7 , with later extensions to more complex shapes approximated by Fourier series @cite_21 . We focus on more recent, deep learning-based solutions in this paper. | {
"cite_N": [
"@cite_21",
"@cite_7"
],
"mid": [
"2167075312",
"2102796633"
],
"abstract": [
"This paper presents the following four advances in iris recognition: 1) more disciplined methods for detecting and faithfully modeling the iris inner and outer boundaries with active contours, lead... |
1901.01028 | 2906683466 | This paper offers three new, open-source, deep learning-based iris segmentation methods, and the methodology how to use irregular segmentation masks in a conventional Gabor-wavelet-based iris recognition. To train and validate the methods, we used a wide spectrum of iris images acquired by different teams and different... | Jalilian and Uhl @cite_18 proposed the first, known to us, deep learning-based iris segmentation method. They used several types of convolutional encoder-decoder network trained on ND-Iris-0405, IITD and CASIA-Iris-Ageing-v5 datasets with manually annotated ground-truth segmentation masks. The authors reported better p... | {
"cite_N": [
"@cite_18",
"@cite_14",
"@cite_19",
"@cite_23",
"@cite_10"
],
"mid": [
"2741839910",
"1869924930",
"2200208292",
"121937170",
"2118282683"
],
"abstract": [
"As a considerable breakthrough in artificial intelligence, deep learning has gained great succe... |
1901.01028 | 2906683466 | This paper offers three new, open-source, deep learning-based iris segmentation methods, and the methodology how to use irregular segmentation masks in a conventional Gabor-wavelet-based iris recognition. To train and validate the methods, we used a wide spectrum of iris images acquired by different teams and different... | Arsalan al @cite_1 adapted the VGG-Face network to segmentation of visible-light iris images acquired for NICE-II benchmark and MICHE dataset. The proposed version of VGG has two output neurons, and thus the segmentation is expressed as a binary classification problem (iris non-iris) defined for local image patches. Th... | {
"cite_N": [
"@cite_28",
"@cite_1"
],
"mid": [
"2802806477",
"2765685479"
],
"abstract": [
"The recent advancements in computer vision have opened new horizons for deploying biometric recognition algorithms in mobile and handheld devices. Similarly, iris recognition is now much needed in ... |
1901.01028 | 2906683466 | This paper offers three new, open-source, deep learning-based iris segmentation methods, and the methodology how to use irregular segmentation masks in a conventional Gabor-wavelet-based iris recognition. To train and validate the methods, we used a wide spectrum of iris images acquired by different teams and different... | Lozej al @cite_13 re-trained the U-Net architecture @cite_9 , with different settings of hyperparametres, on CASIA benchmark and made this model publicly available to the research community. | {
"cite_N": [
"@cite_9",
"@cite_13"
],
"mid": [
"2952232639",
"2891396062"
],
"abstract": [
"There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the stro... |
1901.01028 | 2906683466 | This paper offers three new, open-source, deep learning-based iris segmentation methods, and the methodology how to use irregular segmentation masks in a conventional Gabor-wavelet-based iris recognition. To train and validate the methods, we used a wide spectrum of iris images acquired by different teams and different... | Bazrafkan al @cite_11 proposed a few newly designed convolutional neural networks, working in parallel for end-to-end iris segmentation. Initially trained on NIR images from BATH800 and CASIA-Thousand-v4, these structures were fine-tuned for visible-light images collected for UBIRIS and MobBio benchmarks. This paper ad... | {
"cite_N": [
"@cite_11"
],
"mid": [
"2774581827"
],
"abstract": [
"With the increasing imaging and processing capabilities of today's mobile devices, user authentication using iris biometrics has become feasible. However, as the acquisition conditions become more unconstrained and as image qualit... |
1901.01028 | 2906683466 | This paper offers three new, open-source, deep learning-based iris segmentation methods, and the methodology how to use irregular segmentation masks in a conventional Gabor-wavelet-based iris recognition. To train and validate the methods, we used a wide spectrum of iris images acquired by different teams and different... | Bezerra al @cite_0 proposed to use Generative Adversarial Networks @cite_6 in iris segmentation for the first time, in addition to previously used fully convolutional neural networks. These solutions were evaluated on NIR images from BioSec, CASIA-Iris-Interval-v3, CASIA-Iris-Thousand-v4 and IITD datasets, as well as o... | {
"cite_N": [
"@cite_0",
"@cite_6"
],
"mid": [
"2891787817",
"2099471712"
],
"abstract": [
"The iris can be considered as one of the most important biometric traits due to its high degree of uniqueness. Iris-based biometrics applications depend mainly on the iris segmentation whose suitabi... |
1901.01028 | 2906683466 | This paper offers three new, open-source, deep learning-based iris segmentation methods, and the methodology how to use irregular segmentation masks in a conventional Gabor-wavelet-based iris recognition. To train and validate the methods, we used a wide spectrum of iris images acquired by different teams and different... | the only previous solution that was open-sourced is the one offered by Lozej al @cite_13 , and this was in the form of the network weights, no previous work assessed the resulting deep learning-based segmentation from the matching perspective, relying instead only on comparison to manually-annotated segmentation. | {
"cite_N": [
"@cite_13"
],
"mid": [
"2891396062"
],
"abstract": [
"Iris segmentation is an important research topic that received significant attention from the research community over the years. Traditional iris segmentation techniques have typically been focused on hand-crafted procedures that,... |
1901.01015 | 2907285302 | In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing nee... | @cite_36 proposed one of the first approaches to learn visual relationships using CNN. CNN @cite_36 computes an embedding space such that similar examples have similar embeddings and vice versa. @cite_21 uses loss on Siamese CNN to learn embedding for face verification. One of the recent prominent works using CNNs for ... | {
"cite_N": [
"@cite_30",
"@cite_38",
"@cite_35",
"@cite_36",
"@cite_48",
"@cite_9",
"@cite_21",
"@cite_6",
"@cite_0"
],
"mid": [
"2096733369",
"2789546350",
"2963775347",
"2171590421",
"2689134854",
"2788212895",
"2062677035",
"2794497862",
... |
1901.01015 | 2907285302 | In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing nee... | Another method for obtaining an embedding for an object is utilizing a traditional softmax layer @cite_40 @cite_13 , wherein a fully-connected (embedding) layer is added prior to the softmax-loss layer. Each identity is considered as a separate category and the number of categories is equal to the number of identities ... | {
"cite_N": [
"@cite_22",
"@cite_41",
"@cite_3",
"@cite_40",
"@cite_13"
],
"mid": [
"2800513603",
"2962887033",
"2339172597",
"2342611082",
""
],
"abstract": [
"Metric learning aims to construct an embedding where two extracted features corresponding to the same ide... |
1901.01015 | 2907285302 | In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing nee... | : Fine grained vehicle classification is a closely related problem to vehicle re-identification. Notable works for vehicle classification are @cite_32 @cite_23 @cite_44 @cite_10 @cite_24 @cite_50 . The general task is to predict vehicle , BMW-i3-2016, Toyota-Camry-1996. Vehicle re-identification is a relatively finer g... | {
"cite_N": [
"@cite_32",
"@cite_44",
"@cite_24",
"@cite_23",
"@cite_50",
"@cite_10"
],
"mid": [
"2294126139",
"2605117450",
"2475242006",
"2028563077",
"1958236864",
"196211074"
],
"abstract": [
"",
"Fine-grained car recognition aims to recognize the ca... |
