Hours before the first votes were cast in New Hampshire's first-in-the-nation presidential primary, Tucker Carlson assessed the state of the Democratic field, outlining the fall of former Vice President Joe Biden and rise of Sen. Bernie Sanders, I-Vt. "Like all political leaders, Democrats in Washington sought to control the outcome of this whole process to the degree that they could. And they wrote a detailed script almost a year ago, way back in the spring. They decided that this election would amount to an Obama restoration. By beating Trump, we could all return to the world before Trump," Carlson said Monday night on "Tucker Carlson Tonight."

Change is constant in cybersecurity -- continual, rapid, dynamic change. It's impossible to maintain an effective defensive posture without constantly evolving. Security measures that worked in the past will not be effective today, and today's security controls will not be effective tomorrow. Many factors contribute to this rapid pace of change. Attacks are on the rise, and they are getting more advanced, persistent and stealthy each day, with some attackers even leveraging artificial intelligence (AI) to power their campaigns.

The White House's proposed budget for fiscal year 2021 includes sizable increases in federal funding for AI and quantum computing projects. All told, the overall increase in federal R&D funding is 6% compared to fiscal year 2020, reaching $142.2 billion. Although much of the funding is aimed at R&D and infrastructure investments like $25 million to begin research for a quantum internet, the tenor of the briefing included defense tones. In espousing the need to invest in quantum information science (QIS), an official on the call said the U.S. needs to stay ahead of China and Europe, which are investing in their own quantum computing projects. U.S. CTO Michael Kratsios was clear about American values as they pertain to AI. "I think with regards to some of our adversaries and others around the world [that] utilize this technology, it's imperative that the U.S. continues to lead in technologies like AI," he said.

Amazon wants Donald Trump to submit to questioning over the tech company's losing bid for a $10bn military contract. The Pentagon awarded the cloud computing project to Microsoft in October. Amazon later sued, arguing that Trump's interference and bias against the company harmed Amazon's chances. Amazon was considered an early frontrunner for a project that Pentagon officials have described as critical to advancing the US military's technological advantage over adversaries. The project, known as Joint Enterprise Defense Infrastructure, or Jedi, will store and process vast amounts of classified data, allowing the US military to improve communications with soldiers on the battlefield and use artificial intelligence to speed up its war planning and fighting capabilities.

Contextual bandit algorithms are applied in a wide range of domains, from advertising to recommender systems, from clinical trials to education. In many of these domains, malicious agents may have incentives to attack the bandit algorithm to induce it to perform a desired behavior. For instance, an unscrupulous ad publisher may try to increase their own revenue at the expense of the advertisers; a seller may want to increase the exposure of their products, or thwart a competitor's advertising campaign. In this paper, we study several attack scenarios and show that a malicious agent can force a linear contextual bandit algorithm to pull any desired arm $T - o(T)$ times over a horizon of $T$ steps, while applying adversarial modifications to either rewards or contexts that only grow logarithmically as $O(\log T)$. We also investigate the case when a malicious agent is interested in affecting the behavior of the bandit algorithm in a single context (e.g., a specific user). We first provide sufficient conditions for the feasibility of the attack and we then propose an efficient algorithm to perform the attack. We validate our theoretical results on experiments performed on both synthetic and real-world datasets.

Local robustness ensures that a model classifies all inputs within an $\epsilon$-ball consistently, which precludes various forms of adversarial inputs. In this paper, we present a fast procedure for checking local robustness in feed-forward neural networks with piecewise linear activation functions. The key insight is that such networks partition the input space into a polyhedral complex such that the network is linear inside each polyhedral region; hence, a systematic search for decision boundaries within the regions around a given input is sufficient for assessing robustness. Crucially, we show how these regions can be analyzed using geometric projections instead of expensive constraint solving, thus admitting an efficient, highly-parallel GPU implementation at the price of incompleteness, which can be addressed by falling back on prior approaches. Empirically, we find that incompleteness is not often an issue, and that our method performs one to two orders of magnitude faster than existing robustness-certification techniques based on constraint solving.

Due to their high degree of expressiveness, neural networks have recently been used as surrogate models for mapping inputs of an engineering system to outputs of interest. Once trained, neural networks are computationally inexpensive to evaluate and remove the need for repeated evaluations of computationally expensive models in uncertainty quantification applications. However, given the highly parameterized construction of neural networks, especially deep neural networks, accurate training often requires large amounts of simulation data that may not be available in the case of computationally expensive systems. In this paper, to alleviate this issue for uncertainty propagation, we explore the application of transfer learning techniques using training data generated from both high- and low-fidelity models. We explore two strategies for coupling these two datasets during the training procedure, namely, the standard transfer learning and the bi-fidelity weighted learning. In the former approach, a neural network model mapping the inputs to the outputs of interest is trained based on the low-fidelity data. The high-fidelity data is then used to adapt the parameters of the upper layer(s) of the low-fidelity network, or train a simpler neural network to map the output of the low-fidelity network to that of the high-fidelity model. In the latter approach, the entire low-fidelity network parameters are updated using data generated via a Gaussian process model trained with a small high-fidelity dataset. The parameter updates are performed via a variant of stochastic gradient descent with learning rates given by the Gaussian process model. Using three numerical examples, we illustrate the utility of these bi-fidelity transfer learning methods where we focus on accuracy improvement achieved by transfer learning over standard training approaches.

Data-driven predictive models are increasingly used to inform decisions that hold important consequences for individuals and society. As a result, decision makers are often obliged, even legally required, to provide explanations about their decisions. In this context, it has been increasingly argued that these explanations should help individuals understand what would have to change for these decisions to be beneficial ones. However, there has been little discussion on the possibility that individuals may use the above counterfactual explanations to invest effort strategically in order to maximize their chances of receiving a beneficial decision. In this paper, our goal is to find policies and counterfactual explanations that are optimal in terms of utility in such a strategic setting. To this end, we first show that, given a pre-defined policy, the problem of finding the optimal set of counterfactual explanations is NP-hard. However, we further show that the corresponding objective is nondecreasing and satisfies submodularity. Therefore, a standard greedy algorithm offers an approximation factor of $(1-1/e)$ at solving the problem. Additionally, we also show that the problem of jointly finding both the optimal policy and set of counterfactual explanations reduces to maximizing a non-monotone submodular function. As a result, we can use a recent randomized algorithm to solve the problem, which offers an approximation factor of $1/e$. Finally, we illustrate our theoretical findings by performing experiments on synthetic and real lending data.

To train AI and ML models, large datasets are required. But this data is often inaccessible in healthcare settings, due to a variety of patient privacy laws and institutional policies. To address this obstacle and make data-sharing easier, several open-source projects have emerged. "The ability to use that large amount of data to inform our clinical decision-making is huge," said Joyce Sensmeier, vice president of informatics, technology, and innovation at HIMSS, about the open-source initiatives. The ability to leverage this already-captured data is a great next step, she said.

The Committee has today published its report and recommendations to government to ensure that high standards of conduct are upheld as technologically assisted decision making is adopted more widely across the public sector. Artificial intelligence – and in particular, machine learning – will transform the way public sector organisations make decisions and deliver public services. Adherence to high public standards will help fully realise the benefits of AI in public service delivery. By ensuring that AI is subject to appropriate safeguards and regulations, the public can have confidence that new technologies will be used in a way that upholds the Seven Principles of Public Life. We concluded that the Principles remain a valid guide for public sector practice as AI is deployed across government.