The Latest Papers about AI
-
Object Detection and Tracking Algorithms for Vehicle Counting: A Comparative Analysis
2020-07-31 17:49:27Vishal Mandal, Yaw Adu-GyamfiAbstract
tract: The rapid advancement in the field of deep learning and high performance computing has highly augmented the scope of video based vehicle counting system. In this paper, the authors deploy several state of the art object detection and tracking algorithms to detect and track different classes of vehicles in their regions of interest (ROI). The goal of correctly detecting and tracking vehicles' in their ROI is to obtain an accurate vehicle count. Multiple combinations of object detection models coupled with different tracking systems are applied to access the best vehicle counting framework. The models' addresses challenges associated to different weather conditions, occlusion and low-light settings and efficiently extracts vehicle information and trajectories through its computationally rich training and feedback cycles. The automatic vehicle counts resulting from all the model combinations are validated and compared against the manually counted ground truths of over 9 hours' traffic video data obtained from the Louisiana Department of Transportation and Development. Experimental results demonstrate that the combination of CenterNet and Deep SORT, Detectron2 and Deep SORT, and YOLOv4 and Deep SORT produced the best overall counting percentage for all vehicles.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16198
PDF
-
Designing Neural Speaker Embeddings with Meta Learning
2020-07-31 17:47:36Manoj Kumar, Tae Jin-Park, Somer Bishop, Shrikanth NarayananAbstract
tract: Neural speaker embeddings trained using classification objectives have demonstrated state-of-the-art performance in multiple applications. Typically, such embeddings are trained on an out-of-domain corpus on a single task e.g., speaker classification, albeit with a large number of classes (speakers). In this work, we reformulate embedding training under the meta-learning paradigm. We redistribute the training corpus as an ensemble of multiple related speaker classification tasks, and learn a representation that generalizes better to unseen speakers. First, we develop an open source toolkit to train x-vectors that is matched in performance with pre-trained Kaldi models for speaker diarization and speaker verification applications. We find that different bottleneck layers in the architecture variedly favor different applications. Next, we use two meta-learning strategies, namely prototypical networks and relation networks, to improve over the x-vector embeddings. Our best performing model achieves a relative improvement of 12.37% and 7.11% in speaker error on the DIHARD II development corpus and the AMI meeting corpus, respectively. We analyze improvements across different domains in the DIHARD corpus. Notably, on the challenging child speech domain, we study the relation between child age and the diarization performance. Further, we show reductions in equal error rate for speaker verification on the SITW corpus (7.68%) and the VOiCES challenge corpus (8.78%). We observe that meta-learning particularly offers benefits in challenging acoustic conditions and recording setups encountered in these corpora. Our experiments illustrate the applicability of meta-learning as a generalized learning paradigm for training deep neural speaker embeddings.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16196
PDF
-
Palm Vein Identification based on hybrid features selection model
2020-07-31 17:46:57Mohammed Hamzah Abed, Ali H. Alsaeedi, Ali D. Alfoudi, Abayomi M. Otebolaku, Yasmeen Sajid RazooqiAbstract
tract: Palm vein identification (PVI) is a modern biometric security technique used for increasing security and authentication systems. The key characteristics of palm vein patterns include, its uniqueness to each individual, unforgettable, non-intrusive and cannot be taken by an unauthorized person. However, the extracted features from the palm vein pattern are huge with high redundancy. In this paper, we propose a combine model of two-Dimensional Discrete Wavelet Transform, Principal Component Analysis (PCA), and Particle Swarm Optimization (PSO) (2D-DWTPP) to enhance prediction of vein palm patterns. The 2D-DWT Extracts features from palm vein images, PCA reduces the redundancy in palm vein features. The system has been trained in selecting high reverent features based on the wrapper model. The PSO feeds wrapper model by an optimal subset of features. The proposed system uses four classifiers as an objective function to determine VPI which include Support Vector Machine (SVM), K Nearest Neighbor (KNN), Decision Tree (DT) and Naïve Bayes (NB). The empirical result shows the proposed system Iit satisfied best results with SVM. The proposed 2D-DWTPP model has been evaluated and the results shown remarkable efficiency in comparison with Alexnet and classifier without feature selection. Experimentally, our model has better accuracy reflected by (98.65) while Alexnet has (63.5) and applied classifier without feature selection has (78.79).
