Search

Agile quadrotor flight in challenging environments has the potential to revolutionize shipping, transportation, and search and rescue applications. Nonlinear model predictive control (NMPC) has recently shown promising results for agile quadrotor control, but relies on highly accurate models for maximum performance. Hence, model uncertainties in the form of unmodeled complex aerodynamic effects, varying payloads and parameter mismatch will degrade overall system performance. In this paper, we propose L1-NMPC, a novel hybrid adaptive NMPC to learn model uncertainties online and immediately compensate for them, drastically improving performance over the non-adaptive baseline with minimal computational overhead. Our proposed architecture generalizes to many different environments from which we evaluate wind, unknown payloads, and highly agile flight conditions. The proposed method demonstrates immense flexibility and robustness, with more than 90% tracking error reduction over non-adaptive NMPC under large unknown disturbances and without any gain tuning. In addition, the same controller with identical gains can accurately fly highly agile racing trajectories exhibiting top speeds of 70 km/h, offering tracking performance improvements of around 50% relative to the non-adaptive NMPC baseline., 8 Pages, 6 figures, Accepted RAL 2021

State-of-the-art face restoration methods employ deep convolutional neural networks (CNNs) to learn a mapping between degraded and sharp facial patterns by exploring local appearance knowledge. However, most of these methods do not well exploit facial structures and identity information, and only deal with task-specific face restoration (e.g., face super-resolution or deblurring). In this paper, we propose cross-tasks and cross-models plug-and-play 3D facial priors to explicitly embed the network with the sharp facial structures for general face restoration tasks. Our 3D priors are the first to explore 3D morphable knowledge based on the fusion of parametric descriptions of face attributes (e.g., identity, facial expression, texture, illumination, and face pose). Furthermore, the priors can easily be incorporated into any network and are very efficient in improving the performance and accelerating the convergence speed. Firstly, a 3D face rendering branch is set up to obtain 3D priors of salient facial structures and identity knowledge. Secondly, for better exploiting this hierarchical information (i.e., intensity similarity, 3D facial structure, and identity content), a spatial attention module is designed for the image restoration problems. Extensive face restoration experiments including face super-resolution and deblurring demonstrate that the proposed 3D priors achieve superior face restoration results over the state-of-the-art algorithms.

Predictability is a prerequisite for effective human control of artificial intelligence (AI). For example, the inability to predict the malfunctioning of AI impedes timely human intervention. In this paper, we employ a computerized navigation task, a lunar lander game, to investigate empirically how AI's predictability compares to humans' predictability. To this end, we ask participants to guess whether the landings of a spaceship in this game performed by AI and humans will succeed. We show that humans are worse at predicting AI performance than at predicting human performance in this environment. Significantly, participants underestimate the differences in the relative predictability of AI and, at times, overestimate their prediction skills. We link the difference in predictability to differences in the approaches, i.e. different landing patterns, employed by AI and humans to succeed in the task. These results highlight important differences in perception of AI and human with implications for human-computer interaction.

Fault-tolerant control is crucial for safety-critical systems, such as quadrotors. State-of-art flight controllers can stabilize and control a quadrotor even when subjected to the complete loss of a rotor. However, these methods rely on external sensors, such as GPS or motion capture systems, for state estimation. To the best of our knowledge, this has not yet been achieved with only onboard sensors. In this paper, we propose the first algorithm that combines fault-tolerant control and onboard vision-based state estimation to achieve position control of a quadrotor subjected to complete failure of one rotor. Experimental validations show that our approach is able to accurately control the position of a quadrotor during a motor failure scenario, without the aid of any external sensors. The primary challenge to vision-based state estimation stems from the inevitable high-speed yaw rotation (over 20 rad/s) of the damaged quadrotor, causing motion blur to cameras, which is detrimental to visual inertial odometry (VIO). We compare two types of visual inputs to the vision-based state estimation algorithm: standard frames and events. Experimental results show the advantage of using an event camera especially in low light environments due to its inherent high dynamic range and high temporal resolution. We believe that our approach will render autonomous quadrotors safer in both GPS denied or degraded environments. We release both our controller and VIO algorithm open source., Comment: 8 pages, 10 figures

Neural networks predictions are unreliable when the input sample is out of the training distribution or corrupted by noise. Being able to detect such failures automatically is fundamental to integrate deep learning algorithms into robotics. Current approaches for uncertainty estimation of neural networks require changes to the network and optimization process, typically ignore prior knowledge about the data, and tend to make over-simplifying assumptions which underestimate uncertainty. To address these limitations, we propose a novel framework for uncertainty estimation. Based on Bayesian belief networks and Monte-Carlo sampling, our framework not only fully models the different sources of prediction uncertainty, but also incorporates prior data information, e.g. sensor noise. We show theoretically that this gives us the ability to capture uncertainty better than existing methods. In addition, our framework has several desirable properties: (i) it is agnostic to the network architecture and task; (ii) it does not require changes in the optimization process; (iii) it can be applied to already trained architectures. We thoroughly validate the proposed framework through extensive experiments on both computer vision and control tasks, where we outperform previous methods by up to 23% in accuracy., Accepted for publication in the Robotics and Automation Letters 2020, and for presentation at the International Conference on Robotics and Automation (ICRA) 2020

Humans race drones faster than algorithms, despite being limited to a fixed camera angle, body rate control, and response latencies in the order of hundreds of milliseconds. A better understanding of the ability of human pilots of selecting appropriate motor commands from highly dynamic visual information may provide key insights for solving current challenges in vision-based autonomous navigation. This paper investigates the relationship between human eye movements, control behavior, and flight performance in a drone racing task. We collected a multimodal dataset from 21 experienced drone pilots using a highly realistic drone racing simulator, also used to recruit professional pilots. Our results show task-specific improvements in drone racing performance over time. In particular, we found that eye gaze tracks future waypoints (i.e., gates), with first fixations occurring on average 1.5 seconds and 16 meters before reaching the gate. Moreover, human pilots consistently looked at the inside of the future flight path for lateral (i.e., left and right turns) and vertical maneuvers (i.e., ascending and descending). Finally, we found a strong correlation between pilots eye movements and the commanded direction of quadrotor flight, with an average visual-motor response latency of 220 ms. These results highlight the importance of coordinated eye movements in human-piloted drone racing. We make our dataset publicly available., Comment: 8 pages, 6 figures

