What is a good visual representation for autonomous agents? We address this question in the context of semantic visual navigation, which is the problem of a robot finding its way through a complex environment to a target object, e.g. go to the refrigerator. Instead of acquiring a metric semantic map of an environment and using planning for navigation, our approach learns navigation policies on top of representations that capture spatial layout and semantic contextual cues. We propose to using high level semantic and contextual features including segmentation and detection masks obtained by off-the-shelf state-of-the-art vision as observations and use deep network to learn the navigation policy. This choice allows using additional data, from orthogonal sources, to better train different parts of the model the representation extraction is trained on large standard vision datasets while the navigation component leverages large synthetic environments for training. This combination of real and synthetic is possible because equitable feature representations are available in both (e.g., segmentation and detection masks), which alleviates the need for domain adaptation. Both the representation and the navigation policy can be readily applied to real non-synthetic environments as demonstrated on the Active Vision Dataset [1]. Our approach gets successfully to the target in 54% of the cases in unexplored environments, compared to 46% for non-learning based approach, and 28% for the learning-based baseline.
http://arxiv.org/abs/1805.06066
Removing undesired reflections from images taken through the glass is of great importance in computer vision. It serves as a means to enhance the image quality for aesthetic purposes as well as to preprocess images in machine learning and pattern recognition applications. We propose a convex model to suppress the reflection from a single input image. Our model implies a partial differential equation with gradient thresholding, which is solved efficiently using Discrete Cosine Transform. Extensive experiments on synthetic and real-world images demonstrate that our approach achieves desirable reflection suppression results and dramatically reduces the execution time compared to the state of the art.
https://arxiv.org/abs/1903.03889
Many robotic applications require the agent to perform long-horizon tasks in partially observable environments. In such applications, decision making at any step can depend on observations received far in the past. Hence, being able to properly memorize and utilize the long-term history is crucial. In this work, we propose a novel memory-based policy, named Scene Memory Transformer (SMT). The proposed policy embeds and adds each observation to a memory and uses the attention mechanism to exploit spatio-temporal dependencies. This model is generic and can be efficiently trained with reinforcement learning over long episodes. On a range of visual navigation tasks, SMT demonstrates superior performance to existing reactive and memory-based policies by a margin.
https://arxiv.org/abs/1903.03878
It is incredibly easy for a system designer to misspecify the objective for an autonomous system (“robot’’), thus motivating the desire to have the robot learn the objective from human behavior instead. Recent work has suggested that people have an interest in the robot performing well, and will thus behave pedagogically, choosing actions that are informative to the robot. In turn, robots benefit from interpreting the behavior by accounting for this pedagogy. In this work, we focus on misspecification: we argue that robots might not know whether people are being pedagogic or literal and that it is important to ask which assumption is safer to make. We cast objective learning into the more general form of a common-payoff game between the robot and human, and prove that in any such game literal interpretation is more robust to misspecification. Experiments with human data support our theoretical results and point to the sensitivity of the pedagogic assumption.
http://arxiv.org/abs/1903.03877
We present a solution for the goal of extracting a video from a single motion blurred image to sequentially reconstruct the clear views of a scene as beheld by the camera during the time of exposure. We first learn motion representation from sharp videos in an unsupervised manner through training of a convolutional recurrent video autoencoder network that performs a surrogate task of video reconstruction. Once trained, it is employed for guided training of a motion encoder for blurred images. This network extracts embedded motion information from the blurred image to generate a sharp video in conjunction with the trained recurrent video decoder. As an intermediate step, we also design an efficient architecture that enables real-time single image deblurring and outperforms competing methods across all factors: accuracy, speed, and compactness. Experiments on real scenes and standard datasets demonstrate the superiority of our framework over the state-of-the-art and its ability to generate a plausible sequence of temporally consistent sharp frames.
http://arxiv.org/abs/1804.02913
Word embeddings are widely used in NLP for a vast range of tasks. It was shown that word embeddings derived from text corpora reflect gender biases in society. This phenomenon is pervasive and consistent across different word embedding models, causing serious concern. Several recent works tackle this problem, and propose methods for significantly reducing this gender bias in word embeddings, demonstrating convincing results. However, we argue that this removal is superficial. While the bias is indeed substantially reduced according to the provided bias definition, the actual effect is mostly hiding the bias, not removing it. The gender bias information is still reflected in the distances between “gender-neutralized” words in the debiased embeddings, and can be recovered from them. We present a series of experiments to support this claim, for two debiasing methods. We conclude that existing bias removal techniques are insufficient, and should not be trusted for providing gender-neutral modeling.
https://arxiv.org/abs/1903.03862
One of the challenging aspects of incorporating deep neural networks into robotic systems is the lack of uncertainty measures associated with their output predictions. Recent work has identified aleatoric and epistemic as two types of uncertainty in the output of deep neural networks, and provided methods for their estimation. However, these methods have had limited success when applied to the object detection task. This paper introduces, BayesOD, a Bayesian approach for estimating the uncertainty in the output of deep object detectors, which reformulates the neural network inference and Non-Maximum suppression components of standard object detectors from a Bayesian perspective. As a result, BayesOD provides uncertainty estimates associated with detected object instances, which allows the deep object detector to be treated as any other sensor in a robotic system. BayesOD is shown to be capable of reliably identifying erroneous detection output instances using their estimated uncertainty measure. The estimated uncertainty measures are also shown to be better correlated with the correctness of a detection than the state of the art methods available in literature.