Abstract (translated)
URL
https://arxiv.org/abs/2007.16195
PDF
-
SimulEval: An Evaluation Toolkit for Simultaneous Translation
2020-07-31 17:44:41Xutai Ma, Mohammad Javad Dousti, Changhan Wang, Jiatao Gu, Juan PinoAbstract
tract: Simultaneous translation on both text and speech focuses on a real-time and low-latency scenario where the model starts translating before reading the complete source input. Evaluating simultaneous translation models is more complex than offline models because the latency is another factor to consider in addition to translation quality. The research community, despite its growing focus on novel modeling approaches to simultaneous translation, currently lacks a universal evaluation procedure. Therefore, we present SimulEval, an easy-to-use and general evaluation toolkit for both simultaneous text and speech translation. A server-client scheme is introduced to create a simultaneous translation scenario, where the server sends source input and receives predictions for evaluation and the client executes customized policies. Given a policy, it automatically performs simultaneous decoding and collectively reports several popular latency metrics. We also adapt latency metrics from text simultaneous translation to the speech task. Additionally, SimulEval is equipped with a visualization interface to provide better understanding of the simultaneous decoding process of a system. SimulEval has already been extensively used for the IWSLT 2020 shared task on simultaneous speech translation. Code will be released upon publication.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16193
PDF
-
Self-supervised learning through the eyes of a child
2020-07-31 17:33:45A. Emin Orhan, Vaibhav V. Gupta, Brenden M. LakeAbstract
tract: Within months of birth, children have meaningful expectations about the world around them. How much of this early knowledge can be explained through generic learning mechanisms applied to sensory data, and how much of it requires more substantive innate inductive biases? Addressing this fundamental question in its full generality is currently infeasible, but we can hope to make real progress in more narrowly defined domains, such as the development of high-level visual categories, thanks to improvements in data collecting technology and recent progress in deep learning. In this paper, our goal is to achieve such progress by utilizing modern self-supervised deep learning methods and a recent longitudinal, egocentric video dataset recorded from the perspective of several young children (Sullivan et al., 2020). Our results demonstrate the emergence of powerful, high-level visual representations from developmentally realistic natural videos using generic self-supervised learning objectives.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16189
PDF
-
Ultra-light deep MIR by trimming lottery tickets
2020-07-31 17:30:28Philippe Esling, Theis Bazin, Adrien Bitton, Tristan Carsault, Ninon DevisAbstract
tract: Current state-of-the-art results in Music Information Retrieval are largely dominated by deep learning approaches. These provide unprecedented accuracy across all tasks. However, the consistently overlooked downside of these models is their stunningly massive complexity, which seems concomitantly crucial to their success. In this paper, we address this issue by proposing a model pruning method based on the lottery ticket hypothesis. We modify the original approach to allow for explicitly removing parameters, through structured trimming of entire units, instead of simply masking individual weights. This leads to models which are effectively lighter in terms of size, memory and number of operations. We show that our proposal can remove up to 90% of the model parameters without loss of accuracy, leading to ultra-light deep MIR models. We confirm the surprising result that, at smaller compression ratios (removing up to 85% of a network), lighter models consistently outperform their heavier counterparts. We exhibit these results on a large array of MIR tasks including audio classification, pitch recognition, chord extraction, drum transcription and onset estimation. The resulting ultra-light deep learning models for MIR can run on CPU, and can even fit on embedded devices with minimal degradation of accuracy.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16187
PDF
-
Diet deep generative audio models with structured lottery
2020-07-31 16:43:10Philippe Esling, Ninon Devis, Adrien Bitton, Antoine Caillon, Axel Chemla--Romeu-Santos, Constance DouwesAbstract