Autonomous car racing is a major challenge in robotics. It raises fundamental problems for classical approaches such as planning minimum-time trajectories under uncertain dynamics and controlling the car at the limits of its handling. Besides, the requirement of minimizing the lap time, which is a sparse objective, and the difficulty of collecting training data from human experts have also hindered researchers from directly applying learning-based approaches to solve the problem. In the present work, we propose a learning-based system for autonomous car racing by leveraging a high-fidelity physical car simulation, a course-progress proxy reward, and deep reinforcement learning. We deploy our system in Gran Turismo Sport, a world-leading car simulator known for its realistic physics simulation of different race cars and tracks, which is even used to recruit human race car drivers. Our trained policy achieves autonomous racing performance that goes beyond what had been achieved so far by the built-in AI, and, at the same time, outperforms the fastest driver in a dataset of over 50,000 human players., Comment: Accepted for Publication at the IEEE Robotics and Automation Letters (RA-L) 2021, and International Conference on Robots and Automation (ICRA) 2021

Segmentation of anatomical structures is a fundamental image analysis task for many applications in the medical field. Deep learning methods have been shown to perform well, but for this purpose large numbers of manual annotations are needed in the first place, which necessitate prohibitive levels of resources that are often unavailable. In an active learning framework of selecting informed samples for manual labeling, expert clinician time for manual annotation can be optimally utilized, enabling the establishment of large labeled datasets for machine learning. In this paper, we propose a novel method that combines representativeness with uncertainty in order to estimate ideal samples to be annotated, iteratively from a given dataset. Our novel representativeness metric is based on Bayesian sampling, by using information-maximizing autoencoders. We conduct experiments on a shoulder magnetic resonance imaging (MRI) dataset for the segmentation of four musculoskeletal tissue classes. Quantitative results show that the annotation of representative samples selected by our proposed querying method yields an improved segmentation performance at each active learning iteration, compared to a baseline method that also employs uncertainty and representativeness metrics. For instance, with only 10% of the dataset annotated, our method reaches within 5% of Dice score expected from the upper bound scenario of all the dataset given as annotated (an impractical scenario due to resource constraints), and this gap drops down to a mere 2% when less than a fifth of the dataset samples are annotated. Such active learning approach to selecting samples to annotate enables an optimal use of the expert clinician time, being often the bottleneck in realizing machine learning solutions in medicine., Comment: 10 pages, 7 figures

While supervised object detection and segmentation methods achieve impressive accuracy, they generalize poorly to images whose appearance significantly differs from the data they have been trained on. To address this when annotating data is prohibitively expensive, we introduce a self-supervised detection and segmentation approach that can work with single images captured by a potentially moving camera. At the heart of our approach lies the observation that object segmentation and background reconstruction are linked tasks, and that, for structured scenes, background regions can be re-synthesized from their surroundings, whereas regions depicting the moving object cannot. We encode this intuition into a self-supervised loss function that we exploit to train a proposal-based segmentation network. To account for the discrete nature of the proposals, we developed a Monte Carlo-based training strategy that allows the algorithm to explore the large space of object proposals. We apply our method to human detection and segmentation in images that visually depart from those of standard benchmarks and outperform existing self-supervised methods.

Neuromorphic chips that use spikes to encode information could provide fast and energy-efficient computing for ubiquitous embedded systems. A bio-plausible spike-timing solution for training spiking neural networks that makes the most of sparsity is implemented on the BrainScaleS-2 hardware platform.

Methods for processing point cloud information have seen a great success in collider physics applications. One recent breakthrough in machine learning is the usage of Transformer networks to learn semantic relationships between sequences in language processing. In this work, we apply a modified Transformer network called Point Cloud Transformer as a method to incorporate the advantages of the Transformer architecture to an unordered set of particles resulting from collision events. To compare the performance with other strategies, we study jet-tagging applications for highly-boosted particles., Comment: 12 pages, 3 figures

Data analysis in high-dimensional spaces aims at obtaining a synthetic description of a data set, revealing its main structure and its salient features. We here introduce an approach providing this description in the form of a topography of the data, namely a human-readable chart of the probability density from which the data are harvested. The approach is based on an unsupervised extension of Density Peak clustering and a non-parametric density estimator that measures the probability density in the manifold containing the data. This allows finding automatically the number and the height of the peaks of the probability density, and the depth of the "valleys" separating them. Importantly, the density estimator provides a measure of the error, which allows distinguishing genuine density peaks from density fluctuations due to finite sampling. The approach thus provides robust and visual information about the density peaks' height, their statistical reliability, and their hierarchical organization, offering a conceptually powerful extension of the standard clustering partitions. We show that this framework is particularly useful in the analysis of complex data sets., There is a Supplementary Information document in the ancillary files folder

While the field of algorithmic fairness has brought forth many ways to measure and improve the fairness of machine learning models, these findings are still not widely used in practice. We suspect that one reason for this is that the field of algorithmic fairness came up with a lot of definitions of fairness, which are difficult to navigate. The goal of this paper is to provide data scientists with an accessible introduction to group fairness metrics and to give some insight into the philosophical reasoning for caring about these metrics. We will do this by considering in which sense socio-demographic groups are compared for making a statement on fairness., Comment: Accepted full paper at SDS2021, the 8th Swiss Conference on Data Science