https://arxiv.org/abs/1903.03838
As the computational power of toady’s devices increases, real-time physically-based rendering becomes possible, and is rapidly gaining attention across a variety of domains. These include gaming, where physically-based rendering enhances immersion and overall entertainment experience, all the way to medicine, where it constitutes a powerful tool for intuitive volumetric data visualization. However, leveraging the obvious benefits of physically-based rendering (also referred to as photo-realistic rendering) remains challenging on embedded devices such as optical see-through head-mounted displays because of their limited computational power, and restricted memory usage and power consumption. We propose methods that aim at overcoming these limitations, fueling the implementation of real-time physically-based rendering on embedded devices. We navigate the compromise between memory requirement, computational power, and image quality to achieve reasonable rendering results by introducing a flexible representation of plenoptic functions and adapting a fast approximation algorithm for image generation from our plenoptic functions. We conclude by discussing potential applications and limitations of the proposed method.
https://arxiv.org/abs/1903.03837
Modelling of contact-rich tasks is challenging and cannot be entirely solved using classical control approaches due to the difficulty of constructing an analytic description of the contact dynamics. Additionally, in a manipulation task like food-cutting, purely learning-based methods such as Reinforcement Learning, require either a vast amount of data that is expensive to collect on a real robot, or a highly realistic simulation environment, which is currently not available. This paper presents a data-driven control approach that employs a recurrent neural network to model the dynamics for a Model Predictive Controller. We extend on previous work that was limited to torque-controlled robots by incorporating Force/Torque sensor measurements and formulate the control problem so that it can be applied to the more common velocity controlled robots. We evaluate the performance on objects used for training, as well as on unknown objects, by means of the cutting rates achieved and demonstrate that the method can efficiently treat different cases with only one dynamic model. Finally we investigate the behavior of the system during force-critical instances of cutting and illustrate its adaptive behavior in difficult cases.
https://arxiv.org/abs/1903.03831
Safe autonomous landing in urban cities is a necessity for the growing Unmanned Aircraft Systems (UAS) industry. In urgent situations, building rooftops, particularly flat rooftops, can provide local safe landing zones for small UAS. This paper investigates the real-time identification and selection of safe landing zones on rooftops based on LiDAR and camera sensor feedback. A visual high fidelity simulated city is constructed in the Unreal game engine, with particular attention paid to accurately generating rooftops and the common obstructions found thereon, e.g., ac units, water towers, air vents. AirSim, a robotic simulator plugin for Unreal, offers drone simulation and control and is capable of outputting video and LiDAR sensor data streams from the simulated Unreal world. A neural network is trained on randomized simulated cities to provide a pixel classification model. A novel algorithm is presented which finds the optimum obstacle-free landing position on nearby rooftops by fusing LiDAR and vision data.
https://arxiv.org/abs/1903.03829
We introduce Interpolation Consistency Training (ICT), a simple and computation efficient algorithm for training Deep Neural Networks in the semi-supervised learning paradigm. ICT encourages the prediction at an interpolation of unlabeled points to be consistent with the interpolation of the predictions at those points. In classification problems, ICT moves the decision boundary to low-density regions of the data distribution. Our experiments show that ICT achieves state-of-the-art performance when applied to standard neural network architectures on the CIFAR-10 and SVHN benchmark datasets.
http://arxiv.org/abs/1903.03825
Locomotion planning for legged systems requires reasoning about suitable contact schedules. The contact sequence and timings constitute a hybrid dynamical system and prescribe a subset of achievable motions. State-of-the-art approaches cast motion planning as an optimal control problem. In order to decrease computational complexity, one common strategy separates footstep planning from motion optimization and plans contacts using heuristics. In this paper, we propose to learn contact schedule selection from high-level task descriptors using Bayesian optimization. A bi-level optimization is defined in which a Gaussian process model predicts the performance of trajectories generated by a motion planning nonlinear program. The agent, therefore, retains the ability to reason about suitable contact schedules, while explicit computation of the corresponding gradients is avoided. We delineate the algorithm in its general form and provide results for planning single-legged hopping. Our method is capable of learning contact schedule transitions that align with human intuition. It performs competitively against a heuristic baseline in predicting task appropriate contact schedules.
https://arxiv.org/abs/1903.03823
A great deal of work aims to discover general purpose models of image interest or memorability for visual search and information retrieval. This paper argues that image interest is often domain and user specific, and that mechanisms for learning about this domain-specific image interest as quickly as possible, while limiting the amount of data-labelling required, are often more useful to end-users. Specifically, this paper is concerned with the small to medium-sized data regime regularly faced by practising data scientists, who are often required to build turnkey models for end-users with domain-specific challenges. This work uses pairwise image comparisons to reduce the labelling burden on these users, and shows that Gaussian process smoothing in image feature space can be used to build probabilistic models of image interest extremely quickly for a wide range of problems, and performs similarly to recent deep learning approaches trained using pairwise ranking losses. The Gaussian process model used in this work interpolates image interest inferred using a Bayesian ranking approach over image features extracted using a pre-trained convolutional neural network. This probabilistic approach produces image interests paired with uncertainties that can be used to identify images for which additional labelling is required and measure inference convergence. Results obtained on five distinct datasets reinforce recent findings that pre-trained convolutional neural networks can be used to extract useful representations applicable across multiple domains, and highlight the fact that domain-specific image interest does not always correlate with concepts like image memorability.