tract: Deep learning models have provided extremely successful solutions in most audio application fields. However, the high accuracy of these models comes at the expense of a tremendous computation cost. This aspect is almost always overlooked in evaluating the quality of proposed models. However, models should not be evaluated without taking into account their complexity. This aspect is especially critical in audio applications, which heavily relies on specialized embedded hardware with real-time constraints. In this paper, we build on recent observations that deep models are highly overparameterized, by studying the lottery ticket hypothesis on deep generative audio models. This hypothesis states that extremely efficient small sub-networks exist in deep models and would provide higher accuracy than larger models if trained in isolation. However, lottery tickets are found by relying on unstructured masking, which means that resulting models do not provide any gain in either disk size or inference time. Instead, we develop here a method aimed at performing structured trimming. We show that this requires to rely on global selection and introduce a specific criterion based on mutual information. First, we confirm the surprising result that smaller models provide higher accuracy than their large counterparts. We further show that we can remove up to 95% of the model weights without significant degradation in accuracy. Hence, we can obtain very light models for generative audio across popular methods such as Wavenet, SING or DDSP, that are up to 100 times smaller with commensurate accuracy. We study the theoretical bounds for embedding these models on Raspberry Pi and Arduino, and show that we can obtain generative models on CPU with equivalent quality as large GPU models. Finally, we discuss the possibility of implementing deep generative audio models on embedded platforms.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16170
PDF
-
Imitative Planning using Conditional Normalizing Flow
2020-07-31 16:32:23Shubhankar Agarwal, Harshit Sikchi, Cole Gulino, Eric WilkinsonAbstract
tract: We explore the application of normalizing flows for improving the performance of trajectory planning for autonomous vehicles (AVs). Normalizing flows provide an invertible mapping from a known prior distribution to a potentially complex, multi-modal target distribution and allow for fast sampling with exact PDF inference. By modeling a trajectory planner's cost manifold as an energy function we learn a scene conditioned mapping from the prior to a Boltzmann distribution over the AV control space. This mapping allows for control samples and their associated energy to be generated jointly and in parallel. We propose using neural autoregressive flow (NAF) as part of an end-to-end deep learned system that allows for utilizing sensors, map, and route information to condition the flow mapping. Finally, we demonstrate the effectiveness of our approach on real world datasets over IL and hand constructed trajectory sampling techniques.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16162
PDF
-
Paying Per-label Attention for Multi-label Extraction from Radiology Reports
2020-07-31 16:11:09Patrick Schrempf, Hannah Watson, Shadia Mikhael, Maciej Pajak, Matúš Falis, Aneta Lisowska, Keith W. Muir, David Harris-Birtill, Alison Q. O'NeilAbstract
tract: Training medical image analysis models requires large amounts of expertly annotated data which is time-consuming and expensive to obtain. Images are often accompanied by free-text radiology reports which are a rich source of information. In this paper, we tackle the automated extraction of structured labels from head CT reports for imaging of suspected stroke patients, using deep learning. Firstly, we propose a set of 31 labels which correspond to radiographic findings (e.g. hyperdensity) and clinical impressions (e.g. haemorrhage) related to neurological abnormalities. Secondly, inspired by previous work, we extend existing state-of-the-art neural network models with a label-dependent attention mechanism. Using this mechanism and simple synthetic data augmentation, we are able to robustly extract many labels with a single model, classified according to the radiologist's reporting (positive, uncertain, negative). This approach can be used in further research to effectively extract many labels from medical text.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16152
PDF
-
HMCNAS: Neural Architecture Search using Hidden Markov Chains and Bayesian Optimization
2020-07-31 16:04:08Vasco Lopes, Luís A. AlexandreAbstract