1901.01015 | 2907285302 | In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing nee... | : Some notable approaches prior to deep learning are @cite_5 @cite_39 . Popular deep learning approaches for vehicle re-identification are @cite_42 @cite_18 @cite_33 @cite_38 @cite_15 @cite_9 @cite_28 @cite_45 @cite_29 @cite_13 @cite_19 . @cite_15 proposed fusion of handcrafted features color, texture along with high l... | {
"cite_N": [
"@cite_38",
"@cite_18",
"@cite_33",
"@cite_28",
"@cite_9",
"@cite_42",
"@cite_29",
"@cite_39",
"@cite_19",
"@cite_45",
"@cite_5",
"@cite_15",
"@cite_13"
],
"mid": [
"2789546350",
"2470322391",
"2779954854",
"2749235995",
"278821... |
1901.01015 | 2907285302 | In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing nee... | @cite_9 presents a structured deep learning loss comprising a classification loss term (based on vehicle model) as well as coarse and fine grained ranking terms. @cite_18 proposed a modification of triplet loss by replacing anchor samples with corresponding class center in order to suppress effects of using poor anchor... | {
"cite_N": [
"@cite_9",
"@cite_18",
"@cite_33"
],
"mid": [
"2788212895",
"2470322391",
"2779954854"
],
"abstract": [
"Vehicle re-identification (re-ID) is to identify the same vehicle across different cameras. It’s a significant but challenging topic, which has received little att... |
1901.01015 | 2907285302 | In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing nee... | In a recent work @cite_38 , the authors propose to include group-based sub-clustering in a triplet loss framework. This helps in explicitly dealing with intra-class variations of vehicle identification problem. During training an online grouping method is used to cluster samples within each identity into disparate clus... | {
"cite_N": [
"@cite_38"
],
"mid": [
"2789546350"
],
"abstract": [
"The widespread use of surveillance cameras toward smart and safe cities poses the critical but challenging problem of vehicle reidentification (Re-ID). The state-of-the-art research work performed vehicle Re-ID relying on deep met... |
1901.01015 | 2907285302 | In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing nee... | @cite_29 proposes to use a view-point synthesis approach to predict embedding for unknown views given a true view image. These synthetic embeddings for unknown views are generated using bi-directional LSTM @cite_26 . The complete network is trained using a combination of contrastive, reconstruction and generative adver... | {
"cite_N": [
"@cite_26",
"@cite_29",
"@cite_52",
"@cite_19",
"@cite_46"
],
"mid": [
"",
"2802810579",
"1710476689",
"2799251491",
"2147527908"
],
"abstract": [
"",
"Vehicle re-identification (re-ID) is an area that has received far less attention in the compute... |
1901.01015 | 2907285302 | In this paper we tackle the problem of vehicle re-identification in a camera network utilizing triplet embeddings. Re-identification is the problem of matching appearances of objects across different cameras. With the proliferation of surveillance cameras enabling smart and safer cities, there is an ever-increasing nee... | @cite_45 develops a framework utilizing keypoint annotations on vehicles to learn viewpoint invariant features from a CNN. To further enhance the retrieval of matching vehicles the authors use probabilistic spatio-temporal regularization using random variables representing camera transition probabilities. The authors d... | {
"cite_N": [
"@cite_28",
"@cite_45"
],
"mid": [
"2749235995",
"2776879428"
],
"abstract": [
"Vehicle re-identification is an important problem and has many applications in video surveillance and intelligent transportation. It gains increasing attention because of the recent advances of pe... |
1901.01229 | 2951350728 | A new mechanism for efficiently solving the Markov decision processes (MDPs) is proposed in this paper. We introduce the notion of reachability landscape where we use the Mean First Passage Time (MFPT) as a means to characterize the reachability of every state in the state space. We show that such reachability characte... | Decision-making in uncertain environments is a basic problem in the area of artificial intelligence @cite_15 @cite_8 , and Markov decision processes (MDPs) have become very popular for modeling non-deterministic planning problems with full observability @cite_14 @cite_10 . | {
"cite_N": [
"@cite_15",
"@cite_14",
"@cite_10",
"@cite_8"
],
"mid": [
"",
"2119567691",
"1978942630",
"2737668828"
],
"abstract": [
"",
"From the Publisher: The past decade has seen considerable theoretical and applied research on Markov decision processes, as well as... |
1901.01229 | 2951350728 | A new mechanism for efficiently solving the Markov decision processes (MDPs) is proposed in this paper. We introduce the notion of reachability landscape where we use the Mean First Passage Time (MFPT) as a means to characterize the reachability of every state in the state space. We show that such reachability characte... | An important related heuristic for efficiently solving MDPs is the prioritized sweeping @cite_9 , which has been broadly employed to further speed up the value iteration process. This heuristic evaluates each state and obtains a score based on the state's contribution to the convergence, and then prioritizes sorts all ... | {
"cite_N": [
"@cite_9",
"@cite_3",
"@cite_7"
],
"mid": [
"2048226872",
"2121733891",
"2159420891"
],
"abstract": [
"We present a new algorithm, prioritized sweeping, for efficient prediction and control of stochastic Markov systems. Incremental learning methods such as temporal di... |
1901.01229 | 2951350728 | A new mechanism for efficiently solving the Markov decision processes (MDPs) is proposed in this paper. We introduce the notion of reachability landscape where we use the Mean First Passage Time (MFPT) as a means to characterize the reachability of every state in the state space. We show that such reachability characte... | Important related frameworks for solving MDPs also include compact representations such as linear function representation and approximation @cite_5 @cite_14 used in the policy iteration algorithms. The linear equation based techniques (detailed formulation is provided in the paper) do not exploit regions of uniformity ... | {
"cite_N": [
"@cite_5",
"@cite_14",
"@cite_16"
],
"mid": [
"2028145673",
"2119567691",
"1997477668"
],
"abstract": [
"",
"From the Publisher: The past decade has seen considerable theoretical and applied research on Markov decision processes, as well as the growing use of thes... |
1901.01229 | 2951350728 | A new mechanism for efficiently solving the Markov decision processes (MDPs) is proposed in this paper. We introduce the notion of reachability landscape where we use the Mean First Passage Time (MFPT) as a means to characterize the reachability of every state in the state space. We show that such reachability characte... | Another relevant strategy is called real-time dynamic programming (RTDP) @cite_11 where states are not treated uniformly. Specifically, in each DP iteration, only a subset of most important states might be explored, and the selection of the subset of states are usually built on and related to agent's exploration experi... | {
"cite_N": [
"@cite_13",
"@cite_17",
"@cite_12",
"@cite_11"
],
"mid": [
"108082272",
"2089561656",
"2397240726",
"2009533501"
],
"abstract": [
"RTDP is a recent heuristic-search DP algorithm for solving non-deterministic planning problems with full observability. In relati... |
1901.01172 | 2890696839 | Two-level indexes have been widely used to handle trajectories of moving objects that are constrained to a network. The top-level of these indexes handles the spatial dimension, whereas the bottom level handles the temporal dimension. The latter turns out to be an instance of the interval-intersection problem, but it h... | Like FNR-tree and MON-tree, we focus on these two-level indexes. To solve the spatial problem, that is, the representation of the network in space (two-dimensional plane), aforementioned structures use a 2D R-tree, storing the segments of the network as lines. With the spatial problem solved, time has to be associated ... | {