http://arxiv.org/abs/1706.05850
The online programing services, such as Github,TopCoder, and EduCoder, have promoted a lot of social interactions among the service users. However, the existing social interactions is rather limited and inefficient due to the rapid increasing of source-code repositories, which is difficult to explore manually. The emergence of source-code mining provides a promising way to analyze those source codes, so that those source codes can be relatively easy to understand and share among those service users. Among all the source-code mining attempts,program classification lays a foundation for various tasks related to source-code understanding, because it is impossible for a machine to understand a computer program if it cannot classify the program correctly. Although numerous machine learning models, such as the Natural Language Processing (NLP) based models and the Abstract Syntax Tree (AST) based models, have been proposed to classify computer programs based on their corresponding source codes, the existing works cannot fully characterize the source codes from the perspective of both the syntax and semantic information. To address this problem, we proposed a Graph Neural Network (GNN) based model, which integrates data flow and function call information to the AST,and applies an improved GNN model to the integrated graph, so as to achieve the state-of-art program classification accuracy. The experiment results have shown that the proposed work can classify programs with accuracy over 97%.
http://arxiv.org/abs/1903.03804
We present a method to address the challenging problem of segmentation of multi-modality isointense infant brain MR images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). Our method is based on context-guided, multi-stream fully convolutional networks (FCN), which after training, can directly map a whole volumetric data to its volume-wise labels. In order to alleviate the poten-tial gradient vanishing problem during training, we designed multi-scale deep supervision. Furthermore, context infor-mation was used to further improve the performance of our method. Validated on the test data of the MICCAI 2017 Grand Challenge on 6-month infant brain MRI segmentation (iSeg-2017), our method achieved an average Dice Overlap Coefficient of 95.4%, 91.6% and 89.6% for CSF, GM and WM, respectively.
http://arxiv.org/abs/1711.10212
Normalization methods improve both optimization and generalization of ConvNets. To further boost performance, the recently-proposed switchable normalization (SN) provides a new perspective for deep learning: it learns to select different normalizers for different convolution layers of a ConvNet. However, SN uses softmax function to learn importance ratios to combine normalizers, leading to redundant computations compared to a single normalizer. This work addresses this issue by presenting Sparse Switchable Normalization (SSN) where the importance ratios are constrained to be sparse. Unlike $\ell_1$ and $\ell_0$ constraints that impose difficulties in optimization, we turn this constrained optimization problem into feed-forward computation by proposing SparsestMax, which is a sparse version of softmax. SSN has several appealing properties. (1) It inherits all benefits from SN such as applicability in various tasks and robustness to a wide range of batch sizes. (2) It is guaranteed to select only one normalizer for each normalization layer, avoiding redundant computations. (3) SSN can be transferred to various tasks in an end-to-end manner. Extensive experiments show that SSN outperforms its counterparts on various challenging benchmarks such as ImageNet, Cityscapes, ADE20K, and Kinetics.
https://arxiv.org/abs/1903.03793
Three-dimensional Morphable Models (3DMMs) are powerful statistical tools for representing the 3D surfaces of an object class. In this context, we identify an interesting question that has previously not received research attention: is it possible to combine two or more 3DMMs that (a) are built using different templates that perhaps only partly overlap, (b) have different representation capabilities and (c) are built from different datasets that may not be publicly-available? In answering this question, we make two contributions. First, we propose two methods for solving this problem: i. use a regressor to complete missing parts of one model using the other, ii. use the Gaussian Process framework to blend covariance matrices from multiple models. Second, as an example application of our approach, we build a new face-and-head shape model that combines the variability and facial detail of the LSFM with the full head modelling of the LYHM. The resulting combined shape model achieves state-of-the-art performance and outperforms existing head models by a large margin. Finally, as an application experiment, we reconstruct full head representations from single, unconstrained images by utilizing our proposed large-scale model in conjunction with the FaceWarehouse blendshapes for handling expressions.
https://arxiv.org/abs/1903.03785
Achieving good speed and accuracy trade-off on target platform is very important in deploying deep neural networks. Most existing automatic architecture search approaches only pursue high performance but ignores such an important factor. In this work, we propose an algorithm “Partial Order Pruning” to prune architecture search space with partial order assumption, quickly lift the boundary of speed/accuracy trade-off on target platform, and automatically search the architecture with the best speed and accuracy trade-off. Our algorithm explicitly take profile information about the inference speed on target platform into consideration. With the proposed algorithm, we present several “Dongfeng” networks that provide high accuracy and fast inference speed on various application GPU platforms. By further searching decoder architecture, our DF-Seg real-time segmentation models yields state-of-the-art speed/accuracy trade-off on both embedded device and high-end GPU.
https://arxiv.org/abs/1903.03777
Large scale knowledge graph embedding has attracted much attention from both academia and industry in the field of Artificial Intelligence. However, most existing methods concentrate solely on fact triples contained in the given knowledge graph. Inspired by the fact that logic rules can provide a flexible and declarative language for expressing rich background knowledge, it is natural to integrate logic rules into knowledge graph embedding, to transfer human knowledge to entity and relation embedding, and strengthen the learning process. In this paper, we propose a novel logic rule-enhanced method which can be easily integrated with any translation based knowledge graph embedding model, such as TransE . We first introduce a method to automatically mine the logic rules and corresponding confidences from the triples. And then, to put both triples and mined logic rules within the same semantic space, all triples in the knowledge graph are represented as first-order logic. Finally, we define several operations on the first-order logic and minimize a global loss over both of the mined logic rules and the transformed first-order logics. We conduct extensive experiments for link prediction and triple classification on three datasets: WN18, FB166, and FB15K. Experiments show that the rule-enhanced method can significantly improve the performance of several baselines. The highlight of our model is that the filtered Hits@1, which is a pivotal evaluation in the knowledge inference task, has a significant improvement (up to 700% improvement).