tract: Neural Architecture Search has achieved state-of-the-art performance in a variety of tasks, out-performing human-designed networks. However, many assumptions, that require human definition, related with the problems being solved or the models generated are still needed: final model architectures, number of layers to be sampled, forced operations, small search spaces, which ultimately contributes to having models with higher performances at the cost of inducing bias into the system. In this paper, we propose HMCNAS, which is composed of two novel components: i) a method that leverages information about human-designed models to autonomously generate a complex search space, and ii) an Evolutionary Algorithm with Bayesian Optimization that is capable of generating competitive CNNs from scratch, without relying on human-defined parameters or small search spaces. The experimental results show that the proposed approach results in competitive architectures obtained in a very short time. HMCNAS provides a step towards generalizing NAS, by providing a way to create competitive models, without requiring any human knowledge about the specific task.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16149
PDF
-
A Compositional Model of Consciousness based on Consciousness-Only
Abstract
tract: Scientific studies of consciousness rely on objects whose existence is independent of any consciousness. This theoretical-assumption leads to the "hard problem" of consciousness. We avoid this problem by assuming consciousness to be fundamental, and the main feature of consciousness is characterized as being other-dependent. We set up a framework which naturally subsumes the other-dependent feature by defining a compact closed category where morphisms represent conscious processes. These morphisms are a composition of a set of generators, each being specified by their relations with other generators, and therefore other-dependent. The framework is general enough, i.e. parameters in the morphisms take values in arbitrary commutative semi-rings, from which any finitely dimensional system can be dealt with. Our proposal fits well into a compositional model of consciousness and is an important step forward that addresses both the hard problem of consciousness and the combination problem of (proto)-panpsychism.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16138
PDF
-
Computer-aided Tumor Diagnosis in Automated Breast Ultrasound using 3D Detection Network
2020-07-31 15:25:07Junxiong Yu, Chaoyu Chen, Xin Yang, Yi Wang, Dan Yan, Jianxing Zhang, Dong NiAbstract
tract: Automated breast ultrasound (ABUS) is a new and promising imaging modality for breast cancer detection and diagnosis, which could provide intuitive 3D information and coronal plane information with great diagnostic value. However, manually screening and diagnosing tumors from ABUS images is very time-consuming and overlooks of abnormalities may happen. In this study, we propose a novel two-stage 3D detection network for locating suspected lesion areas and further classifying lesions as benign or malignant tumors. Specifically, we propose a 3D detection network rather than frequently-used segmentation network to locate lesions in ABUS images, thus our network can make full use of the spatial context information in ABUS images. A novel similarity loss is designed to effectively distinguish lesions from background. Then a classification network is employed to identify the located lesions as benign or malignant. An IoU-balanced classification loss is adopted to improve the correlation between classification and localization task. The efficacy of our network is verified from a collected dataset of 418 patients with 145 benign tumors and 273 malignant tumors. Experiments show our network attains a sensitivity of 97.66% with 1.23 false positives (FPs), and has an area under the curve(AUC) value of 0.8720.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16133
PDF
-
Robust Benchmarking for Machine Learning of Clinical Entity Extraction
2020-07-31 15:14:05Monica Agrawal, Chloe O'Connell, Yasmin Fatemi, Ariel Levy, David SontagAbstract
tract: Clinical studies often require understanding elements of a patient's narrative that exist only in free text clinical notes. To transform notes into structured data for downstream use, these elements are commonly extracted and normalized to medical vocabularies. In this work, we audit the performance of and indicate areas of improvement for state-of-the-art systems. We find that high task accuracies for clinical entity normalization systems on the 2019 n2c2 Shared Task are misleading, and underlying performance is still brittle. Normalization accuracy is high for common concepts (95.3%), but much lower for concepts unseen in training data (69.3%). We demonstrate that current approaches are hindered in part by inconsistencies in medical vocabularies, limitations of existing labeling schemas, and narrow evaluation techniques. We reformulate the annotation framework for clinical entity extraction to factor in these issues to allow for robust end-to-end system benchmarking. We evaluate concordance of annotations from our new framework between two annotators and achieve a Jaccard similarity of 0.73 for entity recognition and an agreement of 0.83 for entity normalization. We propose a path forward to address the demonstrated need for the creation of a reference standard to spur method development in entity recognition and normalization.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16127