"cite_N": [
"@cite_25",
"@cite_2"
],
"mid": [
"2149906774",
"1579902393"
],
"abstract": [
"This introduction to computational geometry focuses on algorithms. Motivation is provided from the application areas as all techniques are related to particular applications in robotics, graphics, ... |
1901.01172 | 2890696839 | Two-level indexes have been widely used to handle trajectories of moving objects that are constrained to a network. The top-level of these indexes handles the spatial dimension, whereas the bottom level handles the temporal dimension. The latter turns out to be an instance of the interval-intersection problem, but it h... | Another difference with previous solutions is that our approach uncouples the network from the trajectories. This model known as Network-Matched has been successfully used @cite_10 @cite_18 , but without using compact data structures in its implementation. Our approach has the advantage that mapping trajectories to a n... | {
"cite_N": [
"@cite_18",
"@cite_10"
],
"mid": [
"2116949673",
"2010567219"
],
"abstract": [
"In traffic research, management, and planning a number of path-based analyses are heavily used, e.g., for computing turn-times, evaluating green waves, or studying traffic flow. These analyses req... |
1907.05007 | 2958457519 | With a growing demand for the search by image, many works have studied the task of fashion instance-level image retrieval (FIR). Furthermore, the recent works introduce a concept of fashion attribute manipulation (FAM) which manipulates a specific attribute (e.g color) of a fashion item while maintaining the rest of th... | The recent instance-level image retrieval methods have shown a dramatic increase in performance using advances in metric learning @cite_29 @cite_21 @cite_1 @cite_15 @cite_0 . The metric learning focuses on the way it calculates loss by making a pair in an effective way. The instance ID is required for making a pair, an... | {
"cite_N": [
"@cite_22",
"@cite_36",
"@cite_29",
"@cite_21",
"@cite_1",
"@cite_3",
"@cite_0",
"@cite_15",
"@cite_11"
],
"mid": [
"2892607309",
"2157732827",
"2544587078",
"2096733369",
"2555897561",
"2143183660",
"2924876209",
"2883348239",
... |
1907.05007 | 2958457519 | With a growing demand for the search by image, many works have studied the task of fashion instance-level image retrieval (FIR). Furthermore, the recent works introduce a concept of fashion attribute manipulation (FAM) which manipulates a specific attribute (e.g color) of a fashion item while maintaining the rest of th... | Recently, a concept of interactive search in fashion domain has been introduced @cite_10 @cite_14 @cite_8 . The main idea is that users manipulate attributes of a query, and the search engine finds a fashion item with the desired attributes ( green striped dress). Although attribute manipulation is particularly useful ... | {
"cite_N": [
"@cite_14",
"@cite_8",
"@cite_5",
"@cite_10",
"@cite_12"
],
"mid": [
"2893666197",
"2889720216",
"2735001949",
"2155855695",
"2798951647"
],
"abstract": [
"",
"In this paper, we introduce an attribute-based interactive image search which can levera... |
1907.05045 | 2957442170 | Logic programming languages such as Datalog have become popular as Domain Specific Languages (DSLs) for solving large-scale, real-world problems, in particular, static program analysis and network analysis. The logic specifications which model analysis problems, process millions of tuples of data and contain hundreds o... | Debugging for logic programming languages has a long history, with work having been done on algorithmic debugging strategies since the 1980s @cite_10 @cite_8 . These works present a framework for the algorithmic debugging method for Prolog programs, where a system asks the user questions about the intended model of the... | {
"cite_N": [
"@cite_10",
"@cite_8"
],
"mid": [
"1608059426",
"1514468887"
],
"abstract": [
"A meta-program, regardless of the nature of the programming language, is a program whose data denotes another (object) program. The importance of meta-programming can be gauged from its large numbe... |
1907.05045 | 2957442170 | Logic programming languages such as Datalog have become popular as Domain Specific Languages (DSLs) for solving large-scale, real-world problems, in particular, static program analysis and network analysis. The logic specifications which model analysis problems, process millions of tuples of data and contain hundreds o... | Our method also fits into the established frameworks for provenance in Datalog @cite_23 and debugging for Datalog @cite_32 . The proof trees generated by our method are analogous to the computation graphs presented in @cite_32 , and are equally effective for debugging. They can also be seen as an extension of -provenan... | {
"cite_N": [
"@cite_27",
"@cite_32",
"@cite_23"
],
"mid": [
"1552694902",
"1600055947",
"2167541073"
],
"abstract": [
"With the proliferation of database views and curated databases, the issue of data provenance - where a piece of data came from and the process by which it arrived... |
1907.05045 | 2957442170 | Logic programming languages such as Datalog have become popular as Domain Specific Languages (DSLs) for solving large-scale, real-world problems, in particular, static program analysis and network analysis. The logic specifications which model analysis problems, process millions of tuples of data and contain hundreds o... | Debugging Datalog specifications is not the only use case for provenance, with user-guided approaches @cite_40 @cite_22 @cite_3 @cite_5 for program analysis also relying on tracking the origins of data. In @cite_3 @cite_40 @cite_22 , a user may tag certain static analysis alarms, to increase or decrease their importanc... | {
"cite_N": [
"@cite_5",
"@cite_40",
"@cite_22",
"@cite_3"
],
"mid": [
"2050680750",
"2762682773",
"2079877139",
"2798352717"
],
"abstract": [
"A central task for a program analysis concerns how to efficiently find a program abstraction that keeps only information relevant ... |
1907.05190 | 2952522529 | Not all types of supervision signals are created equal: Different types of feedback have different costs and effects on learning. We show how self-regulation strategies that decide when to ask for which kind of feedback from a teacher (or from oneself) can be cast as a learning-to-learn problem leading to improved cost... | Further connections between our work on learning with multiple feedback types can be drawn to various extensions of reinforcement learning by multiple tasks @cite_27 , multiple loss functions @cite_28 , or multiple policies @cite_0 . | {
"cite_N": [
"@cite_28",
"@cite_27",
"@cite_0"
],
"mid": [
"2963780286",
"2551887912",
"2785940258"
],
"abstract": [
"Teaching is critical to human society: it is with teaching that prospective students are educated and human civilization can be inherited and advanced. A good teac... |
1907.04954 | 2954915885 | This paper presents work on modelling the social psychological aspect of socialization in the case of a computationally creative master-apprentice system. In each master-apprentice pair, the master, a genetic algorithm, is seen as a parent for its apprentice, which is an NMT based sequence-to-sequence model. The effect... | Research on an agent community consisting of self-organizing maps @cite_14 , although outside of the computational creativity paradigm, presents a way of simulating the emergence of language. The agents are capable of meaning negotiation and converging into a common language to communicate about edibility of different ... | {
"cite_N": [
"@cite_14"
],
"mid": [
"2187002499"
],
"abstract": [
"In this article, we present a model of a cognitive system, or an agent, with the following properties: it can perceive its environment, it can move in its environment, it can perform some simple actions, and it can send and receiv... |