http://arxiv.org/abs/1903.03772
Currently, many intelligence systems contain the texts from multi-sources, e.g., bulletin board system (BBS) posts, tweets and news. These texts can be comparative'' since they may be semantically correlated and thus provide us with different perspectives toward the same topics or events. To better organize the multi-sourced texts and obtain more comprehensive knowledge, we propose to study the novel problem of Mutual Clustering on Comparative Texts (MCCT), which aims to cluster the comparative texts simultaneously and collaboratively. The MCCT problem is difficult to address because 1) comparative texts usually present different data formats and structures and thus they are hard to organize, and 2) there lacks an effective method to connect the semantically correlated comparative texts to facilitate clustering them in an unified way. To this aim, in this paper we propose a Heterogeneous Information Network-based Text clustering framework HINT. HINT first models multi-sourced texts (e.g. news and tweets) as heterogeneous information networks by introducing the sharedanchor texts’’ to connect the comparative texts. Next, two similarity matrices based on HINT as well as a transition matrix for cross-text-source knowledge transfer are constructed. Comparative texts clustering are then conducted by utilizing the constructed matrices. Finally, a mutual clustering algorithm is also proposed to further unify the separate clustering results of the comparative texts by introducing a clustering consistency constraint. We conduct extensive experimental on three tweets-news datasets, and the results demonstrate the effectiveness and robustness of the proposed method in addressing the MCCT problem.
https://arxiv.org/abs/1903.03762
Indoor scenes exhibit rich hierarchical structure in 3D object layouts. Many tasks in 3D scene understanding can benefit from reasoning jointly about the hierarchical context of a scene, and the identities of objects. We present a variational denoising recursive autoencoder (VDRAE) that generates and iteratively refines a hierarchical representation of 3D object layouts, interleaving bottom-up encoding for context aggregation and top-down decoding for propagation. We train our VDRAE on large-scale 3D scene datasets to predict both instance-level segmentations and a 3D object detections from an over-segmentation of an input point cloud. We show that our VDRAE improves object detection performance on real-world 3D point cloud datasets compared to baselines from prior work.
https://arxiv.org/abs/1903.03757
The IEEE Low-Power Image Recognition Challenge (LPIRC) is an annual competition started in 2015 that encourages joint hardware and software solutions for computer vision systems with low latency and power. Track 1 of the competition in 2018 focused on the innovation of software solutions with fixed inference engine and hardware. This decision allows participants to submit models online and not worry about building and bringing custom hardware on-site, which attracted a historically large number of submissions. Among the diverse solutions, the winning solution proposed a quantization-friendly framework for MobileNets that achieves an accuracy of 72.67% on the holdout dataset with an average latency of 27ms on a single CPU core of Google Pixel2 phone, which is superior to the best real-time MobileNet models at the time.
http://arxiv.org/abs/1803.08607
Visual tracking is fragile in some difficult scenarios, for instance, appearance ambiguity and variation, occlusion can easily degrade most of visual trackers to some extent. In this paper, visual tracking is empowered with wireless positioning to achieve high accuracy while maintaining robustness. Fundamentally different from the previous works, this study does not involve any specific wireless positioning algorithms. Instead, we use the confidence region derived from the wireless positioning Cramer-Rao bound (CRB) as the search region of visual trackers. The proposed framework is low-cost and very simple to implement, yet readily leads to enhanced and robustified visual tracking performance in difficult scenarios as corroborated by our experimental results. Most importantly, it is utmost valuable for the practioners to pre-evaluate how effectively can the wireless resources available at hand alleviate the visual tracking pains.
https://arxiv.org/abs/1903.03736
Dynamic scenes that contain both object motion and egomotion are a challenge for monocular visual odometry (VO). Another issue with monocular VO is the scale ambiguity, i.e. these methods cannot estimate scene depth and camera motion in real scale. Here, we propose a learning based approach to predict camera motion parameters directly from optic flow, by marginalizing depthmap variations and outliers. This is achieved by learning a sparse overcomplete basis set of egomotion in an autoencoder network, which is able to eliminate irrelevant components of optic flow for the task of camera parameter or motionfield estimation. The model is trained using a sparsity regularizer and a supervised egomotion loss, and achieves the state-of-the-art performances on trajectory prediction and camera rotation prediction tasks on KITTI and Virtual KITTI datasets, respectively. The sparse latent space egomotion representation learned by the model is robust and requires only 5% of the hidden layer neurons to maintain the best trajectory prediction accuracy on KITTI dataset. Additionally, in presence of depth information, the proposed method demonstrates faithful object velocity prediction for wide range of object sizes and speeds by global compensation of predicted egomotion and a divisive normalization procedure.
https://arxiv.org/abs/1903.03731
Explainability and effectiveness are two key aspects for building recommender systems. Prior efforts mostly focus on incorporating side information to achieve better recommendation performance. However, these methods have some weaknesses: (1) prediction of neural network-based embedding methods are hard to explain and debug; (2) symbolic, graph-based approaches (e.g., meta path-based models) require manual efforts and domain knowledge to define patterns and rules, and ignore the item association types (e.g. substitutable and complementary). In this paper, we propose a novel joint learning framework to integrate \textit{induction of explainable rules from knowledge graph} with \textit{construction of a rule-guided neural recommendation model}. The framework encourages two modules to complement each other in generating effective and explainable recommendation: 1) inductive rules, mined from item-centric knowledge graphs, summarize common multi-hop relational patterns for inferring different item associations and provide human-readable explanation for model prediction; 2) recommendation module can be augmented by induced rules and thus have better generalization ability dealing with the cold-start issue. Extensive experiments\footnote{Code and data can be found at: \url{https://github.com/THUIR/RuleRec}} show that our proposed method has achieved significant improvements in item recommendation over baselines on real-world datasets. Our model demonstrates robust performance over “noisy” item knowledge graphs, generated by linking item names to related entities.