PDF
-
Exploring Image Enhancement for Salient Object Detection in Low Light Images
2020-07-31 15:09:03Xin Xu, Shiqin Wang, Zheng Wang, Xiaolong Zhang, Ruimin HuAbstract
tract: Low light images captured in a non-uniform illumination environment usually are degraded with the scene depth and the corresponding environment lights. This degradation results in severe object information loss in the degraded image modality, which makes the salient object detection more challenging due to low contrast property and artificial light influence. However, existing salient object detection models are developed based on the assumption that the images are captured under a sufficient brightness environment, which is impractical in real-world scenarios. In this work, we propose an image enhancement approach to facilitate the salient object detection in low light images. The proposed model directly embeds the physical lighting model into the deep neural network to describe the degradation of low light images, in which the environment light is treated as a point-wise variate and changes with local content. Moreover, a Non-Local-Block Layer is utilized to capture the difference of local content of an object against its local neighborhood favoring regions. To quantitative evaluation, we construct a low light Images dataset with pixel-level human-labeled ground-truth annotations and report promising results on four public datasets and our benchmark dataset.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16124
PDF
-
COLD: Towards the Next Generation of Pre-Ranking System
2020-07-31 15:06:43Zhe Wang, Liqin Zhao, Biye Jiang, Guorui Zhou, Xiaoqiang Zhu, Kun GaiAbstract
tract: Multi-stage cascade architecture exists widely in many industrial systems such as recommender systems and online advertising, which often consists of sequential modules including matching, pre-ranking, ranking, etc. For a long time, it is believed pre-ranking is just a simplified version of the ranking module, considering the larger size of the candidate set to be ranked. Thus, efforts are made mostly on simplifying ranking model to handle the explosion of computing power for online inference. In this paper, we rethink the challenge of the pre-ranking system from an algorithm-system co-design view. Instead of saving computing power with restriction of model architecture which causes loss of model performance, here we design a new pre-ranking system by joint optimization of both the pre-ranking model and the computing power it costs. We name it COLD (Computing power cost-aware Online and Lightweight Deep pre-ranking system). COLD beats SOTA in three folds: (i) an arbitrary deep model with cross features can be applied in COLD under a constraint of controllable computing power cost. (ii) computing power cost is explicitly reduced by applying optimization tricks for inference acceleration. This further brings space for COLD to apply more complex deep models to reach better performance. (iii) COLD model works in an online learning and severing manner, bringing it excellent ability to handle the challenge of the data distribution shift. Meanwhile, the fully online pre-ranking system of COLD provides us with a flexible infrastructure that supports efficient new model developing and online A/B testing.Since 2019, COLD has been deployed in almost all products involving the pre-ranking module in the display advertising system in Alibaba, bringing significant improvements.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16122
PDF
-
Pop Music Transformer: Beat-based Modeling and Generation of Expressive Pop Piano Compositions
Abstract
tract: A great number of deep learning based models have been recently proposed for automatic music composition. Among these models, the Transformer stands out as a prominent approach for generating expressive classical piano performance with a coherent structure of up to one minute. The model is powerful in that it learns abstractions of data on its own, without much human-imposed domain knowledge or constraints. In contrast with this general approach, this paper shows that Transformers can do even better for music modeling, when we improve the way a musical score is converted into the data fed to a Transformer model. In particular, we seek to impose a metrical structure in the input data, so that Transformers can be more easily aware of the beat-bar-phrase hierarchical structure in music. The new data representation maintains the flexibility of local tempo changes, and provides hurdles to control the rhythmic and harmonic structure of music. With this approach, we build a Pop Music Transformer that composes Pop piano music with better rhythmic structure than existing Transformer models.
