Safety Controller Synthesis for Collaborative Robots Mario Gleirscher, Radu Calinescu Assuring Autonomy International Programme, University of Y ork, Y ork, UK Department of Computer Science, University of Y ork, Y ork, UK mario.gleirscher,radu.calinescu@york.ac.uk Abstract --In human-robot collaboration (HRC), software-based automatic safety controllers (ASCs) are used in various forms (e.g. Complex robotic tasks and increasingly close human-robot interaction pose new challenges to ASC developers and certification authorities. Key among these challenges is the need to assure the correctness of ASCs under reasonably weak assumptions. T o address this need, we introduce and evaluate a tool-supported ASC synthesis method for HRC in manufacturing. Our ASC synthesis is: (i) informed by the manufacturing process, risk analysis, and regulations; (ii) formally verified against correctness criteria; and (iii) selected from a design space of feasible controllers according to a set of optimality criteria. The synthesised ASC can detect the occurrence of hazards, move the process into a safe state, and, in certain circumstances, return the process to an operational state from which it can resume its original task. I NTRODUCTION An effective collaboration between industrial robot systems (IRSs) and humans [1], [2] can leverage their complementary skills, but is difficult to achieve because of uncontrolled hazards and unexploited sensing, tracking, and safety measures [3]. Such hazards have been studied since the 1970s, resulting in elaborate risk taxonomies based on workspaces, tasks, and human body regions [2], [4]-[10]. The majority are impact hazards (e.g. Addressing these hazards involves the examination of each mode of operation (e.g. I, a variety of safety measures [3] can prevent or mitigate hazards and accidents by reducing the probability of their occurrence and the severity of their consequences . There are functional measures using electronic equipment (e.g.

Bayesian optimization (BO) is an efficient and flexible global optimization framework that is applicable to a very wide range of engineering applications. To leverage the capability of the classical BO, many extensions, including multi-objective, multi-fidelity, parallelization, latent-variable model, have been proposed to improve the limitation of the classical BO framework. In this work, we propose a novel multi-objective (MO) extension, called srMO-BO-3GP, to solve the MO optimization problems in a sequential setting. Three different Gaussian processes (GPs) are stacked together, where each of the GP is assigned with a different task: the first GP is used to approximate the single-objective function, the second GP is used to learn the unknown constraints, and the third GP is used to learn the uncertain Pareto frontier. At each iteration, a MO augmented Tchebycheff function converting MO to single-objective is adopted and extended with a regularized ridge term, where the regularization is introduced to smoothen the single-objective function. Finally, we couple the third GP along with the classical BO framework to promote the richness and diversity of the Pareto frontier by the exploitation and exploration acquisition function. The proposed framework is demonstrated using several numerical benchmark functions, as well as a thermomechanical finite element model for flip-chip package design optimization.

Voting rules may fail to implement the will of the society when only some voters actively participate, and/or in the presence of sybil (fake or duplicate) voters. Here we aim to address social choice in the presence of sybils and voter abstention. To do so we assume the status-quo (Reality) as an ever-present distinguished alternative, and study Reality Enforcing voting rules, which add virtual votes in support of the status-quo. We measure the tradeoff between safety and liveness (the ability of active honest voters to maintain/change the status-quo, respectively) in a variety of domains, and show that the Reality Enforcing voting rule is optimal in this respect.

Developing secure machine learning models from adversarial examples is challenging as various methods are continually being developed to generate adversarial attacks. In this work, we propose an evolutionary approach to automatically determine Image Processing Techniques Sequence (IPTS) for detecting malicious inputs. Accordingly, we first used a diverse set of attack methods including adaptive attack methods (on our defense) to generate adversarial samples from the clean dataset. A detection framework based on a genetic algorithm (GA) is developed to find the optimal IPTS, where the optimality is estimated by different fitness measures such as Euclidean distance, entropy loss, average histogram, local binary pattern and loss functions. The "image difference" between the original and processed images is used to extract the features, which are then fed to a classification scheme in order to determine whether the input sample is adversarial or clean. This paper described our methodology and performed experiments using multiple data-sets tested with several adversarial attacks. For each attack-type and dataset, it generates unique IPTS. A set of IPTS selected dynamically in testing time which works as a filter for the adversarial attack. Our empirical experiments exhibited promising results indicating the approach can efficiently be used as processing for any AI model.

