Abstract

With improved digital literacy and cheaper data connections, crowdsourcing has become more prevalent. Crowdsourcing tasks such as image tagging, language translation, etc. have contributors with varied skill level. Question and answers, software development/testing and product design are more serious tasks and require expert contributions. With increasing security threats, crowdsourcing in ’security’ domain is seen as an upcoming trend. In the survey, opportunities and challenges in crowdsourcing security are analyzed based on a survey conducted with 229 participants. The results provide insights on motivation, quality mechanisms and improvement areas for increased participation. ’Question and Answers’ that includes information sharing and ’software development and testing’ that includes Bug Bounty are identified as top crowdsourcing tasks by security specialists although they are concerned about security and privacy of information. Among quality attributes analyzed, contributor ’credibility’ is identified as a key are

Abstract

The problem of estimating scene-depth from a single im- age has seen great progress lately. Recent unsupervised methods are based on view-synthesis and learn depth by minimizing photometric reconstruction error. In this pa- per, we introduce Structured Adversarial Training (StrAT) to this problem. We generate multiple novel views using depth (or disparity), with the stereo-baseline changing in an increasing order. Adversarial training that goes from easy examples to harder ones produces richer losses and better models. The impact of StrAT is shown to exceed tra- ditional data augmentation using random new views. The combination of an adversarial framework, multiview learn- ing, and structured adversarial training produces state-of- the-art performance on unsupervised depth estimation for monocular images. The StrAT framework can benefit sev- eral problems that use adversarial training

Abstract

The current procedure for assessing emergency situational awareness in the USA requires Police, Paramedic and Fire First Responders to go on site. This observation is termed “size up”. It is effective, but not efficient. Using advanced information and communication for size-up would save Responders time and allow them to begin actual response faster- possibly saving lives. The commercial software that exists for size up is not in wide use. We have developed a software application called an Indoor-Outdoor Viewer for size-up for emergencies so that Responders can get information about the incident before they arrive at the site. It is an easy-to-use Viewer that shows building surroundings and their interiors, and allows Responders to mark the map to organize their approach. It will be compatible with real-time emergency data streams.

Abstract

Cognitive radio is an innovative technology that possesses profound potential to mitigate spectrum scarcity in the next-generation wireless networks. In this technology, unlicensed users are capable of accessing unused portions of the spectrum, popularly known as white spaces, for communication that enhances that spectrum usage. Each of these unlicensed user nodes must incorporate spectrum sensor to detect white space for transmitting and receiving the information. Such spectrum sensors must be portable and it needs to support wide bandwidth suitable for contemporary wireless networks. This paper presents a new frequency-domain energy-detector for spectrum sensing that is hardware efficient. We propose VLSI architecture for this detector where the area and power consuming FFT unit has been replaced by an iterative FFT- computation unit with simple architecture. This area-efficient design also improves the performance of detector in real-world noisy environment. In addition, we extend the conventional single-band detection-approach to a wide-band detection ap- proach and present its VLSI architecture. Performance analysis of such detection in AWGN channel environment showed that the detection probability of 0.16 is achieved at a SNR of -20 dB. Hardware implementation of the proposed energy detector is carried out using FPGA and its real world testing plus verification is performed using USRP software-defined radio- platform. It could operate with the bandwidth as high as 150 MHz with a sensing time of 7 ms. Compared to the state-of- the-art implementation, our prototype consumes 46% lesser hardware resources and does not require any memory in its architecture. Keywords-Cognitive radio; spectrum sensing; energy det

Abstract

Long-short term memory networks(LSTM) models have shown considerable performance on variety of problems dealing with sequential data. In this paper, we propose a variant of Long-Term Recurrent Convolutional Network(LRCN) to detect road intersection. We call this network as IntersectNet. We pose road intersection detection as binary classification task over sequence of frames. The model combines deep hierarchical visual feature extractor with recurrent sequence model. The model is end to end trainable with capability of capturing the temporal dynamics of the system. We exploit this capability to identify road intersection in a sequence of temporally consistent images. The model has been rigorously trained and tested on various different datasets. We think that our findings could be useful to model behavior of autonomous agent in the real-world.

