The problem of adversarial robustness has been studied extensively for neural networks. However, for boosted decision trees and decision stumps there are almost no results, even though they are widely used in practice (e.g. We show in this paper that for boosted decision stumps the \textit{exact} min-max robust loss and test error for an $l_\infty$-attack can be computed in $O(T\log T)$ time per input, where $T$ is the number of decision stumps and the optimal update step of the ensemble can be done in $O(n 2\,T\log T)$, where $n$ is the number of data points. For boosted trees we show how to efficiently calculate and optimize an upper bound on the robust loss, which leads to state-of-the-art robust test error for boosted trees on MNIST (12.5\% for $\epsilon_\infty 0.3$), FMNIST (23.2\% for $\epsilon_\infty 0.1$), and CIFAR-10 (74.7\% for $\epsilon_\infty 8/255$). Moreover, the robust test error rates we achieve are competitive to the ones of provably robust convolutional networks. Papers published at the Neural Information Processing Systems Conference.

While over half of organizations use artificial intelligence or machine learning in their security stack, nearly 60 percent are still more confident in cyberthreat findings verified by humans over AI, according to WhiteHat Security. The survey responses of 102 industry professionals at RSA Conference 2020 reflect the need for security organizations to incorporate both AI- and human-centric offerings, especially in the application security space. Three-quarters of respondents use an application security tool, and more than 40 percent of those application security solutions use both AI-based and human-based verification. AI and machine learning have provided several advantages for cybersecurity professionals overall the past several years, especially in the face of the technology talent gap, which has left 45 percent of respondents' companies lacking a sufficiently staffed cybersecurity team. More than 70 percent of respondents agree that AI-based tools made their cybersecurity teams more efficient by eliminating over 55 percent of mundane tasks.

We propose functional adversarial attacks, a novel class of threat models for crafting adversarial examples to fool machine learning models. Unlike a standard lp-ball threat model, a functional adversarial threat model allows only a single function to be used to perturb input features to produce an adversarial example. For example, a functional adversarial attack applied on colors of an image can change all red pixels simultaneously to light red. Such global uniform changes in images can be less perceptible than perturbing pixels of the image individually. For simplicity, we refer to functional adversarial attacks on image colors as ReColorAdv, which is the main focus of our experiments.

To deflect adversarial attacks, a range of "certified" classifiers have been proposed. In addition to labeling an image, certified classifiers produce (when possible) a certificate guaranteeing that the input image is not an $\ell_p$-bounded adversarial example. We present a new attack that exploits not only the labelling function of a classifier, but also the certificate generator. The proposed method applies large perturbations that place images far from a class boundary while maintaining the imperceptibility property of adversarial examples. The proposed "Shadow Attack" causes certifiably robust networks to mislabel an image and simultaneously produce a "spoofed" certificate of robustness.

Event extraction (EE) is one of the core information extraction tasks, whose purpose is to automatically identify and extract information about incidents and their actors from texts. This may be beneficial to several domains such as knowledge bases, question answering, information retrieval and summarization tasks, to name a few. The problem of extracting event information from texts is longstanding and usually relies on elaborately designed lexical and syntactic features, which, however, take a large amount of human effort and lack generalization. More recently, deep neural network approaches have been adopted as a means to learn underlying features automatically. However, existing networks do not make full use of syntactic features, which play a fundamental role in capturing very long-range dependencies. Also, most approaches extract each argument of an event separately without considering associations between arguments which ultimately leads to low efficiency, especially in sentences with multiple events. To address the two above-referred problems, we propose a novel joint event extraction framework that aims to extract multiple event triggers and arguments simultaneously by introducing shortest dependency path (SDP) in the dependency graph. We do this by eliminating irrelevant words in the sentence, thus capturing long-range dependencies. Also, an attention-based graph convolutional network is proposed, to carry syntactically related information along the shortest paths between argument candidates that captures and aggregates the latent associations between arguments; a problem that has been overlooked by most of the literature. Our results show a substantial improvement over state-of-the-art methods.

Artificial intelligence often raises concerns about privacy, bias and trickery in areas such as facial recognition and deep fake videos. But amidst the outbreak of the novel coronavirus, some technology companies and scientists are looking to AI for a positive impact instead. "AI and high tech in general have gotten something of a bad rap recently, but this crisis shows how AI can potentially do a world of good," said Oren Etzioni, CEO of Seattle's Allen Institute for Artificial Intelligence (AI2) and a University of Washington computer science professor. Etzioni was speaking on a call Monday organized by the White House Office of Science and Technology Policy, as part of an announcement of a project called the COVID-19 Open Research Dataset, aka CORD-19. The initiative, building on AI2's Semantic Scholar project, uses natural language processing to analyze scientific papers about coronavirus, including the novel coronavirus that causes COVID-19.

The White House has urged AI experts to analyze a dataset of 29,000 scholarly articles about coronavirus that could offer insights into how to manage the pandemic. These questions have been published on Kaggle, a machine learning community owned by Google. The entire COVID-19 Open Research Dataset (CORD-19) has been made available on SemanticScholar, a free, nonprofit, academic search engine. The collection will be updated whenever new research is published in archival services and peer-reviewed publications. "Decisive action from America's science and technology enterprise is critical to prevent, detect, treat, and develop solutions to COVID-19," said Michael Kratsios, the USA's Chief Technology Officer.

In a very important way, adaptive analytics processes help companies cope with constantly changing external forces, such as new regulatory compliance requirements, market dynamics and competitive pressures. Analytics are also critical for monitoring internal functions and change programs to provide context, transparency and agility for identifying issues and responding to potential problems before they become emergencies. They enable organizations to become adaptable and react and pro act to constantly changing market forces such as Brexit, US/China Trade, Mideast and others.

Like'digital transformation', innovation is an over-used and under-examined term. This applies within business generally, but more especially within the public sector, where there are limits to the amount of disruption and risk that it is considered acceptable to carry within the public domain. Further, a range of questions arises when government'innovates'. These include building the culture and incentives for innovation; understanding what innovation in the digital era is actually about (clue: it's not simply about having a new idea); handling the public-private sector relationship differently; scaling innovations; and handling the politics that inevitably surround changes of almost any kind to public services. The opportunity to chair the second Public Sector Innovation Conference on 25 February was a great opportunity to reflect on these, and many of the other tensions and opportunities that surround ongoing modernisation of public services, and benefit from a really high-quality speaker lineup.

Police in China are using Robocop-style helmets embedded with AI to spot someone with a fever from 16 feet away. A restaurant in L.A. has been checking people's temperatures at the door with an infrared noncontact thermometer. And a hotel near Texas Medical Center in Houston just deployed germ-zapping robots to sanitize guest rooms and common areas. In the war against the spread of the coronavirus, tech gadgets and telemedicine services are getting fast-tracked to the front lines. It's been a whirlwind few months for Dr. Samir Qamar.