1907.04954 | 2954915885 | This paper presents work on modelling the social psychological aspect of socialization in the case of a computationally creative master-apprentice system. In each master-apprentice pair, the master, a genetic algorithm, is seen as a parent for its apprentice, which is an NMT based sequence-to-sequence model. The effect... | The papers discussed in this section, as well as other similar previously conducted work @cite_7 @cite_0 @cite_22 , study mostly the collaboration of agents that have an equal social status, in contrast to our case where the social status is hierarchical. Therefore we find that there's need for conducting the study pre... | {
"cite_N": [
"@cite_0",
"@cite_22",
"@cite_7"
],
"mid": [
"2952184798",
"",
"1841291977"
],
"abstract": [
"One particular challenge in AI is the computational modelling and simulation of creativity. Feedback and learning from experience are key aspects of the creative process. Her... |
1907.04569 | 2959551295 | Road markings provide guidance to traffic participants and enforce safe driving behaviour, understanding their semantic meaning is therefore paramount in (automated) driving. However, producing the vast quantities of road marking labels required for training state-of-the-art deep networks is costly, time-consuming, and... | Road marking segmentation as demonstrated in @cite_31 is closest to the application of this paper. The authors train a network for semantic road marking segmentation and improve their results by predicting the vanishing point simultaneously. In contrast to this paper, they require thousands of hand-labelled images, whi... | {
"cite_N": [
"@cite_31",
"@cite_34",
"@cite_20"
],
"mid": [
"2964332990",
"2909971279",
"2890657615"
],
"abstract": [
"In this paper, we propose a unified end-to-end trainable multi-task network that jointly handles lane and road marking detection and recognition that is guided by... |
1907.04569 | 2959551295 | Road markings provide guidance to traffic participants and enforce safe driving behaviour, understanding their semantic meaning is therefore paramount in (automated) driving. However, producing the vast quantities of road marking labels required for training state-of-the-art deep networks is costly, time-consuming, and... | However, virtual data lacks the richness and complexity of the real world. A possible alternative is to augment real-world data. For the task of semantic segmentation this means either generating new, photo-realistic images from semantic labels @cite_26 @cite_5 @cite_7 or enriching semantic labels with virtually-genera... | {
"cite_N": [
"@cite_26",
"@cite_4",
"@cite_7",
"@cite_8",
"@cite_28",
"@cite_6",
"@cite_0",
"@cite_19",
"@cite_2",
"@cite_5",
"@cite_25"
],
"mid": [
"2899412645",
"",
"",
"2903103701",
"",
"2743627947",
"",
"2963201472",
"",
"",
... |
1907.04569 | 2959551295 | Road markings provide guidance to traffic participants and enforce safe driving behaviour, understanding their semantic meaning is therefore paramount in (automated) driving. However, producing the vast quantities of road marking labels required for training state-of-the-art deep networks is costly, time-consuming, and... | * Recently, several approaches have been introduced for more complex scene manipulation, beyond simple augmentation. Additional sensor modalities are used in @cite_29 to offer the flexibility (e.g. different view points) of a virtual simulator, while generating data with the fidelity and richness of real-world images. ... | {
"cite_N": [
"@cite_35",
"@cite_29",
"@cite_21"
],
"mid": [
"2894520414",
"2950317706",
"2912124074"
],
"abstract": [
"Synthetic data has proved increasingly useful in both training and testing machine learning models such as neural networks. The major problem in synthetic data ge... |
1907.04868 | 2959020461 | We are interested in the task of generating multi-instrumental music scores. The Transformer architecture has recently shown great promise for the task of piano score generation; here we adapt it to the multi-instrumental setting. Transformers are complex, high-dimensional language models which are capable of capturing... | Music generation has been an active area of research for decades. Most early work involved manually encoding musical rules into generative systems or rearranging fragments of human-composed music; see @cite_13 for an extensive overview. Recent research has favored machine learning systems which automatically extract pa... | {
"cite_N": [
"@cite_13"
],
"mid": [
"1556624199"
],
"abstract": [
"Algorithmic composition composing by means of formalizable methods has a century old tradition not only in occidental music history. This is the first book to provide a detailed overview of prominent procedures of algorithmic comp... |
1907.04868 | 2959020461 | We are interested in the task of generating multi-instrumental music scores. The Transformer architecture has recently shown great promise for the task of piano score generation; here we adapt it to the multi-instrumental setting. Transformers are complex, high-dimensional language models which are capable of capturing... | Other research focuses on the multi-instrumental setting and seeks to provide systems which can with human-composed material @cite_31 @cite_12 @cite_11 @cite_0 . Unlike the system we develop here, these approaches all require complex inference procedures to generate music without human input. Recent work @cite_18 @cite... | {
"cite_N": [
"@cite_18",
"@cite_28",
"@cite_9",
"@cite_1",
"@cite_0",
"@cite_2",
"@cite_31",
"@cite_25",
"@cite_12",
"@cite_11"
],
"mid": [
"2964289981",
"2792210438",
"2963681776",
"2894295011",
"2902184207",
"2962942158",
"2161850243",
"",... |
1907.04669 | 2960010413 | When predictive models are used to support complex and important decisions, the ability to explain a model's reasoning can increase trust, expose hidden biases, and reduce vulnerability to adversarial attacks. However, attempts at interpreting models are often ad hoc and application-specific, and the concept of interpr... | Many interpretable machine learning approaches involve optimizing some characteristics of the model as proxies for interpretability. Examples include sparsity for linear models @cite_14 , number of splits for decision trees @cite_29 , number of subspace features for case-based reasoning @cite_26 , or depth for rule lis... | {
"cite_N": [
"@cite_14",
"@cite_26",
"@cite_4",
"@cite_7",
"@cite_29",
"@cite_1",
"@cite_6",
"@cite_24",
"@cite_27",
"@cite_5",
"@cite_17"
],
"mid": [
"1523985187",
"2963673242",
"2282821441",
"2613463286",
"1594031697",
"2962861173",
"",
... |
1907.04669 | 2960010413 | When predictive models are used to support complex and important decisions, the ability to explain a model's reasoning can increase trust, expose hidden biases, and reduce vulnerability to adversarial attacks. However, attempts at interpreting models are often ad hoc and application-specific, and the concept of interpr... | In the specific case of linear models, the typical interpretability proxy of sparsity (small number of nonzero coefficients) has been a topic of extensive study over the past twenty years @cite_14 . Sparse regression models can be trained using heuristics such as LASSO @cite_18 , stagewise regression @cite_10 or least-... | {
"cite_N": [
"@cite_18",
"@cite_14",
"@cite_19",
"@cite_23",
"@cite_15",
"@cite_10",
"@cite_12"
],
"mid": [
"2135046866",
"1523985187",
"2963351303",
"2588168955",
"2164878629",
"1885924565",
"2063978378"
],
"abstract": [
"SUMMARY We propose a new m... |
1907.04669 | 2960010413 | When predictive models are used to support complex and important decisions, the ability to explain a model's reasoning can increase trust, expose hidden biases, and reduce vulnerability to adversarial attacks. However, attempts at interpreting models are often ad hoc and application-specific, and the concept of interpr... | Training low-complexity models often affects predictive accuracy, and the tradeoff between the two can be difficult to quantify @cite_9 . Similarly, the limitations of an ex post explanation relative to the original black box model can be difficult to explain to users @cite_16 . And it is not clear that practitioners a... | {