Chronic diseases can greatly benefit from bioelectronic medicine approaches. Neuromorphic electronic circuits present ideal characteristics for the development of brain-inspired low-power implantable processing systems that can be interfaced with biological systems. These circuits, therefore, represent a promising additional tool in the tool-set of bioelectronic medicine. In this paper, we describe the main features of neuromorphic circuits that are ideally suited for continuously monitoring the physiological parameters of the body and interact with them in real-time. We propose examples of computational primitives that can be used for real-time pattern generation and present a neuromorphic implementation of neural oscillators for the generation of sequence activation patterns. We demonstrate the features of such systems with an implementation of a three-phase network that models the dynamics of the respiratory Central Pattern Generator (CPG) and the heart chambers rhythm, and that could be used to build an adaptive pacemaker., Comment: 4 pages; 7 figures; to be published in 10th International IEEE EMBS Conference On Neural Engineering (NER'21), Special Session on Bioelectronics medicine

In this letter, we study the effects that perception latency has on the maximum speed a robot can reach to safely navigate through an unknown cluttered environment. We provide a general analysis that can serve as a baseline for future quantitative reasoning for design tradeoffs in autonomous robot navigation. We consider the case where the robot is modeled as a linear second-order system with bounded input and navigates through static obstacles. Also, we focus on a scenario where the robot wants to reach a target destination in as little time as possible, and therefore cannot change its longitudinal velocity to avoid obstacles. We show how the maximum latency that the robot can tolerate to guarantee safety is related to the desired speed, the range of its sensing pipeline, and the actuation limitations of the platform (i.e., the maximum acceleration it can produce). As a particular case study, we compare monocular and stereo frame-based cameras against novel, low-latency sensors, such as event cameras, in the case of quadrotor flight. To validate our analysis, we conduct experiments on a quadrotor platform equipped with an event camera to detect and avoid obstacles thrown towards the robot. To the best of our knowledge, this is the first theoretical work in which perception and actuation limitations are jointly considered to study the performance of a robotic platform in high-speed navigation.

Spiking neural networks have shown great promise for the design of low-power sensory-processing and edge-computing hardware platforms. However, implementing on-chip learning algorithms on such architectures is still an open challenge, especially for multi-layer networks that rely on the back-propagation algorithm. In this paper, we present a spike-based learning method that approximates back-propagation using local weight update mechanisms and which is compatible with mixed-signal analog/digital neuromorphic circuits. We introduce a network architecture that enables synaptic weight update mechanisms to back-propagate error signals across layers and present a network that can be trained to distinguish between two spike-based patterns that have identical mean firing rates, but different spike-timings. This work represents a first step towards the design of ultra-low power mixed-signal neuromorphic processing systems with on-chip learning circuits that can be trained to recognize different spatio-temporal patterns of spiking activity (e.g. produced by event-based vision or auditory sensors)., 2020 2nd IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS)

This paper presents a novel system for autonomous, vision-based drone racing combining learned data abstraction, nonlinear filtering, and time-optimal trajectory planning. The system has successfully been deployed at the first autonomous drone racing world championship: the 2019 AlphaPilot Challenge. Contrary to traditional drone racing systems, which only detect the next gate, our approach makes use of any visible gate and takes advantage of multiple, simultaneous gate detections to compensate for drift in the state estimate and build a global map of the gates. The global map and drift-compensated state estimate allow the drone to navigate through the race course even when the gates are not immediately visible and further enable to plan a near time-optimal path through the race course in real time based on approximate drone dynamics. The proposed system has been demonstrated to successfully guide the drone through tight race courses reaching speeds up to 8m/s and ranked second at the 2019 AlphaPilot Challenge., Autonomous Robots, 46 (1), ISSN:0929-5593, ISSN:1573-7527

Once an academic venture, autonomous driving has received unparalleled corporate funding in the last decade. Still, the operating conditions of current autonomous cars are mostly restricted to ideal scenarios. This means that driving in challenging illumination conditions such as night, sunrise, and sunset remains an open problem. In these cases, standard cameras are being pushed to their limits in terms of low light and high dynamic range performance. To address these challenges, we propose, DSEC, a new dataset that contains such demanding illumination conditions and provides a rich set of sensory data. DSEC offers data from a wide-baseline stereo setup of two color frame cameras and two high-resolution monochrome event cameras. In addition, we collect lidar data and RTK GPS measurements, both hardware synchronized with all camera data. One of the distinctive features of this dataset is the inclusion of high-resolution event cameras. Event cameras have received increasing attention for their high temporal resolution and high dynamic range performance. However, due to their novelty, event camera datasets in driving scenarios are rare. This work presents the first high-resolution, large-scale stereo dataset with event cameras. The dataset contains 53 sequences collected by driving in a variety of illumination conditions and provides ground truth disparity for the development and evaluation of event-based stereo algorithms., Comment: IEEE Robotics and Automation Letters

This paper proposes a fully-connected network training architecture called EILE targeting incremental learning on edge. By using a novel reconfigurable processing element (PE) architecture, EILE avoids explicit transposition of weight matrices required for backpropagation to preserve the same efficient memory access pattern for both the forward (FP) and backward propagation (BP) phases. Experimental results on a Zynq XC7Z100 FPGA with 64 PEs show that EILE achieves 19.2 GOp/s peak throughput and maintains nearly 100 % PE utilization efficiency for both FP and BP with batch sizes from 1 to 32. EILE’s small on-chip memory footprint and scalability to match any available off-chip memory bandwidth makes it an attractive ASIC architecture for energy-constrained training.

Event cameras are novel vision sensors that report per-pixel brightness changes as a stream of asynchronous "events". They offer significant advantages compared to standard cameras due to their high temporal resolution, high dynamic range and lack of motion blur. However, events only measure the varying component of the visual signal, which limits their ability to encode scene context. By contrast, standard cameras measure absolute intensity frames, which capture a much richer representation of the scene. Both sensors are thus complementary. However, due to the asynchronous nature of events, combining them with synchronous images remains challenging, especially for learning-based methods. This is because traditional recurrent neural networks (RNNs) are not designed for asynchronous and irregular data from additional sensors. To address this challenge, we introduce Recurrent Asynchronous Multimodal (RAM) networks, which generalize traditional RNNs to handle asynchronous and irregular data from multiple sensors. Inspired by traditional RNNs, RAM networks maintain a hidden state that is updated asynchronously and can be queried at any time to generate a prediction. We apply this novel architecture to monocular depth estimation with events and frames where we show an improvement over state-of-the-art methods by up to 30% in terms of mean absolute depth error. To enable further research on multimodal learning with events, we release EventScape, a new dataset with events, intensity frames, semantic labels, and depth maps recorded in the CARLA simulator.