Abstract

Accurate localization of other traffic participants is a vital task in autonomous driving systems. State-of-the-art systems employ a combination of sensing modalities such as RGB cameras and LiDARs for localizing traffic participants, but monocular localization demonstrations have been confined to plain roads. We demonstrate — to the best of our knowledge — the first results for monocular object localization and shape estimation on surfaces that are non-coplanar with the moving ego vehicle mounted with a monocular camera. We approximate road surfaces by local planar patches and use semantic cues from vehicles in the scene to initialize a local bundle-adjustment like procedure that simultaneously estimates the 3D pose and shape of the vehicles, and the orientation of the local ground plane on which the vehicle stands. We also demonstrate that our approach transfers from synthetic to real data, without any hyperparameter-/fine-tuning. We evaluate the proposed approach on the KITTI and SYNTHIA-SF benchmarks, for a variety of road plane configurations. The proposed approach significantly improves the state-of-the-art

Abstract

We are given an environment with some objects (a city block area) and mobile agents moving in the environment. An agent (hider) can hide behind an object to be not seen by other agents (seekers) through their line of sight (visibility). The aim of hiders is not to be caught for the longest time, and the aim of the seekers is catch all of them in the shortest period of time. We formulate the problem by using visibility based map abstractions. Agents plan their moves by utilizing multi-armed bandits UCB reward update model. We evaluate our abstractions and strategies by simulating the game under various different scenarios.

Abstract

Motion planning for manipulators under task space constraints is difficult as it constrains the joint config- urations to always lie on an implicitly defined manifold. It is possible to view task constrained motion planning as an opti- mization problem with non-linear equality constraints, which can be solved by general non-linear optimization techniques. In this paper, we present a novel custom optimizer which exploits the underlying structure present in many task constraints. At the core of our approach are some simple reformulations, which when coupled with the method of alternating projection, leads to an efficient convex optimization based routine for computing a feasible solution to the task constraints. We subsequently build on this result and use the concept of Augmented Lagrangian to guide the feasible solutions towards those that also minimize the user defined cost function. We show that the proposed optimizer is fully distributive and thus, can be easily parallelized. We validate our formulation on some common robotic benchmark problems. In particular, we show that the proposed optimizer achieves cyclic motion in the joint space corresponding to a similar nature trajectory in the task space. Furthermore, as a baseline, we compare the proposed optimizer with an off-the-shelf non-linear solver provide in open source package SciPy. We show that for similar task constraint residuals and smoothness cost, it can be upto more than three times faster than the SciPy alternative

Abstract

Recently there have been a large number of studies on embedding large-scale information networks using low-dimensional, neighborhood and community aware node representations. Though the performance of these embedding models have been better than traditional methods for graph mining applications, little is known about what these representations encode, or why a particular node representation works better for certain tasks. Our work presented here constitutes the first step in decoding the black-box of vector embeddings of nodes by evaluating their effectiveness in encoding elementary properties of a node such as page rank, degree, closeness centrality, clustering coefficient, etc. We believe that a node representation is effective for an application only if it encodes the application-specific elementary properties of nodes. To unpack the elementary properties encoded in a node representation, we evaluate the representations on the accuracy with which they can model each of these properties. Our extensive study of three state-of-the-art node representation models (DeepWalk, node2vec and LINE) on four different tasks and six diverse graphs reveal that node2vec and LINE best encode the network properties of sparse and dense graphs respectively. We correlate the model performance obtained for elementary property prediction tasks with the high-level downstream applications such as link prediction and node classification, and visualize the task performance vector of each model to understand the semantic similarity between the embeddings learned by various models. Our first study of the node embedding models for outlier detection reveals that node2vec and DeepWalk identify outliers well for sparse and dense graphs respectively. Our analysis highlights that the proposed elementary property prediction tasks help in unearthing the important features responsible for the given node embedding model to perform well for a given downstream task. This understanding would facilitate in picking the right model for a given downstream task.

Abstract

We use automatic speech recognition to assess spoken English learner pronunciation based on the authentic intelligibility of the learners' spoken responses determined from support vector machine (SVM) classifier or deep learning neural network model predictions of transcription correctness. Using numeric features produced by PocketSphinx alignment mode and many recognition passes searching for the substitution and deletion of each expected phoneme and insertion of unexpected phonemes in sequence, the SVM models achieve 82% agreement with the accuracy of Amazon Mechanical Turk crowdworker transcriptions, up from 75% reported by multiple independent researchers. Using such features with SVM classifier probability prediction models can help computer-aided pronunciation teaching (CAPT) systems provide intelligibility remediation.