"cite_N": [
"@cite_22",
"@cite_8",
"@cite_9",
"@cite_3",
"@cite_16"
],
"mid": [
"2439568532",
"2119315254",
"2160455305",
"2594475271",
"2807015674"
],
"abstract": [
"Supervised machine learning models boast remarkable predictive capabilities. But can you trust yo... |
1907.04685 | 2960705509 | Learning robot control policies from physics simulations is of great interest to the robotics community as it may render the learning process faster, cheaper, and safer by alleviating the need for expensive real-world experiments. However, the direct transfer of learned behavior from simulation to reality is a major ch... | Hobbs & Hepenstal @cite_7 proved for linear programs that optimization is optimistically biased, given that there are errors in estimating the objective function coefficients. Furthermore, they demonstrated the optimistic bias of a nonlinear program, and mentioned the effect of errors on the parameters of linear constr... | {
"cite_N": [
"@cite_7",
"@cite_6",
"@cite_24",
"@cite_16",
"@cite_10",
"@cite_11"
],
"mid": [
"1990277704",
"2000953623",
"1995878217",
"2621667117",
"2076635261",
"1832379062"
],
"abstract": [
"Does optimization systematically lead to solutions that appear... |
1907.04685 | 2960705509 | Learning robot control policies from physics simulations is of great interest to the robotics community as it may render the learning process faster, cheaper, and safer by alleviating the need for expensive real-world experiments. However, the direct transfer of learned behavior from simulation to reality is a major ch... | Physics simulations have already been used successfully in robot learning. Traditionally, simulators are operating on a single nominal model, which makes the direct transfer of policies from simulation to reality highly vulnerable to model uncertainties and biases. Thus, model-based control in most cases relies on fine... | {
"cite_N": [
"@cite_5",
"@cite_12",
"@cite_17"
],
"mid": [
"1978161072",
"1481659984",
"2001685400"
],
"abstract": [
"The reality gap, which often makes controllers evolved in simulation inefficient once transferred onto the physical robot, remains a critical issue in evolutionary... |
1907.04685 | 2960705509 | Learning robot control policies from physics simulations is of great interest to the robotics community as it may render the learning process faster, cheaper, and safer by alleviating the need for expensive real-world experiments. However, the direct transfer of learned behavior from simulation to reality is a major ch... | As done in @cite_20 @cite_18 @cite_23 @cite_1 @cite_2 we use the domain parameter distribution as a prior which ensures the physical plausibility of each parameter. Note that specifying this distribution in the current state-of-the-art requires the researcher to make design decisions. Chebotar al @cite_14 presented a p... | {
"cite_N": [
"@cite_18",
"@cite_14",
"@cite_1",
"@cite_23",
"@cite_2",
"@cite_20"
],
"mid": [
"2529477964",
"2897345632",
"2767050701",
"",
"2963614114",
"2205975260"
],
"abstract": [
"Sample complexity and safety are major challenges when learning policies... |
1907.04685 | 2960705509 | Learning robot control policies from physics simulations is of great interest to the robotics community as it may render the learning process faster, cheaper, and safer by alleviating the need for expensive real-world experiments. However, the direct transfer of learned behavior from simulation to reality is a major ch... | There is a large consensus that further increasing the simulator's accuracy alone will not bridge the reality gap. Instead, the idea of domain randomization has recently gained momentum. The common characteristic of such approaches is the perturbation of the parameters which determine the physics simulator and the stat... | {
"cite_N": [
"@cite_12"
],
"mid": [
"1481659984"
],
"abstract": [
"The pitfalls of naive robot simulations have been recognised for areas such as evolutionary robotics. It has been suggested that carefully validated simulations with a proper treatment of noise may overcome these problems. This pa... |
1907.04685 | 2960705509 | Learning robot control policies from physics simulations is of great interest to the robotics community as it may render the learning process faster, cheaper, and safer by alleviating the need for expensive real-world experiments. However, the direct transfer of learned behavior from simulation to reality is a major ch... | Wang al @cite_26 proposed sampling initial states, external disturbances, goals, as well as actuator noise from probability distributions and learned walking policies in simulation. Regarding robot RL , recent domain randomization methods focus on perturbing the parameters defining the system dynamics. Approaches cover... | {
"cite_N": [
"@cite_30",
"@cite_18",
"@cite_26",
"@cite_8",
"@cite_1",
"@cite_20"
],
"mid": [
"2785389871",
"2529477964",
"1966784014",
"2623289472",
"2767050701",
"2205975260"
],
"abstract": [
"Model-free reinforcement learning (RL) methods are succeeding ... |
1907.04685 | 2960705509 | Learning robot control policies from physics simulations is of great interest to the robotics community as it may render the learning process faster, cheaper, and safer by alleviating the need for expensive real-world experiments. However, the direct transfer of learned behavior from simulation to reality is a major ch... | Another approach of learning robust policies in simulation is to apply adversarial disturbances to the training process. Mandleka al @cite_28 proposed physically plausible perturbations by randomly deciding when to add a rescaled gradient of the expected return. Pinto al @cite_19 introduced the idea of a second agent w... | {
"cite_N": [
"@cite_28",
"@cite_19"
],
"mid": [
"2773691349",
"2602963933"
],
"abstract": [
"Policy search methods in reinforcement learning have demonstrated success in scaling up to larger problems beyond toy examples. However, deploying these methods on real robots remains challenging ... |
1907.04889 | 2960851860 | Persistent cycles, especially the minimal ones, are useful geometric features functioning as augmentations for the intervals in the purely topological persistence diagrams (also termed as barcodes). In our earlier work, we showed that computing minimal 1-dimensional persistent cycles (persistent 1-cycles) for finite in... | In terms of computing minimal cycles for homology groups, two problems are of most interest: the localization problem and the minimal basis problem. The localization problem asks for computing a minimal cycle in a homology class and the minimal basis problem asks for computing a set of generating cycles for a homology ... | {
"cite_N": [
"@cite_11",
"@cite_22",
"@cite_8",
"@cite_23",
"@cite_10",
"@cite_17"
],
"mid": [
"2066108505",
"2093766264",
"1976067114",
"2029958010",
"2118382603",
"1991566340"
],
"abstract": [
"We describe the first algorithms to compute minimum cuts in s... |
1907.04889 | 2960851860 | Persistent cycles, especially the minimal ones, are useful geometric features functioning as augmentations for the intervals in the purely topological persistence diagrams (also termed as barcodes). In our earlier work, we showed that computing minimal 1-dimensional persistent cycles (persistent 1-cycles) for finite in... | In this work, we use graph cuts and their duality extensively. The duality of cuts on a planar graph and separating cycles on the dual graph has long been utilized to efficiently compute maximal flows and minimal cuts on planar graphs, a topic for which Chambers et al. @cite_11 provide a comprehensive review. In their ... | {
"cite_N": [
"@cite_17",
"@cite_11"
],
"mid": [
"1991566340",
"2066108505"
],
"abstract": [
"We address the problem of localizing homology classes, namely, finding the cycle representing a given class with the most concise geometric measure. We focus on the volume measure, that is, the 1-... |