Quadrotors are agile. Unlike most other machines, they can traverse extremely complex environments at high speeds. To date, only expert human pilots have been able to fully exploit their capabilities. Autonomous operation with on-board sensing and computation has been limited to low speeds. State-of-the-art methods generally separate the navigation problem into subtasks: sensing, mapping, and planning. While this approach has proven successful at low speeds, the separation it builds upon can be problematic for high-speed navigation in cluttered environments. Indeed, the subtasks are executed sequentially, leading to increased processing latency and a compounding of errors through the pipeline. Here we propose an end-to-end approach that can autonomously fly quadrotors through complex natural and man-made environments at high speeds, with purely onboard sensing and computation. The key principle is to directly map noisy sensory observations to collision-free trajectories in a receding-horizon fashion. This direct mapping drastically reduces processing latency and increases robustness to noisy and incomplete perception. The sensorimotor mapping is performed by a convolutional network that is trained exclusively in simulation via privileged learning: imitating an expert with access to privileged information. By simulating realistic sensor noise, our approach achieves zero-shot transfer from simulation to challenging real-world environments that were never experienced during training: dense forests, snow-covered terrain, derailed trains, and collapsed buildings. Our work demonstrates that end-to-end policies trained in simulation enable high-speed autonomous flight through challenging environments, outperforming traditional obstacle avoidance pipelines., Comment: 16 pages (+7 supplementary)

Enacting an AI system typically requires three iterative phases where AI engineers are in command: selection and preparation of the data, selection and configuration of algorithmic tools, and fine-tuning of the different parameters on the basis of intermediate results. Our main hypothesis is that these phases involve practices with ethical questions. This paper maps these ethical questions and proposes a way to address them in light of a neo-republican understanding of freedom, defined as absence of domination. We thereby identify different types of responsibility held by AI engineers and link them to concrete suggestions on how to improve professional practices. This paper contributes to the literature on AI and ethics by focusing on the work necessary to configure AI systems, thereby offering an input to better practices and an input for societal debates.

This paper describes the submission by the team from the Department of Computational Linguistics, Zurich University, to the Multilingual Grapheme-to-Phoneme Conversion (G2P) Task 1 of the SIGMORPHON 2021 challenge in the low and medium settings. The submission is a variation of our 2020 G2P system, which serves as the baseline for this year’s challenge. The system is a neural transducer that operates over explicit edit actions and is trained with imitation learning. For this challenge, we experimented with the following changes: a) emitting phoneme segments instead of single character phonemes, b) input character dropout, c) a mogrifier LSTM decoder (Melis et al., 2019), d) enriching the decoder input with the currently attended input character, e) parallel BiLSTM encoders, and f) an adaptive batch size scheduler. In the low setting, our best ensemble improved over the baseline, however, in the medium setting, the baseline was stronger on average, although for certain languages improvements could be observed.

Since 2015 all citizens in Taiwan and since 2018 also residents can use the online platform called “Join”. The platform allows registered users to propose, discuss and review governmental policies. The ongoing practice currently makes Join one of the most innovative and successful cases of digital policy co-creation worldwide. However, due to the language barrier little is known about its contents. Internationally available research results are scarce. This article provides an overview of the platform’s functionalities as well as an overview of the research published to date. In addition, a first descriptive inroad to online petition dynamics is made. The case studies are based on the dataset available on the government’s Open Data portal. The goal of this paper is to lay the groundwork and provide the context for a more comprehensive quantitative analysis of online petition signing dynamics in Taiwan.

Item does not contain fulltext Deep Artificial Neural Networks (ANNs) employ a simplified analog neuron model that mimics the rate transfer function of integrate-and-fire neurons. In Spiking Neural Networks (SNNs), the predominant information transmission method is based on rate codes. This code is inefficient from a hardware perspective because the number of transmitted spikes is proportional to the encoded analog value. Alternate codes such as temporal codes that are based on single spikes are difficult to scale up for large networks due to their sensitivity to spike timing noise. Here we present a study of an encoding scheme based on temporal spike patterns. This scheme inherits the efficiency of temporal codes but retains the robustness of rate codes. The pattern code is evaluated on MNIST, CIFAR-10, and ImageNet image classification tasks. We compare the network performance of ANNs, rate-coded SNNs, and temporal-coded SNNs, using the classification error and operation count as performance metrics. We also estimate the power consumption of the digital logic needed for the operations associated with each encoding type, and the impact of the bit precision of the weights and activations. On ImageNet, the temporal pattern code achieves up to 35× reduction in the estimated power consumption compared to the rate-coded SNN, and 42× compared to the ANN. The classification error of the pattern-coded SNN is increased by

Businesses were moving during the past decades to-ward full digital models, which made companies face new threats and cyberattacks affecting their services and, consequently, their profits. To avoid negative impacts, companies' investments in cybersecurity are increasing considerably. However, Small and Medium-sized Enterprises (SMEs) operate on small budgets, minimal technical expertise, and few personnel to address cybersecurity threats. In order to address such challenges, it is essential to promote novel approaches that can intuitively present cybersecurity-related technical information.This paper introduces SecBot, a cybersecurity-driven conversational agent (i.e., chatbot) for the support of cybersecurity planning and management. SecBot applies concepts of neural networks and Natural Language Processing (NLP), to interact and extract information from a conversation. SecBot can (a) identify cyberattacks based on related symptoms, (b) indicate solutions and configurations according to business demands, and (c) provide insightful information for the decision on cybersecurity investments and risks. A formal description had been developed to describe states, transitions, a language, and a Proof-of-Concept (PoC) implementation. A case study and a performance evaluation were conducted to provide evidence of the proposed solution's feasibility and accuracy.