Deep neural networks are capable of training fast and generalizing well within many domains. Despite their promising performance, deep networks have shown sensitivities to perturbations of their inputs (e.g., adversarial examples) and their learned feature representations are often difficult to interpret, raising concerns about their true capability and trustworthiness. Recent work in adversarial training, a form of robust optimization in which the model is optimized against adversarial examples, demonstrates the ability to improve performance sensitivities to perturbations and yield feature representations that are more interpretable. Adversarial training, however, comes with an increased computational cost over that of standard (i.e., nonrobust) training, rendering it impractical for use in large-scale problems. Recent work suggests that a fast approximation to adversarial training shows promise for reducing training time and maintaining robustness in the presence of perturbations bounded by the infinity norm. In this work, we demonstrate that this approach extends to the Euclidean norm and preserves the human-aligned feature representations that are common for robust models. Additionally, we show that using a distributed training scheme can further reduce the time to train robust deep networks. Fast adversarial training is a promising approach that will provide increased security and explainability in machine learning applications for which robust optimization was previously thought to be impractical.

By all appearances, May Mobility was a scrappy success story. The autonomous transportation startup made its debut at Y Combinator's demo day in 2017, with a team that had been working on driverless tech since the third U.S. Defense Advanced Research Projects Agency (DARPA) Grand Challenge in 2017. Within the span of a few years, May had a roster of paying customers in Michigan, Ohio, and Rhode Island as it raised tens of millions in venture capital from investors including Toyota and BMW. But on the inside looking out, it was a different story. May engineers struggled to maintain and upgrade the company's vehicle platform, at one point spending months attempting to install an air conditioning system in the depths of summer.

Organizations are deluged with billions of security events every day, far too many for human analysts to cope with. But security analysts have a powerful ally in their corner: machine learning is tipping the advantage toward defenders. Machine learning (ML) is changing the approach of organizations to threat detection and how they adapt and adopt cybersecurity processes. The idea is not just to identify and prevent threats, but to mitigate them as well. ML has the power to comprehend threats in real time, to understand the infrastructure of a company and its network design and attack vectors, and to protect and defend it with human talent and machine power. The algorithm--the machine--is capable of the unthinkable when it comes to data mining, data crunching, and data correlation, since it does what it does best tirelessly, without complaint, and without having a bad day.

Combat drone engagement is becoming a usual practice. The unmanned aerial vehicles (UAV) fight air defense and often win. Swarm tactic of drone engagement has emerged. It brings drone control to the foreground, as the usual operator-drone scheme often fails. Other technologies are necessary, the online Army Standard publication said.

Cyberattacks on the likes of several tech giants have brought to the fore the challenge of bridging the skills gap in the cybersecurity space in India. And, artificial intelligence being the latest buzzword of the tech industry, is being touted as one of the key solutions to the cybersecurity skills gap. According to a report, it is estimated that there will be 3.5 million unfilled cybersecurity jobs globally by the year 2021. And therefore, companies are struggling to find adequate qualified people to assist in creating an intelligent cybersecurity framework. The challenge has become apparent in the last five to ten years with a sharp increase in cyberattacks, all the way from ransomware to zero-day malware to now sneaky crypto-mining attacks.

China announced in 2017 its ambition to become the world leader in artificial intelligence (AI) by 2030. While the US still leads in absolute terms, China appears to be making more rapid progress than either the US or the EU, and central and local government spending on AI in China is estimated to be in the tens of billions of dollars. The move has led – at least in the West – to warnings of a global AI arms race and concerns about the growing reach of China's authoritarian surveillance state. But treating China as a "villain" in this way is both overly simplistic and potentially costly. While there are undoubtedly aspects of the Chinese government's approach to AI that are highly concerning and rightly should be condemned, it's important that this does not cloud all analysis of China's AI innovation.