What goals should a multi-goal reinforcement learning agent pursue during training in long-horizon tasks? When the desired (test time) goal distribution is too distant to offer a useful learning signal, we argue that the agent should not pursue unobtainable goals. Instead, it should set its own intrinsic goals that maximize the entropy of the historical achieved goal distribution. We propose to optimize this objective by having the agent pursue past achieved goals in sparsely explored areas of the goal space, which focuses exploration on the frontier of the achievable goal set. We show that our strategy achieves an order of magnitude better sample efficiency than the prior state of the art on long-horizon multi-goal tasks including maze navigation and block stacking.

In the age of big data, companies and governments are increasingly using algorithms to inform hiring decisions, employee management, policing, credit scoring, insurance pricing, and many more aspects of our lives. AI systems can help us make evidence-driven, efficient decisions, but can also confront us with unjustified, discriminatory decisions wrongly assumed to be accurate because they are made automatically and quantitatively. It is becoming evident that these technological developments are consequential to people's fundamental human rights. Despite increasing attention to these urgent challenges in recent years, technical solutions to these complex socio-ethical problems are often developed without empirical study of societal context and the critical input of societal stakeholders who are impacted by the technology. On the other hand, calls for more ethically- and socially-aware AI often fail to provide answers for how to proceed beyond stressing the importance of transparency, explainability, and fairness. Bridging these socio-technical gaps and the deep divide between abstract value language and design requirements is essential to facilitate nuanced, context-dependent design choices that will support moral and social values. In this paper, we bridge this divide through the framework of Design for Values, drawing on methodologies of Value Sensitive Design and Participatory Design to present a roadmap for proactively engaging societal stakeholders to translate fundamental human rights into context-dependent design requirements through a structured, inclusive, and transparent process.

Data augmentation techniques have become standard practice in deep learning, as it has been shown to greatly improve the generalisation abilities of models. These techniques rely on different ideas such as invariance-preserving transformations (e.g, expert-defined augmentation), statistical heuristics (e.g, Mixup), and learning the data distribution (e.g, GANs). However, in the adversarial settings it remains unclear under what conditions such data augmentation methods reduce or even worsen the misclassification risk. In this paper, we therefore analyse the effect of different data augmentation techniques on the adversarial risk by three measures: (a) the well-known risk under adversarial attacks, (b) a new measure of prediction-change stress based on the Laplacian operator, and (c) the influence of training examples on prediction. The results of our empirical analysis disprove the hypothesis that an improvement in the classification performance induced by a data augmentation is always accompanied by an improvement in the risk under adversarial attack. Further, our results reveal that the augmented data has more influence than the non-augmented data, on the resulting models. Taken together, our results suggest that general-purpose data augmentations that do not take into the account the characteristics of the data and the task, must be applied with care.

"The World Economic Forum is spearheading a multistakeholder, evidence-based policy project in partnership with the Government of New Zealand. The project aims at co-designing actionable governance frameworks for AI regulation. It is structured around three focus areas: 1) obtaining of a social licence for the use of AI through an inclusive national conversation; 2) the development of in-house understanding of AI to produce well-informed policies; and 3) the effective mitigation of risks associated with AI systems to maximize their benefits." Reimagining Regulation for the Age of AI – aim to create enabling frameworks that support the operationalization of the ethical use of artificial intelligence. This is all about how do you create transparency and build trust between humans, robots and the companies controling them?

Julia is a comparably new language that aimed to have the performance of C and simplicity of Python. Having the ability to perform data analysis without much trouble while shipping the code with competitive performance, Julia is expected to be a powerful tool in FinTech businesses. But I think it also have some great potentials regarding to the current trends in Cybersecurity. In this article, I will explain why Julia can also be a great tool for the future of Cybersecurity. Meanwhile, I will share how I wrote a Julia script on my Mac computer to crack a cipher text encrypted with Caesar Code Shift and Columnar Transposition.

The World Economic Forum's Frameworks for Reimagining Regulation at the Age of AI seek to address the need for upgrading our existing regulatory environment to ensure the trustworthy design and deployment of AI. These frameworks provide Governments with innovative approaches and tools for regulating AI that can be scaled.

You are receiving this notification due to your interest in advanced physical security technology solutions.

A number of articles have been published recently predicting that the largest cyberattack in history is destined to happen soon, one of the main underlying factors behind this assertion is the overnight explosion of the enterprise attack surface and large increase in noted hacks that we've witnessed during the COVID-19 Pandemic. As an example, a recent Forbes article by Stephen McBride claims that, "The Largest Cyberattack In History Could Happen Within Six Months." Although it is entirely possible and even potentially probable that the largest cyber security breach in history is right around the corner, it is also entirely avoidable. Solutions to protect networks in the changing enterprise cyber landscape we are witnessing due to events like COVID-19 do exist, but they are not your typical legacy tools utilizing "AI" that is based on human labeling or Supervised Learning Algorithms which most companies are relying on for their cybersecurity now. According to McBride, switching to remote work, with employees now sharing computers with their loved ones, who are using them for everything from zoom get-togethers to school work, on such a massive scale has caused the attack surface to grow by an astounding 500 percent, virtually overnight.

Back when we started WIRED magazine, it was all digital, all the time. In Silicon Valley, bodies were treated like the somewhat inconvenient and sometimes embarrassing things that needed to be fueled and occasionally rested so that they could support big heads that housed big ideas about the future. Human biology wasn't exactly on our radar, except in science fiction, where pandemics always seemed du jour. Jane Metcalfe is the founder, with Louis Rossetto, of WIRED. After a stint as the president of TCHO Chocolate, she created NEO.LIFE to track the ways we are changing as we bring an engineering mindset to our own biology.

A wave of protests over law enforcement abuses has highlighted concerns over artificial intelligence programs like facial recognition that critics say …