We present a new algorithm for computing pure-strategy $\varepsilon$-Bayes-Nash equilibria ($\varepsilon$-BNEs) in combinatorial auctions with continuous value and action spaces. An essential innovation of our algorithm is to separate the algorithm's search phase (for finding the $\varepsilon$-BNE) from the verification phase (for computing the $\varepsilon$). Using this approach, we obtain an algorithm that is both very fast and provides theoretical guarantees on the $\varepsilon$ it finds. Our main technical contribution is a verification method which allows us to upper bound the $\varepsilon$ across the whole continuous value space without making assumptions about the mechanism. Using our algorithm, we can now compute $\varepsilon$-BNEs in multi-minded domains that are significantly more complex than what was previously possible to solve. We release our code under an open-source license to enable researchers to perform algorithmic analyses of auctions, to enable bidders to analyze different strategies, and to facilitate many other applications., Comment: 40 pages

Low-cost gas sensors allow for large-scale spatial monitoring of air quality in the environment. However they require calibration before deployment. Methods such as multivariate regression techniques have been applied towards sensor calibration. In this work, we propose instead, the use of deep learning methods, particularly, recurrent neural networks for predicting the gas concentrations based on the outputs of these sensors. This paper presents a first study of using Gated Recurrent Unit (GRU) neural network models for gas concentration prediction. The GRU networks achieve on average, a 44.69% and a 25.17% RMSE improvement in concentration prediction on a gas dataset when compared with Support Vector Regression (SVR) and Multilayer Perceptron (MLP) models respectively. With the current advances in deep network hardware accelerators, these networks can be combined with the sensors for a compact embedded system suitable for edge applications.

Despite the rapid development of robotic control theory, hardware motor controllers still suffer from some disadvantages: they are computationally-intensive and rely on powerful computing systems which are usually implemented using bulky and power-hungry devices. On the other hand, biological motor control systems are power-efficient, light-weight and robust. Neuromorphic engineering sheds a light on how to uncover biological control features that could lead to the design of lower power and less bulky controllers. In this paper, we present a closed-loop motor controller implemented on mixed-signal analog-digital neuromorphic hardware using a spiking neural network. The network performs PI control by encoding target, feedback and error signals using population coding. It continuously calculates the error through the network, which relates the three variables by means of feed-forward inter-population synapses. This biologically plausible and fault-tolerant strategy is ideally suited for the use of neuromorphic hardware that comprises noisy silicon neurons. Here we show how to optimize the network structure to make it robust to both noisy inputs and device mismatch. We provide experimental results showing how the controller can reach 97.1% accuracy with 75.8 ms average latency.

This paper presents a Gated Recurrent Unit (GRU) based recurrent neural network (RNN) accelerator called EdgeDRNN designed for portable edge computing. EdgeDRNN adopts the spiking neural network inspired delta network algorithm to exploit temporal sparsity in RNNs. It reduces off-chip memory access by a factor of up to 10x with tolerable accuracy loss. Experimental results on a 10 million parameter 2-layer GRU-RNN, with weights stored in DRAM, show that EdgeDRNN computes them in under 0.5 ms. With 2.42 W wall plug power on an entry level USB powered FPGA board, it achieves latency comparable with a 92 W Nvidia 1080 GPU. It outperforms NVIDIA Jetson Nano, Jetson TX2 and Intel Neural Compute Stick 2 in latency by 6X. For a batch size of 1, EdgeDRNN achieves a mean effective throughput of 20.2 GOp/s and a wall plug power efficiency that is over 4X higher than all other platforms., Comment: This paper has been accepted for publication at the IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS), Genoa, 2020

Visual Localization is one of the key enabling technologies for autonomous driving and augmented reality. High quality datasets with accurate 6 Degree-of-Freedom (DoF) reference poses are the foundation for benchmarking and improving existing methods. Traditionally, reference poses have been obtained via Structure-from-Motion (SfM). However, SfM itself relies on local features which are prone to fail when images were taken under different conditions, e.g., day/ night changes. At the same time, manually annotating feature correspondences is not scalable and potentially inaccurate. In this work, we propose a semi-automated approach to generate reference poses based on feature matching between renderings of a 3D model and real images via learned features. Given an initial pose estimate, our approach iteratively refines the pose based on feature matches against a rendering of the model from the current pose estimate. We significantly improve the nighttime reference poses of the popular Aachen Day-Night dataset, showing that state-of-the-art visual localization methods perform better (up to $47\%$) than predicted by the original reference poses. We extend the dataset with new nighttime test images, provide uncertainty estimates for our new reference poses, and introduce a new evaluation criterion. We will make our reference poses and our framework publicly available upon publication., 25 pages, 16 figures. Int J Comput Vis (2020)

In modern visual SLAM systems, it is a standard practice to retrieve potential candidate map points from overlapping keyframes for further feature matching or direct tracking. In this work, we argue that keyframes are not the optimal choice for this task, due to several inherent limitations, such as weak geometric reasoning and poor scalability. We propose a voxel-map representation to efficiently retrieve map points for visual SLAM. In particular, we organize the map points in a regular voxel grid. Visible points from a camera pose are queried by sampling the camera frustum in a raycasting manner, which can be done in constant time using an efficient voxel hashing method. Compared with keyframes, the retrieved points using our method are geometrically guaranteed to fall in the camera field-of-view, and occluded points can be identified and removed to a certain extend. This method also naturally scales up to large scenes and complicated multi-camera configurations. Experimental results show that our voxel map representation is as efficient as a keyframe map with 5 keyframes and provides significantly higher localization accuracy (average 46% improvement in RMSE) on the EuRoC dataset. The proposed voxel-map representation is a general approach to a fundamental functionality in visual SLAM and widely applicable.