http://arxiv.org/abs/1903.03714
Age prediction based on appearances of different anatomies in medical images has been clinically explored for many decades. In this paper, we used deep learning to predict a persons age on Chest X-Rays. Specifically, we trained a CNN in regression fashion on a large publicly available dataset. Moreover, for interpretability, we explored activation maps to identify which areas of a CXR image are important for the machine (i.e. CNN) to predict a patients age, offering insight. Overall, amongst correctly predicted CXRs, we see areas near the clavicles, shoulders, spine, and mediastinum being most activated for age prediction, as one would expect biologically. Amongst incorrectly predicted CXRs, we have qualitatively identified disease patterns that could possibly make the anatomies appear older or younger than expected. A further technical and clinical evaluation would improve this work. As CXR is the most commonly requested imaging exam, a potential use case for estimating age may be found in the preventative counseling of patient health status compared to their age-expected average, particularly when there is a large discrepancy between predicted age and the real patient age.
http://arxiv.org/abs/1903.06542
To perform complex tasks, robots must be able to interact with and manipulate their surroundings. One of the key challenges in accomplishing this is robust state estimation during physical interactions, where the state involves not only the robot and the object being manipulated, but also the state of the contact itself. In this work, within the context of planar pushing, we extend previous inference-based approaches to state estimation in several ways. We estimate the robot, object, and the contact state on multiple manipulation platforms configured with a vision-based articulated model tracker, and either a biomimetic tactile sensor or a force-torque sensor. We show how to fuse raw measurements from the tracker and tactile sensors to jointly estimate the trajectory of the kinematic states and the forces in the system via probabilistic inference on factor graphs, in both batch and incremental settings. We perform several benchmarks with our framework and show how performance is affected by incorporating various geometric and physics based constraints, occluding vision sensors, or injecting noise in tactile sensors. We also compare with prior work on multiple datasets and demonstrate that our approach can effectively optimize over multi-modal sensor data and reduce uncertainty to find better state estimates.
https://arxiv.org/abs/1903.03699
In standard reinforcement learning, each new skill requires a manually-designed reward function, which takes considerable manual effort and engineering. Self-supervised goal setting has the potential to automate this process, enabling an agent to propose its own goals and acquire skills that achieve these goals. However, such methods typically rely on manually-designed goal distributions, or heuristics to force the agent to explore a wide range of states. We propose a formal exploration objective for goal-reaching policies that maximizes state coverage. We show that this objective is equivalent to maximizing the entropy of the goal distribution together with goal reaching performance, where goals correspond to entire states. We present an algorithm called Skew-Fit for learning such a maximum-entropy goal distribution, and show that under certain regularity conditions, our method converges to a uniform distribution over the set of possible states, even when we do not know this set beforehand. Skew-Fit enables self-supervised agents to autonomously choose and practice diverse goals. Our experiments show that it can learn a variety of manipulation tasks from images, including opening a door with a real robot, entirely from scratch and without any manually-designed reward function.
https://arxiv.org/abs/1903.03698
We consider the problem of vehicle routing for Autonomous Mobility-on-Demand (AMoD) systems, wherein a fleet of self-driving vehicles provides on-demand mobility in a given environment. Specifically, the task it to compute routes for the vehicles (both customer-carrying and empty travelling) so that travel demand is fulfilled and operational cost is minimized. The routing process must account for congestion effects affecting travel times, as modeled via a volume-delay function (VDF). Route planning with VDF constraints is notoriously challenging, as such constraints compound the combinatorial complexity of the routing optimization process. Thus, current solutions for AMoD routing resort to relaxations of the congestion constraints, thereby trading optimality with computational efficiency. In this paper, we present the first computationally-efficient approach for AMoD routing where VDF constraints are explicitly accounted for. We demonstrate that our approach is faster by at least one order of magnitude with respect to the state of the art, while providing higher quality solutions. From a methodological standpoint, the key technical insight is to establish a mathematical reduction of the AMoD routing problem to the classical traffic assignment problem (a related vehicle-routing problem where empty traveling vehicles are not present). Such a reduction allows us to extend powerful algorithmic tools for traffic assignment, which combine the classic Frank-Wolfe algorithm with modern techniques for pathfinding, to the AMoD routing problem. We provide strong theoretical guarantees for our approach in terms of near-optimality of the returned solution.
https://arxiv.org/abs/1903.03697
Images today are increasingly shared online on social networking sites such as Facebook, Flickr, Foursquare, and Instagram. Despite that current social networking sites allow users to change their privacy preferences, this is often a cumbersome task for the vast majority of users on the Web, who face difficulties in assigning and managing privacy settings. Thus, automatically predicting images’ privacy to warn users about private or sensitive content before uploading these images on social networking sites has become a necessity in our current interconnected world. In this paper, we explore learning models to automatically predict appropriate images’ privacy as private or public using carefully identified image-specific features. We study deep visual semantic features that are derived from various layers of Convolutional Neural Networks (CNNs) as well as textual features such as user tags and deep tags generated from deep CNNs. Particularly, we extract deep (visual and tag) features from four pre-trained CNN architectures for object recognition, i.e., AlexNet, GoogLeNet, VGG-16, and ResNet, and compare their performance for image privacy prediction. Results of our experiments on a Flickr dataset of over thirty thousand images show that the learning models trained on features extracted from ResNet outperform the state-of-the-art models for image privacy prediction. We further investigate the combination of user tags and deep tags derived from CNN architectures using two settings: (1) SVM on the bag-of-tags features; and (2) text-based CNN. Our results show that even though the models trained on the visual features perform better than those trained on the tag features, the combination of deep visual features with image tags shows improvements in performance over the individual feature sets.