Abstract (translated)
URL
https://arxiv.org/abs/2002.00212
PDF
-
Lookahead and Hybrid Sample Allocation Procedures for Multiple Attribute Selection Decisions
Abstract
tract: Attributes provide critical information about the alternatives that a decision-maker is considering. When their magnitudes are uncertain, the decision-maker may be unsure about which alternative is truly the best, so measuring the attributes may help the decision-maker make a better decision. This paper considers settings in which each measurement yields one sample of one attribute for one alternative. When given a fixed number of samples to collect, the decision-maker must determine which samples to obtain, make the measurements, update prior beliefs about the attribute magnitudes, and then select an alternative. This paper presents the sample allocation problem for multiple attribute selection decisions and proposes two sequential, lookahead procedures for the case in which discrete distributions are used to model the uncertain attribute magnitudes. The two procedures are similar but reflect different quality measures (and loss functions), which motivate different decision rules: (1) select the alternative with the greatest expected utility and (2) select the alternative that is most likely to be the truly best alternative. We conducted a simulation study to evaluate the performance of the sequential procedures and hybrid procedures that first allocate some samples using a uniform allocation procedure and then use the sequential, lookahead procedure. The results indicate that the hybrid procedures are effective; allocating many (but not all) of the initial samples with the uniform allocation procedure not only reduces overall computational effort but also selects alternatives that have lower average opportunity cost and are more often truly best.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16119
PDF
-
Physical Adversarial Attack on Vehicle Detector in the Carla Simulator
2020-07-31 15:04:45Tong Wu, Xuefei Ning, Wenshuo Li, Ranran Huang, Huazhong Yang, Yu WangAbstract
tract: In this paper, we tackle the issue of physical adversarial examples for object detectors in the wild. Specifically, we proposed to generate adversarial patterns to be applied on vehicle surface so that it's not recognizable by detectors in the photo-realistic Carla simulator. Our approach contains two main techniques, an \textit{Enlarge-and-Repeat} process and a \textit{Discrete Searching} method, to craft mosaic-like adversarial vehicle textures without access to neither the model weight of the detector nor a differential rendering procedure. The experimental results demonstrate the effectiveness of our approach in the simulator.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16118
PDF
-
Predictability and Fairness in Social Sensing
2020-07-31 15:04:08Ramen Ghosh, Jakub Marecek, Wynita M. Griggs, Matheus Souza, Robert N. ShortenAbstract
tract: In many applications, one may benefit from the collaborative collection of data for sensing a physical phenomenon, which is known as social sensing. We show how to make social sensing (1) predictable, in the sense of guaranteeing that the number of queries per participant will be independent of the initial state, in expectation, even when the population of participants varies over time, and (2) fair, in the sense of guaranteeing that the number of queries per participant will be equalised among the participants, in expectation, even when the population of participants varies over time. In a use case, we consider a large, high-density network of participating parked vehicles. When awoken by an administrative centre, this network proceeds to search for moving, missing entities of interest using RFID-based techniques. We regulate the number and geographical distribution of the parked vehicles that are "Switched On" and thus actively searching for the moving entity of interest. In doing so, we seek to conserve vehicular energy consumption while, at the same time, maintaining good geographical coverage of the city such that the moving entity of interest is likely to be located within an acceptable time frame. Which vehicle participants are "Switched On" at any point in time is determined periodically through the use of stochastic techniques. This is illustrated on the example of a missing Alzheimer's patient in Melbourne, Australia.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16117
PDF
-
Neural Architecture Search as Sparse Supernet
2020-07-31 14:51:52Yan Wu, Aoming Liu, Zhiwu Huang, Siwei Zhang, Luc Van GoolAbstract
tract: This paper aims at enlarging the problem of Neural Architecture Search from Single-Path and Multi-Path Search to automated Mixed-Path Search. In particular, we model the new problem as a sparse supernet with a new continuous architecture representation using a mixture of sparsity constraints, i.e., Sparse Group Lasso. The sparse supernet is expected to automatically achieve sparsely-mixed paths upon a compact set of nodes. To optimize the proposed sparse supernet, we exploit a hierarchical accelerated proximal gradient algorithm within a bi-level optimization framework. Extensive experiments on CIFAR-10, CIFAR-100, Tiny ImageNet and ImageNet demonstrate that the proposed methodology is capable of searching for compact, general and powerful neural architectures.
Abstract (translated)
URL
https://arxiv.org/abs/2007.16112
PDF