Today's autonomous drones have reaction times of tens of milliseconds, which is not enough for navigating fast in complex dynamic environments. To safely avoid fast moving objects, drones need low-latency sensors and algorithms. We departed from state-of-the-art approaches by using event cameras, which are bioinspired sensors with reaction times of microseconds. Our approach exploits the temporal information contained in the event stream to distinguish between static and dynamic objects and leverages a fast strategy to generate the motor commands necessary to avoid the approaching obstacles. Standard vision algorithms cannot be applied to event cameras because the output of these sensors is not images but a stream of asynchronous events that encode per-pixel intensity changes. Our resulting algorithm has an overall latency of only 3.5 milliseconds, which is sufficient for reliable detection and avoidance of fast-moving obstacles. We demonstrate the effectiveness of our approach on an autonomous quadrotor using only onboard sensing and computation. Our drone was capable of avoiding multiple obstacles of different sizes and shapes, at relative speeds up to 10 meters/second, both indoors and outdoors.

Spiking Neural Networks (SNNs) are bio-inspired networks that process information conveyed as temporal spikes rather than numeric values. An example of a sensor providing such data is the event-camera. It only produces an event when a pixel reports a significant brightness change. Similarly, the spiking neuron of an SNN only produces a spike whenever a significant number of spikes occur within a short period of time. Due to their spike-based computational model, SNNs can process output from event-based, asynchronous sensors without any pre-processing at extremely lower power unlike standard artificial neural networks. This is possible due to specialized neuromorphic hardware that implements the highly-parallelizable concept of SNNs in silicon. Yet, SNNs have not enjoyed the same rise of popularity as artificial neural networks. This not only stems from the fact that their input format is rather unconventional but also due to the challenges in training spiking networks. Despite their temporal nature and recent algorithmic advances, they have been mostly evaluated on classification problems. We propose, for the first time, a temporal regression problem of numerical values given events from an event-camera. We specifically investigate the prediction of the 3- DOF angular velocity of a rotating event-camera with an SNN. The difficulty of this problem arises from the prediction of angular velocities continuously in time directly from irregular, asynchronous event-based input. Directly utilising the output of event-cameras without any pre-processing ensures that we inherit all the benefits that they provide over conventional cameras. That is high-temporal resolution, high-dynamic range and no motion blur. To assess the performance of SNNs on this task, we introduce a synthetic event-camera dataset generated from real-world panoramic images and show that we can successfully train an SNN to perform angular velocity regression.

Adding more cameras to SLAM systems improves robustness and accuracy but complicates the design of the visual front-end significantly. Thus, most systems in the literature are tailored for specific camera configurations. In this work, we aim at an adaptive SLAM system that works for arbitrary multi-camera setups. To this end, we revisit several common building blocks in visual SLAM. In particular, we propose an adaptive initialization scheme, a sensor-agnostic, information- theoretic keyframe selection algorithm, and a scalable voxel- based map. These techniques make little assumption about the actual camera setups and prefer theoretically grounded methods over heuristics. We adapt a state-of-the-art visual- inertial odometry with these modifications, and experimental results show that the modified pipeline can adapt to a wide range of camera setups (e.g., 2 to 6 cameras in one experiment) without the need of sensor-specific modifications or tuning.

Authorship attribution (i.e., determining who is the author of a piece of source code) is an established research topic. State-of-the-art results for the authorship attribution problem look promising for the software engineering field, where they could be applied to detect plagiarized code and prevent legal issues. With this article, we first introduce a new language-agnostic approach to authorship attribution of source code. Then, we discuss limitations of existing synthetic datasets for authorship attribution, and propose a data collection approach that delivers datasets that better reflect aspects important for potential practical use in software engineering. Finally, we demonstrate that high accuracy of authorship attribution models on existing datasets drastically drops when they are evaluated on more realistic data. We outline next steps for the design and evaluation of authorship attribution models that could bring the research efforts closer to practical use for software engineering., 13 pages

In this paper, we propose to augment image-based place recognition with structural cues. Specifically, these structural cues are obtained using structure-from-motion, such that no additional sensors are needed for place recognition. This is achieved by augmenting the 2D convolutional neural network (CNN) typically used for image-based place recognition with a 3D CNN that takes as input a voxel grid derived from the structure-from-motion point cloud. We evaluate different methods for fusing the 2D and 3D features and obtain best performance with global average pooling and simple concatenation. On the Oxford RobotCar dataset, the resulting descriptor exhibits superior recognition performance compared to descriptors extracted from only one of the input modalities, including state-of-the-art image-based descriptors. Especially at low descriptor dimensionalities, we outperform state-of-the-art descriptors by up to 90%., Comment: 8 pages, published in RA-L & IROS 2020

L-SCN is a two-layered Service-Centric Networking (SCN) architecture. The L-SCN design splits the network into domains and specifies communication protocols for service provider information propagation. Nodes in a domain receive substantial knowledge about the available resources (e.g., CPU, RAM) and available services within the domain, while the communication between different domains is realized through supernodes. We extend L-SCN with new communication mechanisms, which improve the processing time and provide lower protocol overhead for service request processing. The two proposed mechanisms are named event-driven and provider-driven. The event-driven mechanism propagates service provider information based on an event (e.g., high overload). The provider-driven mechanism propagates service provider information periodically.

We present EKLT, a feature tracking method that leverages the complementarity of event cameras and standard cameras to track visual features with high temporal resolution. Event cameras are novel sensors that output pixel-level brightness changes, called “events”. They offer significant advantages over standard cameras, namely a very high dynamic range, no motion blur, and a latency in the order of microseconds. However, because the same scene pattern can produce different events depending on the motion direction, establishing event correspondences across time is challenging. By contrast, standard cameras provide intensity measurements (frames) that do not depend on motion direction. Our method extracts features on frames and subsequently tracks them asynchronously using events, thereby exploiting the best of both types of data: the frames provide a photometric representation that does not depend on motion direction and the events provide updates with high temporal resolution. In contrast to previous works, which are based on heuristics, this is the first principled method that uses intensity measurements directly, based on a generative event model within a maximum-likelihood framework. As a result, our method produces feature tracks that are more accurate than the state of the art, across a wide variety of scenes.