https://arxiv.org/abs/1903.03695
For enterprise, personal and societal applications, there is now an increasing demand for automated authentication of identity from images using computer vision. However, current authentication technologies are still vulnerable to presentation attacks. We present RoPAD, an end-to-end deep learning model for presentation attack detection that employs unsupervised adversarial invariance to ignore visual distractors in images for increased robustness and reduced overfitting. Experiments show that the proposed framework exhibits state-of-the-art performance on presentation attack detection on several benchmark datasets.
https://arxiv.org/abs/1903.03691
Recent progress in reinforcement learning (RL) using self-game-play has shown remarkable performance on several board games (e.g., Chess and Go) as well as video games (e.g., Atari games and Dota2). It is plausible to consider that RL, starting from zero knowledge, might be able to gradually approximate a winning strategy after a certain amount of training. In this paper, we explore neural Monte-Carlo-Tree-Search (neural MCTS), an RL algorithm which has been applied successfully by DeepMind to play Go and Chess at a super-human level. We try to leverage the computational power of neural MCTS to solve a class of combinatorial optimization problems. Following the idea of Hintikka’s Game-Theoretical Semantics, we propose the Zermelo Gamification (ZG) to transform specific combinatorial optimization problems into Zermelo games whose winning strategies correspond to the solutions of the original optimization problem. The ZG also provides a specially designed neural MCTS. We use a combinatorial planning problem for which the ground-truth policy is efficiently computable to demonstrate that ZG is promising.
http://arxiv.org/abs/1903.03674
This paper describes a system whereby a robot detects and track human-meaningful navigational cues as it navigates in an indoor environment. It is intended as the sensor front-end for a mobile robot system that can communicate its navigational context with human users. From simulated LiDAR scan data we construct a set of 2D occupancy grid bitmaps, then hand-label these with human-scale navigational features such as closed doors, open corridors and intersections. We train a Convolutional Neural Network (CNN) to recognize these features on input bitmaps. In our demonstration system, these features are detected at every time step then passed to a tracking module that does frame-to-frame data association to improve detection accuracy and identify stable unique features. We evaluate the system in both simulation and the real world. We compare the performance of using input occupancy grids obtained directly from LiDAR data, or incrementally constructed with SLAM, and their combination.
https://arxiv.org/abs/1903.03669
To improve efficiency and reduce failures in autonomous vehicles, research has focused on developing robust and safe learning methods that take into account disturbances in the environment. Existing literature in robust reinforcement learning poses the learning problem as a two player game between the autonomous system and disturbances. This paper examines two different algorithms to solve the game, Robust Adversarial Reinforcement Learning and Neural Fictitious Self Play, and compares performance on an autonomous driving scenario. We extend the game formulation to a semi-competitive setting and demonstrate that the resulting adversary better captures meaningful disturbances that lead to better overall performance. The resulting robust policy exhibits improved driving efficiency while effectively reducing collision rates compared to baseline control policies produced by traditional reinforcement learning methods.
https://arxiv.org/abs/1903.03642
Robustness of deep learning models is a property that has recently gained increasing attention. We formally define a notion of robustness for generative adversarial models, and show that, perhaps surprisingly, the GAN in its original form is not robust. Indeed, the discriminator in GANs may be viewed as merely offering “teaching feedback”. Our notion of robustness relies on a dishonest discriminator, or noisy, adversarial interference with its feedback. We explore, theoretically and empirically, the effect of model and training properties on this robustness. In particular, we show theoretical conditions for robustness that are supported by empirical evidence. We also test the effect of regularization. Our results suggest variations of GANs that are indeed more robust to noisy attacks and have more stable training behavior, requiring less regularization in general. Inspired by our theoretical results, we further extend our framework to obtain a class of models related to WGAN, with good empirical performance. Overall, this work introduces a novel perspective on designing GAN models from the viewpoint of robustness.
https://arxiv.org/abs/1802.09700
When creating benchmarks for SAT solvers, we need SAT instances that are easy to build but hard to solve. A recent development in the search for such methods has led to the Balanced SAT algorithm, which can create k-SAT instances with m clauses of high difficulty, for arbitrary k and m. In this paper we introduce the No-Triangle SAT algorithm, a SAT instance generator based on the cluster coefficient graph statistic. We empirically compare the two algorithms by fixing the arity and the number of variables, but varying the number of clauses. The hardest instances that we find are produced by No-Triangle SAT. Furthermore, difficult instances from No-Triangle SAT have a different number of clauses than difficult instances from Balanced SAT, potentially allowing a combination of the two methods to find hard SAT instances for a larger array of parameters.
http://arxiv.org/abs/1903.03592
Distance preserving visualization techniques have emerged as one of the fundamental tools for data analysis. One example are the techniques that arrange data instances into two-dimensional grids so that the pairwise distances among the instances are preserved into the produced layouts. Currently, the state-of-the-art approaches produce such grids by solving assignment problems or using permutations to optimize cost functions. Although precise, such strategies are computationally expensive, limited to small datasets or being dependent on specialized hardware to speed up the process. In this paper, we present a new technique, called Distance-preserving Grid (DGrid), that employs a binary space partitioning process in combination with multidimensional projections to create orthogonal regular grid layouts. Our results show that DGrid is as precise as the existing state-of-the-art techniques whereas requiring only a fraction of the running time and computational resources.