1907.04889 | 2960851860 | Persistent cycles, especially the minimal ones, are useful geometric features functioning as augmentations for the intervals in the purely topological persistence diagrams (also termed as barcodes). In our earlier work, we showed that computing minimal 1-dimensional persistent cycles (persistent 1-cycles) for finite in... | As pointed out earlier, our main focus is the optimality of representative cycles in the persistence framework. Some early works @cite_12 @cite_9 address the representative cycle problem for persistence by computing minimal cycles at the birth points of intervals without considering what actually die at the death point... | {
"cite_N": [
"@cite_9",
"@cite_24",
"@cite_2",
"@cite_12",
"@cite_20"
],
"mid": [
"2227949165",
"2963904897",
"2897738333",
"2150304504",
"2618638433"
],
"abstract": [
"In this work, we discuss the problem of finding optimal cycles for homology groups of simplicial... |
1907.04733 | 2958944212 | We initiate the study of coresets for clustering in graph metrics, i.e., the shortest-path metric of edge-weighted graphs. Such clustering problems (on graph metrics) are essential to data analysis and used for example in road networks and data visualization. Specifically, we consider @math -Clustering, where given a m... | Coresets for clustering in Euclidean spaces @math have been well studied. @cite_3 constructed the first strong coreset for both and with an exponential size dependence on @math . @cite_12 improved the dimensionality dependence to be polynomial for both and . @cite_32 designed coresets for with size independent of @math... | {
"cite_N": [
"@cite_38",
"@cite_26",
"@cite_4",
"@cite_8",
"@cite_32",
"@cite_3",
"@cite_6",
"@cite_19",
"@cite_23",
"@cite_13",
"@cite_12"
],
"mid": [
"2220402431",
"",
"2013967419",
"2016056456",
"2229238337",
"2045964207",
"1537449969",
... |
1907.04840 | 2956434358 | We demonstrate the possibility of what we call sparse learning: accelerated training of deep neural networks that maintain sparse weights throughout training while achieving performance levels competitive with dense networks. We accomplish this by developing sparse momentum, an algorithm which uses exponentially smooth... | : @cite_0 show that "winning lottery tickets" exist for deep neural networks -- sparse initializations which reach similar predictive performance as dense networks and train just as fast. However, finding these winning lottery tickets is computationally expensive and involves multiple prune and re-train cycles starting... | {
"cite_N": [
"@cite_0"
],
"mid": [
"2805003733"
],
"abstract": [
"Neural network pruning techniques can reduce the parameter counts of trained networks by over 90 , decreasing storage requirements and improving computational performance of inference without compromising accuracy. However, contemp... |
1907.04707 | 2958092922 | Graph classification is practically important in many domains. To solve this problem, one usually calculates a low-dimensional representation for each node in the graph with supervised or unsupervised approaches. Most existing approaches consider all the edges between nodes while overlooking whether the edge will bring... | When dealing with graph structured data, it is important to design efficient graph models to learn the embedding for each node. These methods can be categorized into supervised learning methods and unsupervised learning methods depending on whether utilize the information of the training labels. In , convolution-based ... | {
"cite_N": [
"@cite_14",
"@cite_22",
"@cite_33",
"@cite_9",
"@cite_15"
],
"mid": [
"637153065",
"2964321699",
"1662382123",
"2916106175",
"2964015378"
],
"abstract": [
"Deep Learning's recent successes have mostly relied on Convolutional Networks, which exploit fun... |
1907.04707 | 2958092922 | Graph classification is practically important in many domains. To solve this problem, one usually calculates a low-dimensional representation for each node in the graph with supervised or unsupervised approaches. Most existing approaches consider all the edges between nodes while overlooking whether the edge will bring... | To further enhance the performance of GCN, two types of methods are proposed: 1) sampling and 2) attention. There are two kinds of sampling-based methods including GraphSAGE @cite_10 and FastGCN @cite_12 , they introduced node-wise sampling and layer-wise sampling, separately. GraphSAGE computed node representation wit... | {
"cite_N": [
"@cite_10",
"@cite_12"
],
"mid": [
"2962767366",
"2786915849"
],
"abstract": [
"Low-dimensional embeddings of nodes in large graphs have proved extremely useful in a variety of prediction tasks, from content recommendation to identifying protein functions. However, most exist... |
1907.04707 | 2958092922 | Graph classification is practically important in many domains. To solve this problem, one usually calculates a low-dimensional representation for each node in the graph with supervised or unsupervised approaches. Most existing approaches consider all the edges between nodes while overlooking whether the edge will bring... | Another class of methods can enhance the GCN network are using the attention mechanism, GAT @cite_35 first applies the idea of self-attention to graph representation learning. GAT gives different weights to the neighbors of the center node by weighting the similarity between each neighbor and the center node. However, ... | {
"cite_N": [
"@cite_19",
"@cite_35",
"@cite_37",
"@cite_29"
],
"mid": [
"2890703109",
"2766453196",
"2792839479",
""
],
"abstract": [
"Graph Convolutional Networks (GCNs) have become a crucial tool on learning representations of graph vertices. The main challenge of adapti... |
1907.04667 | 2961035135 | Click-through rate (CTR) prediction is a critical task in online advertising systems. Models like Deep Neural Networks (DNNs) are simple but stateless. They consider each target ad independently and cannot directly extract useful information contained in users' historical ad impressions and clicks. In contrast, models ... | CTR prediction has attracted lots of attention from both academia and industry @cite_11 @cite_13 @cite_16 . Generalized linear models, such as Logistic Regression (LR) @cite_2 and Follow-The-Regularized-Leader (FTRL) @cite_6 , have shown decent performance in practice. However, a linear model lacks the ability to learn... | {
"cite_N": [
"@cite_18",
"@cite_8",
"@cite_10",
"@cite_9",
"@cite_1",
"@cite_6",
"@cite_0",
"@cite_2",
"@cite_5",
"@cite_16",
"@cite_13",
"@cite_11"
],
"mid": [
"2443960221",
"",
"2604662567",
"2748787082",
"2512971201",
"2074694452",
"2... |
1907.04667 | 2961035135 | Click-through rate (CTR) prediction is a critical task in online advertising systems. Models like Deep Neural Networks (DNNs) are simple but stateless. They consider each target ad independently and cannot directly extract useful information contained in users' historical ad impressions and clicks. In contrast, models ... | In this paper, we propose Memory Augmented Deep Neural Network (MA-DNN) for CTR prediction. The proposed MA-DNN achieves a good compromise between DNN and RNN. We are aware of recent work like @cite_12 that also utilizes memory networks @cite_17 . However, @cite_12 is proposed for recommender systems and our way of des... | {
"cite_N": [
"@cite_12",
"@cite_17"
],
"mid": [
"2783944588",
"2950527759"
],
"abstract": [
"User preferences are usually dynamic in real-world recommender systems, and a user»s historical behavior records may not be equally important when predicting his her future interests. Existing rec... |
1907.04752 | 2957510905 | We present the first algorithm for regular expression matching that can take advantage of sparsity in the input instance. Our main result is a new algorithm that solves regular expression matching in @math time, where @math is the number of positions in the regular expression, @math is the length of the string, and @ma... | A related construction, by Chang and Paige @cite_26 , considered compact representations of the position automaton that supports efficiently implementing NFA to DFA conversion by subset construction. They presented a linear space representation that supports efficiently computing the set of states @math reachable via c... | {