Fluorescence microscopy allows for a detailed inspection of cells, cellular networks, and anatomical landmarks by staining with a variety of carefully-selected markers visualized as color channels. Quantitative characterization of structures in acquired images often relies on automatic image analysis methods. Despite the success of deep learning methods in other vision applications, their potential for fluorescence image analysis remains underexploited. One reason lies in the considerable workload required to train accurate models, which are normally specific for a given combination of markers, and therefore applicable to a very restricted number of experimental settings. We herein propose Marker Sampling and Excite, a neural network approach with a modality sampling strategy and a novel attention module that together enable (i) flexible training with heterogeneous datasets with combinations of markers and (ii) successful utility of learned models on arbitrary subsets of markers prospectively. We show that our single neural network solution performs comparably to an upper bound scenario where an ensemble of many networks is na\"ively trained for each possible marker combination separately. In addition, we demonstrate the feasibility of this framework in high-throughput biological analysis by revising a recent quantitative characterization of bone marrow vasculature in 3D confocal microscopy datasets and further confirm the validity of our approach on an additional, significantly different dataset of microvessels in fetal liver tissues. Not only can our work substantially ameliorate the use of deep learning in fluorescence microscopy analysis, but it can also be utilized in other fields with incomplete data acquisitions and missing modalities., Comment: Main: 21 pages, 6 figures, 1 table. Supplementary: 5 pages, 7 figures, 3 tables

Motivated by the astonishing capabilities of natural intelligent agents and inspired by theories from psychology, this paper explores the idea that perception gets coupled to 3D properties of the world via interaction with the environment. Existing works for depth estimation require either massive amounts of annotated training data or some form of hard-coded geometrical constraint. This paper explores a new approach to learning depth perception requiring neither of those. Specifically, we train a specialized global-local network architecture with what would be available to a robot interacting with the environment: from extremely sparse depth measurements down to even a single pixel per image. From a pair of consecutive images, our proposed network outputs a latent representation of the observer's motion between the images and a dense depth map. Experiments on several datasets show that, when ground truth is available even for just one of the image pixels, the proposed network can learn monocular dense depth estimation up to 22.5% more accurately than state-of-the-art approaches. We believe that this work, despite its scientific interest, lays the foundations to learn depth from extremely sparse supervision, which can be valuable to all robotic systems acting under severe bandwidth or sensing constraints., Comment: Accepted for Publication at the IEEE Robotics and Automation Letters (RA-L) 2020, and International Conference on Intelligent Robots and Systems (IROS) 2020

Stream Processing (SP) is an important Big Data technology enabling continuous querying of data streams. The stream setting offers the opportunity to exploit synergies and, theoretically, share the access and processing costs between multiple different collaborators. But what should be the monetary contribution of each consumer when they do not trust each other and have varying valuations of the differing outcomes? In this article, we present Collaborative Stream Processing (CSP), a model where the costs, which are set exogenously by providers, are shared between multiple consumers, the collaborators. For this, we identify three important requirements for CSP to establish trust between the collaborators and propose a CSP algorithm, ENCSPA, adhering to these requirements. Based on the collaborators’ outcome valuations and the costs of the raw data streams, ENCSPA computes the payment for each collaborator. At the same time, ENCSPA ensures that no collaborator has an incentive to manipulate the system by providing misinformation about her/his value, budget, or time limit. We show that ENCSPA can calculate payments in a reasonable amount of time for up to one thousand collaborators.

Civilian drones are soon expected to be used in a wide variety of tasks, such as aerial surveillance, delivery, or monitoring of existing architectures. Nevertheless, their deployment in urban environments has so far been limited. Indeed, in unstructured and highly dynamic scenarios, drones face numerous challenges to navigate autonomously in a feasible and safe way. In contrast to traditional “map-localize-plan” methods, this letter explores a data-driven approach to cope with the above challenges. To accomplish this, we propose DroNet: a convolutional neural network that can safely drive a drone through the streets of a city. Designed as a fast eight-layers residual network, DroNet produces two outputs for each single input image: A steering angle to keep the drone navigating while avoiding obstacles, and a collision probability to let the UAV recognize dangerous situations and promptly react to them. The challenge is however to collect enough data in an unstructured outdoor environment such as a city. Clearly, having an expert pilot providing training trajectories is not an option given the large amount of data required and, above all, the risk that it involves for other vehicles or pedestrians moving in the streets. Therefore, we propose to train a UAV from data collected by cars and bicycles, which, already integrated into the urban environment, would not endanger other vehicles and pedestrians. Although trained on city streets from the viewpoint of urban vehicles, the navigation policy learned by DroNet is highly generalizable. Indeed, it allows a UAV to successfully fly at relative high altitudes and even in indoor environments, such as parking lots and corridors. To share our findings with the robotics community, we publicly release all our datasets, code, and trained networks.

In recent years, bacterial populations have been engineered to act as biological sensors able to improve human health by developing novel therapeutics and diagnostics. Nowadays, populations of engineered bacteria can be remotely controlled to perform some medical actions on-demand; however, it brings crucial concerns from the cybersecurity perspective. As an example, one of the first cyberbioattacks has been recently proposed to explore the feasibility of using engineered bacteria to produce a Distributed Denial-of-Service and disrupt the creation of biofilm, a natural protection of bacteria against external agents. With the goal of mitigating the impact of this cyberbioattack, this paper proposes two novel mitigation mechanisms: quorum quenching and amplification. On the one hand, quorum quenching focuses on emitting molecules to block those sent by the cyberbioattack. On the other hand, the amplification approach emits molecules to increase the percentage of those needed to create the biofilm structure. To measure the performance of both mitigation techniques in dynamic scenarios, we have implemented different configurations of the Distributed Denial-of-Service attack and evaluated the channel attenuation and the signal-to-interference-plus-noise (SINR). As a result, we have observed that both approaches reduce the impact caused by the cyberbioattack, detecting differences between them. The quorum quenching mechanism presented better results, although it did not adapt its behavior to different attack configurations, responding statically. In contrast, the amplitude mitigation technique is perfectly adapted to attack configurations with different impacts on biofilm creation.