http://arxiv.org/abs/1903.06262
Much of the literature on robotic perception focuses on the visual modality. Vision provides a global observation of a scene, making it broadly useful. However, in the domain of robotic manipulation, vision alone can sometimes prove inadequate: in the presence of occlusions or poor lighting, visual object identification might be difficult. The sense of touch can provide robots with an alternative mechanism for recognizing objects. In this paper, we study the problem of touch-based instance recognition. We propose a novel framing of the problem as multi-modal recognition: the goal of our system is to recognize, given a visual and tactile observation, whether or not these observations correspond to the same object. To our knowledge, our work is the first to address this type of multi-modal instance recognition problem on such a large-scale with our analysis spanning 98 different objects. We employ a robot equipped with two GelSight touch sensors, one on each finger, and a self-supervised, autonomous data collection procedure to collect a dataset of tactile observations and images. Our experimental results show that it is possible to accurately recognize object instances by touch alone, including instances of novel objects that were never seen during training. Our learned model outperforms other methods on this complex task, including that of human volunteers.
http://arxiv.org/abs/1903.03591
Reinforcement Learning (RL) algorithms can suffer from poor sample efficiency when rewards are delayed and sparse. We introduce a solution that enables agents to learn temporally extended actions at multiple levels of abstraction in a sample efficient and automated fashion. Our approach combines universal value functions and hindsight learning, allowing agents to learn policies belonging to different time scales in parallel. We show that our method significantly accelerates learning in a variety of discrete and continuous tasks.
http://arxiv.org/abs/1805.08180
This paper addresses the problem of inferring unseen cross-domain and cross-modal image-to-image translations between multiple domains and modalities. We assume that only some of the pairwise translations have been seen (i.e. trained) and infer the remaining unseen translations (where training pairs are not available). We propose mix and match networks, an approach where multiple encoders and decoders are aligned in such a way that the desired translation can be obtained by simply cascading the source encoder and the target decoder, even when they have not interacted during the training stage (i.e. unseen). The main challenge lies in the alignment of the latent representations at the bottlenecks of encoder-decoder pairs. We propose an architecture with several tools to encourage alignment, including autoencoders and robust side information and latent consistency losses. We show the benefits of our approach in terms of effectiveness and scalability compared with other pairwise image-to-image translation approaches. We also propose zero-pair cross-modal image translation, a challenging setting where the objective is inferring semantic segmentation from depth (and vice-versa) without explicit segmentation-depth pairs, and only from two (disjoint) segmentation-RGB and depth-segmentation training sets. We observe that certain part of the shared information between unseen domains might not be reachable, so we further propose a variant that leverages pseudo-pairs to exploit all shared information.
https://arxiv.org/abs/1903.04294
Planning high-speed trajectories for UAVs in unknown environments requires extremely fast algorithms able to solve the trajectory generation problem in real-time in order to be able to react quickly to the changing knowledge of the world, but that guarantee safety at all times. The desire of maintaining computational tractability typically leads to optimization problems that do not include the obstacles (collision checks are done on the solutions) or to formulations that use a convex decomposition of the free space and then impose an ad hoc allocation of each interval of the trajectory in a specific polyhedron. Moreover, safety guarantees are usually obtained by having a local planner that plans a trajectory with a final “stop” condition in the free-known space. However, these two decisions typically lead to slow and conservative trajectories. We propose FaSTrap (Fast and Safe Trajectory Planner) to overcome these issues. FasTrap obtains faster trajectories by enabling the local planner to optimize in both free-known and unknown spaces. Safety guarantees are ensured by always having a feasible, safe back-up trajectory in the free-known space at the start of each replanning step. Furthermore, we present a Mixed Integer Quadratic Problem (MIQP) formulation in which the solver can choose the interval allocation and where a heuristics for the time allocation is computed efficiently using the result of the previous replanning iteration. This proposed algorithm is tested both in simulation and in real hardware, showing agile flights in unknown cluttered environments.
http://arxiv.org/abs/1903.03558
The paper addresses the problem of energy compaction of dense 4D light fields by designing geometry-aware local graph-based transforms. Local graphs are constructed on super-rays that can be seen as a grouping of spatially and geometry-dependent angularly correlated pixels. Both non separable and separable transforms are considered. Despite the local support of limited size defined by the super-rays, the Laplacian matrix of the non separable graph remains of high dimension and its diagonalization to compute the transform eigen vectors remains computationally expensive. To solve this problem, we then perform the local spatio-angular transform in a separable manner. We show that when the shape of corresponding super-pixels in the different views is not isometric, the basis functions of the spatial transforms are not coherent, resulting in decreased correlation between spatial transform coefficients. We hence propose a novel transform optimization method that aims at preserving angular correlation even when the shapes of the super-pixels are not isometric. Experimental results show the benefit of the approach in terms of energy compaction. A coding scheme is also described to assess the rate-distortion perfomances of the proposed transforms and is compared to state of the art encoders namely HEVC and JPEG Pleno VM 1.1.
http://arxiv.org/abs/1903.03556
Graph-based transforms have been shown to be powerful tools in terms of image energy compaction. However, when the support increases to best capture signal dependencies, the computation of the basis functions becomes rapidly untractable. This problem is in particular compelling for high dimensional imaging data such as light fields. The use of local transforms with limited supports is a way to cope with this computational difficulty. Unfortunately, the locality of the support may not allow us to fully exploit long term signal dependencies present in both the spatial and angular dimensions in the case of light fields. This paper describes sampling and prediction schemes with local graph-based transforms enabling to efficiently compact the signal energy and exploit dependencies beyond the local graph support. The proposed approach is investigated and is shown to be very efficient in the context of spatio-angular transforms for quasi-lossless compression of light fields.