"cite_N": [
"@cite_0",
"@cite_26"
],
"mid": [
"2604580999",
"1966773178"
],
"abstract": [
"Abstract A linear time algorithm is presented for testing determinism of a regular expression. It is shown that an input word of length n can be matched against a deterministic regular expression o... |
1907.04658 | 2958996718 | The game of Go has a long history in East Asian countries, but the field of Computer Go has yet to catch up to humans until the past couple of years. While the rules of Go are simple, the strategy and combinatorics of the game are immensely complex. Even within the past couple of years, new programs that rely on neural... | Computer Go, the creation of Go-playing agents for computers, has existed as early as 1968 @cite_8 . As mentioned previously, before convolutional neural networks became popular, MCTS was the most powerful method to play Go. These techniques generally required a lot of enhancements and optimizations. For example, MCTS ... | {
"cite_N": [
"@cite_18",
"@cite_7",
"@cite_8",
"@cite_9",
"@cite_12"
],
"mid": [
"1603772156",
"1568143599",
"1598695809",
"2155625359",
"2101101673"
],
"abstract": [
"Recently, Monte-Carlo Tree Search (MCTS) has advanced the field of computer Go substantially. In ... |
1907.04658 | 2958996718 | The game of Go has a long history in East Asian countries, but the field of Computer Go has yet to catch up to humans until the past couple of years. While the rules of Go are simple, the strategy and combinatorics of the game are immensely complex. Even within the past couple of years, new programs that rely on neural... | Previous work has been done on using only convolutional neural network to play Go. They offered boosts over traditional MCTS but were not able to achieve the same level of play as AlphaGo. In 2008, @cite_6 created a Convolutional Neural Network to play Go using an ensemble of networks. They were only able to achieve a ... | {
"cite_N": [
"@cite_22",
"@cite_21",
"@cite_6",
"@cite_0",
"@cite_16",
"@cite_20"
],
"mid": [
"",
"1757796397",
"1589775371",
"",
"2963284097",
"1947291763"
],
"abstract": [
"",
"We present the first deep learning model to successfully learn control pol... |
1907.04662 | 2960490173 | What is a good exploration strategy for an agent that interacts with an environment in the absence of external rewards? Ideally, we would like to get a policy driving towards a uniform state-action visitation (highly exploring) in a minimum number of steps (fast mixing), in order to ease efficient learning of any goal-... | Other works propose to intrinsically motivate the agent towards learning to reach all possible states in the environment @cite_19 . To extend this same idea from the tabular setting to the context of a continuous, high-dimensional state space, @cite_5 employ a generative model to seek for a maximum-entropy goal distrib... | {
"cite_N": [
"@cite_19",
"@cite_5",
"@cite_21"
],
"mid": [
"2293729149",
"2922007426",
"2914261249"
],
"abstract": [
"While intrinsically motivated learning agents hold considerable promise to overcome limitations of more supervised learning systems, quantitative evaluation and th... |
1907.04662 | 2960490173 | What is a good exploration strategy for an agent that interacts with an environment in the absence of external rewards? Ideally, we would like to get a policy driving towards a uniform state-action visitation (highly exploring) in a minimum number of steps (fast mixing), in order to ease efficient learning of any goal-... | Another promising intrinsic objective is to make value out of the exploration phase by acquiring a set of reusable skills, typically formulated by means of the option framework @cite_10 , which can be combined hierarchically to achieve challenging goals. @cite_2 , a set of options is learned by maximizing an intrinsic ... | {
"cite_N": [
"@cite_15",
"@cite_21",
"@cite_10",
"@cite_2"
],
"mid": [
"1601389419",
"2914261249",
"2109910161",
"1486707268"
],
"abstract": [
"A central role in the development process of children is played by self-exploratory activities Through a playful interaction with... |
1907.04580 | 2962430776 | This paper studies the stabilization for a kind of linear and impulse control systems in finite-dimensional spaces, where impulse instants appear periodically. We present several characterizations on the stabilization; show how to design feedback laws; and provide locations for impulse instants to ensure the stabilizat... | In @cite_1 , the author built up a Kalman-type controllability decomposition for the system: @math Based on the decomposition, a necessary condition, as well as a sufficient condition, for the stabilization of the above system was given. Both results are related to some kind of reachability. The stabilization of the ab... | {
"cite_N": [
"@cite_1",
"@cite_12"
],
"mid": [
"2139559232",
"2161442655"
],
"abstract": [
"In this paper, we address an output feedback stabilization problem for a class of linear impulsive systems that accommodate arbitrarily-spaced impulse times and possibly singular state transition m... |
1907.04580 | 2962430776 | This paper studies the stabilization for a kind of linear and impulse control systems in finite-dimensional spaces, where impulse instants appear periodically. We present several characterizations on the stabilization; show how to design feedback laws; and provide locations for impulse instants to ensure the stabilizat... | About the controllability for impulse control systems, we mention works: @cite_10 @cite_13 @cite_19 @cite_16 @cite_9 @cite_14 @cite_15 and the references therein. | {
"cite_N": [
"@cite_14",
"@cite_10",
"@cite_9",
"@cite_19",
"@cite_15",
"@cite_16",
"@cite_13"
],
"mid": [
"2074704816",
"2031217568",
"2581030625",
"2029314834",
"1989542040",
"2045019399",
"2152531925"
],
"abstract": [
"This paper studies the cont... |
1907.04580 | 2962430776 | This paper studies the stabilization for a kind of linear and impulse control systems in finite-dimensional spaces, where impulse instants appear periodically. We present several characterizations on the stabilization; show how to design feedback laws; and provide locations for impulse instants to ensure the stabilizat... | In @cite_9 , the authors studied the controllability for the system: @math (Here @math , @math and @math .) They found @math (defined in ) with @math and @math ) so that for each @math and each @math with @math , the above system is controllable, provided that @math holds Kalman controllability rank condition. This res... | {
"cite_N": [
"@cite_9"
],
"mid": [
"2581030625"
],
"abstract": [
"Abstract This paper studies the approximate and null controllability for impulse controlled systems of heat equations coupled by a pair ( A , B ) of constant matrices. We present a necessary and sufficient condition for the approxi... |
1907.04580 | 2962430776 | This paper studies the stabilization for a kind of linear and impulse control systems in finite-dimensional spaces, where impulse instants appear periodically. We present several characterizations on the stabilization; show how to design feedback laws; and provide locations for impulse instants to ensure the stabilizat... | About optimal control for impulse control systems, we mention the works: @cite_24 @cite_4 @cite_11 @cite_8 @cite_7 and the references therein. | {
"cite_N": [
"@cite_4",
"@cite_7",
"@cite_8",
"@cite_24",
"@cite_11"
],
"mid": [
"2073506225",
"1985491254",
"2127615550",
"1524099673",
"2326361197"
],
"abstract": [
"This paper addresses the problem of impulsive control and optimization for linear dynamical syste... |
1907.04580 | 2962430776 | This paper studies the stabilization for a kind of linear and impulse control systems in finite-dimensional spaces, where impulse instants appear periodically. We present several characterizations on the stabilization; show how to design feedback laws; and provide locations for impulse instants to ensure the stabilizat... | About general theory for impulse systems, we refer readers to @cite_6 @cite_22 @cite_23 and the references therein. | {