Event cameras are bio-inspired vision sensors that output pixel-level brightness changes instead of standard intensity frames. They offer significant advantages over standard cameras, namely a very high dynamic range, no motion blur, and a latency in the order of microseconds. However, because the output is composed of a sequence of asynchronous events rather than actual intensity images, traditional vision algorithms cannot be applied, so that a paradigm shift is needed. We introduce the problem of event-based multi-view stereo (EMVS) for event cameras and propose a solution to it. Unlike traditional MVS methods, which address the problem of estimating dense 3D structure from a set of known viewpoints, EMVS estimates semi-dense 3D structure from an event camera with known trajectory. Our EMVS solution elegantly exploits two inherent properties of an event camera: (1) its ability to respond to scene edges—which naturally provide semi-dense geometric information without any pre-processing operation—and (2) the fact that it provides continuous measurements as the sensor moves. Despite its simplicity (it can be implemented in a few lines of code), our algorithm is able to produce accurate, semi-dense depth maps, without requiring any explicit data association or intensity estimation. We successfully validate our method on both synthetic and real data. Our method is computationally very efficient and runs in real-time on a CPU.

In this paper, we try to understand neural machine translation (NMT) via simplifying NMT architectures and training encoder-free NMT models. In an encoder-free model, the sums of word embeddings and positional embeddings represent the source. The decoder is a standard Transformer or recurrent neural network that directly attends to embeddings via attention mechanisms. Experimental results show (1) that the attention mechanism in encoder-free models acts as a strong feature extractor, (2) that the word embeddings in encoder-free models are competitive to those in conventional models, (3) that non-contextualized source representations lead to a big performance drop, and (4) that encoder-free models have different effects on alignment quality for German-English and Chinese-English., Accepted by RANLP 2019, camera ready version

We propose a simple unsupervised method for extracting pseudo-parallel monolingual sentence pairs from comparable corpora representative of two different text styles, such as news articles and scientific papers. Our approach does not require a seed parallel corpus, but instead relies solely on hierarchical search over pre-trained embeddings of documents and sentences. We demonstrate the effectiveness of our method through automatic and extrinsic evaluation on text simplification from the normal to the Simple Wikipedia. We show that pseudo-parallel sentences extracted with our method not only supplement existing parallel data, but can even lead to competitive performance on their own., Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2019)

When we search online for content, we are constantly exposed to rankings. For example, web search results are presented as a ranking, and online bookstores often show us lists of best-selling books. While popularity-based ranking algorithms (like Google's PageRank) have been extensively studied in previous works, we still lack a clear understanding of their potential systemic consequences. In this work, we fill this gap by introducing a new model of network growth that allows us to compare the properties of the networks generated under the influence of different ranking algorithms. We show that by correcting for the omnipresent age bias of popularity-based ranking algorithms, the resulting networks exhibit a significantly larger agreement between the nodes' inherent quality and their long-term popularity, and a less concentrated popularity distribution. To further promote popularity diversity, we introduce and validate a perturbation of the original rankings where a small number of randomly-selected nodes are promoted to the top of the ranking. Our findings move the first steps toward a model-based understanding of the long-term impact of popularity-based ranking algorithms, and could be used as an informative tool for the design of improved information filtering tools., Main text (pp. 1-27) and Supplementary Material (pp. 28-49)

Speech planning is a sophisticated process. In dialog, it regularly starts in overlap with an incoming turn by a conversation partner. We show that planning spoken responses in overlap with incoming turns is associated with higher processing load than planning in silence. In a dialogic experiment, participants took turns with a confederate describing lists of objects. The confederate’s utterances (to which participants responded) were pre‐recorded and varied in whether they ended in a verb or an object noun and whether this ending was predictable or not. We found that response planning in overlap with sentence‐final verbs evokes larger task‐evoked pupillary responses, while end predictability had no effect. This finding indicates that planning in overlap leads to higher processing load for next speakers in dialog and that next speakers do not proactively modulate the time course of their response planning based on their predictions of turn endings. The turn‐taking system exerts pressure on the language processing system by pushing speakers to plan in overlap despite the ensuing increase in processing load.

Heart activity can be monitored by means of ElectroCardioGram (ECG) measure which is widely used to detect heart diseases due to its non-invasive nature. Trained cardiologists can detect anomalies by visual inspecting recordings of the ECG signals. However, arrhythmias occur intermittently especially in early stages and therefore they can be missed in routine check recordings. We propose a hardware setup that enables the always-on monitoring of ECG signals into wearables. The system exploits a fully event-driven approach for carrying arrhythmia detection and classification employing a bio-inspired spiking neural network. The two staged Spiking Neural Network (SNN) topology comprises a recurrent network of spiking neurons whose output is classified by a cluster of Leaky integrate-and-fire (LIF) neurons that have been supervisely trained to distinguish 17 types of cardiac patterns. We introduce a method for compressing ECG signals into a stream of asynchronous digital events that are used to stimulate the recurrent SNN. Using ablative analysis, we demonstrate the impact of the recurrent SNN and we show an overall classification accuracy of 95% on the PhysioNet Arrhythmia Database provided by the Massachusetts Institute of Technology and Beth Israel Hospital (MIT/BIH). The proposed system has been implemented on an event-driven mixed-signal analog/digital neuromorphic processor. This work contributes to the realization of an energy-efficient, wearable, and accurate multi-class ECG classification system.