http://arxiv.org/abs/1903.03546
Classical deformable registration techniques achieve impressive results and offer a rigorous theoretical treatment, but are computationally intensive since they solve an optimization problem for each image pair. Recently, learning-based methods have facilitated fast registration by learning spatial deformation functions. However, these approaches use restricted deformation models, require supervised labels, or do not guarantee a diffeomorphic (topology-preserving) registration. Furthermore, learning-based registration tools have not been derived from a probabilistic framework that can offer uncertainty estimates. In this paper, we build a connection between classical and learning-based methods. We present a probabilistic generative model and derive an unsupervised learning-based inference algorithm that uses insights from classical registration methods and makes use of recent developments in convolutional neural networks (CNNs). We demonstrate our method on a 3D brain registration task for both images and anatomical surfaces, and provide extensive empirical analyses of the algorithm. Our principled approach results in state of the art accuracy and very fast runtimes, while providing diffeomorphic guarantees. Our implementation is available online at this http URL
http://arxiv.org/abs/1903.03545
Modern neural network-based algorithms are able to produce highly accurate depth estimates from stereo image pairs, nearly matching the reliability of measurements from more expensive depth sensors. However, this accuracy comes with a higher computational cost since these methods use network architectures designed to compute and process matching scores across all candidate matches at all locations, with floating point computations repeated across a match volume with dimensions corresponding to both space and disparity. This leads to longer running times to process each image pair, making them impractical for real-time use in robots and autonomous vehicles. We propose a new stereo algorithm that employs a significantly more efficient network architecture. Our method builds an initial match cost volume using traditional matching costs that are fast to compute, and trains a network to estimate disparity from this volume. Crucially, our network only employs per-pixel and two-dimensional convolution operations: to summarize the match information at each location as a low-dimensional feature vector, and to spatially process these `cost-signature’ features to produce a dense disparity map. Experimental results on the KITTI benchmark show that our method delivers competitive accuracy at significantly higher speeds—running at 48 frames per second on a modern GPU.
http://arxiv.org/abs/1903.04939
The recent advent of automated neural network architecture search led to several methods that outperform state-of-the-art human-designed architectures. However, these approaches are computationally expensive, in extreme cases consuming GPU years. We propose two novel methods which aim to expedite this optimization problem by transferring knowledge acquired from previous tasks to new ones. First, we propose a novel neural architecture selection method which employs this knowledge to identify strong and weak characteristics of neural architectures across datasets. Thus, these characteristics do not need to be rediscovered in every search, a strong weakness of current state-of-the-art searches. Second, we propose a method for learning curve extrapolation to determine if a training process can be terminated early. In contrast to existing work, we propose to learn from learning curves of architectures trained on other datasets to improve the prediction accuracy for novel datasets. On five different image classification benchmarks, we empirically demonstrate that both of our orthogonal contributions independently lead to an acceleration, without any significant loss in accuracy.
http://arxiv.org/abs/1903.03536
Data for human-human spoken dialogues for research and development are currently very limited in quantity, variety, and sources; such data is even scarcer in healthcare. In this work, we investigate fast prototyping of a dialogue comprehension system by leveraging on minimal nurse-to-patient conversations. We propose a framework inspired by nurse-initiated clinical symptom monitoring conversations to construct a simulated human-human dialogue dataset, embodying linguistic characteristics of spoken interactions like thinking aloud, self-contradiction, and topic drift. We then adapt an established bidirectional attention pointer network on this simulated dataset, achieving more than 80% F1 score on a held-out test set from real-world nurse-to-patient conversations. The ability to automatically comprehend conversations in the healthcare domain by exploiting only limited data has implications for improving clinical workflows with automatic summarization of consultations, red flag symptom detection and triaging capabilities. Our prototype demonstrates the feasibility for efficient and effective extraction, retrieval and comprehension of symptom checking information discussed in multi-turn human-human spoken conversations.
http://arxiv.org/abs/1903.03530
We describe the workflow of a digital surface models (DSMs) refinement algorithm using a hybrid conditional generative adversarial network (cGAN) where the generative part consists of two parallel networks merged at the last stage forming a WNet architecture. The inputs to the so-called WNet-cGAN are stereo DSMs and panchromatic (PAN) half-meter resolution satellite images. Fusing these helps to propagate fine detailed information from a spectral image and complete the missing 3D knowledge from a stereo DSM about building shapes. Besides, it refines the building outlines and edges making them more rectangular and sharp.
http://arxiv.org/abs/1903.03519
Although event-based cameras are already commercially available. Vision algorithms based on them are still not common. As a consequence, there are few Hardware Accelerators for them. In this work we present some experiments to create FPGA accelerators for a well-known vision algorithm using event-based cameras. We present a stereo matching algorithm to create a stream of disparity events disparity map and implement several accelerators using the Intel FPGA OpenCL tool-chain. The results show that multiple designs can be easily tested and that a performance speedup of more than 8x can be achieved with simple code transformations.
http://arxiv.org/abs/1903.03509
A wide range of systems exhibit high dimensional incomplete data. Accurate estimation of the missing data is often desired, and is crucial for many downstream analyses. Many state-of-the-art recovery methods involve supervised learning using datasets containing full observations. In contrast, we focus on unsupervised estimation of missing image data, where no full observations are available - a common situation in practice. Unsupervised imputation methods for images often employ a simple linear subspace to capture correlations between data dimensions, omitting more complex relationships. In this work, we introduce a general probabilistic model that describes sparse high dimensional imaging data as being generated by a deep non-linear embedding. We derive a learning algorithm using a variational approximation based on convolutional neural networks and discuss its relationship to linear imputation models, the variational auto encoder, and deep image priors. We introduce sparsity-aware network building blocks that explicitly model observed and missing data. We analyze proposed sparsity-aware network building blocks, evaluate our method on public domain imaging datasets, and conclude by showing that our method enables imputation in an important real-world problem involving medical images. The code is freely available as part of the \verb|neuron| library at this http URL
http://arxiv.org/abs/1903.03503