"cite_N": [
"@cite_23",
"@cite_22",
"@cite_6"
],
"mid": [
"1484739396",
"2043104638",
"2050240106"
],
"abstract": [
"Geared primarily to an audience consisting of mathematically advanced undergraduate or beginning graduate students, this text may additionally be used by engineeri... |
1901.00898 | 2905297777 | This work examines the role of reinforcement learning in reducing the severity of on-road collisions by controlling velocity and steering in situations in which contact is imminent. We construct a model, given camera images as input, that is capable of learning and predicting the dynamics of obstacles, cars and pedestr... | A system capable of early detecting a pedestrian’s intention of crossing the road and performing an evasive maneuver if avoidance by braking is impossible is presented in @cite_13 . However, they rely on the existence of a Road Side Unit placed in dangerous road spots in order to detect pedestrian intention and send th... | {
"cite_N": [
"@cite_13"
],
"mid": [
"2010256323"
],
"abstract": [
"We present an active pedestrian protection system that performs an autonomous lane-keeping evasive maneuver in urban traffic scenarios when collision avoidance by braking is no longer possible. The system focuses on pedestrians st... |
1901.00898 | 2905297777 | This work examines the role of reinforcement learning in reducing the severity of on-road collisions by controlling velocity and steering in situations in which contact is imminent. We construct a model, given camera images as input, that is capable of learning and predicting the dynamics of obstacles, cars and pedestr... | As opposed to most previous research, @cite_7 propose a model-free collision avoidance system using Deep Reinforcement Learning (DRL). They derive a balanced reward function for an autonomous braking system based on DRL, where the action space allows 4 choices: no braking, weak braking, medium and strong. The reward fu... | {
"cite_N": [
"@cite_1",
"@cite_7"
],
"mid": [
"2145339207",
"2586886183"
],
"abstract": [
"An artificial agent is developed that learns to play a diverse range of classic Atari 2600 computer games directly from sensory experience, achieving a performance comparable to that of an expert hu... |
1901.00898 | 2905297777 | This work examines the role of reinforcement learning in reducing the severity of on-road collisions by controlling velocity and steering in situations in which contact is imminent. We construct a model, given camera images as input, that is capable of learning and predicting the dynamics of obstacles, cars and pedestr... | Various metrics are being employed in research as a measure of crash severity, while this makes for grim reading they do provide widely adopted quantitative, data driven models of accident outcomes. They include the Acceleration Severity Index (ASI), Occupant Impact Velocity (OIV) and Delta-V. However, @cite_19 show th... | {
"cite_N": [
"@cite_19",
"@cite_9"
],
"mid": [
"1969207082",
"594023927"
],
"abstract": [
"The occupant impact velocity (OIV) and acceleration severity index (ASI) are competing measures of crash severity used to assess occupant injury risk in full-scale crash tests involving roadside saf... |
1901.00898 | 2905297777 | This work examines the role of reinforcement learning in reducing the severity of on-road collisions by controlling velocity and steering in situations in which contact is imminent. We construct a model, given camera images as input, that is capable of learning and predicting the dynamics of obstacles, cars and pedestr... | In @cite_6 the fatality risk of pedestrians as given by the vehicle's speed on impact using the GIDAS dataset (German In-Depth Accident Study) is studied. The dataset includes data from 2127 pedestrians that were involved in accidents between 1999 and 2007. They present a now widely-adopted approximation of the fatalit... | {
"cite_N": [
"@cite_6"
],
"mid": [
"2062417984"
],
"abstract": [
"Knowledge of the amount of violence tolerated by the human body is essential when developing and implementing pedestrian safety strategies. When estimating the potential benefits of new countermeasures, the pedestrian fatality risk... |
1901.00826 | 2907783554 | The last decade has witnessed an unprecedented growth in the demand for data-driven real-time services. These services are fueled by emerging applications that require rapidly injecting data streams and computing updated analytics results in real-time. In many of such applications, the computing resources are often sha... | The notion of AoI is formally introduced in @cite_14 , where the authors analyze the time average AoI in M M 1, M D 1, and D M 1 systems under the FCFS policy. Since this seminal work, the study on the AoI has attracted a lot of research interests. There is a large body of work that focuses on the analysis of the AoI u... | {
"cite_N": [
"@cite_18",
"@cite_14",
"@cite_22",
"@cite_7",
"@cite_9",
"@cite_17",
"@cite_6",
"@cite_15",
"@cite_13",
"@cite_12",
"@cite_11"
],
"mid": [
"2076618162",
"1993918491",
"2962807737",
"2069689609",
"2809579289",
"1205143274",
"206... |
1901.00826 | 2907783554 | The last decade has witnessed an unprecedented growth in the demand for data-driven real-time services. These services are fueled by emerging applications that require rapidly injecting data streams and computing updated analytics results in real-time. In many of such applications, the computing resources are often sha... | Despite the aforementioned studies on service performance and information freshness, the tradeoff between them has often been neglected in the literature (partially due to the nature of the considered applications), except for the following limited work. In @cite_3 , the tradeoff of performance and freshness has been c... | {
"cite_N": [
"@cite_16",
"@cite_4",
"@cite_3"
],
"mid": [
"2145068111",
"2137536766",
"2031513076"
],
"abstract": [
"Typical Web-database systems receive read-only queries, that generate dynamic Web pages as a response, and write-only updates, that keep information up-to-date. Use... |
1901.00942 | 2907266819 | Decision-making problems can be modeled as combinatorial optimization problems with Constraint Programming formalisms such as Constrained Optimization Problems. However, few Constraint Programming formalisms can deal with both optimization and uncertainty at the same time, and none of them are convenient to model probl... | Although the following papers do not deal with uncertainty, they all focus on solving optimization problems in RTS games, in particular StarCraft. Thus @cite_3 propose to model with the optimal building placement to make a wall at a base entrance in order to make easier its defense. @cite_0 @cite_9 propose a and solver... | {
"cite_N": [
"@cite_0",
"@cite_9",
"@cite_3"
],
"mid": [
"1425820592",
"2406883374",
"1716013926"
],
"abstract": [
"GHOST is a framework to help game developers to model and implement their own optimization problems, or to simply instantiate a problem already encoded in GHOST. Pre... |
1901.00942 | 2907266819 | Decision-making problems can be modeled as combinatorial optimization problems with Constraint Programming formalisms such as Constrained Optimization Problems. However, few Constraint Programming formalisms can deal with both optimization and uncertainty at the same time, and none of them are convenient to model probl... | Beyond Constraint Programming but close enough, @cite_7 use a branch and bound algorithms to optimize build order in the RTS game StarCraft. Like @cite_3 , @cite_6 tackle the problem to optimize a wall-in building placement in StarCraft but through the prism of Answer-Set Programming. | {
"cite_N": [
"@cite_3",
"@cite_6",
"@cite_7"
],
"mid": [
"1716013926",
"1599388449",
"2098487995"
],
"abstract": [
"This paper presents a constraint optimization approach to walling in real-time strategy (RTS) games. Walling is a specific type of spatial reasoning, typically emplo... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.