This paper presents a new approach to a robust Gaussian process (GP) regression. Most existing approaches replace an outlier-prone Gaussian likelihood with a non-Gaussian likelihood induced from a heavy tail distribution, such as the Laplace distribution and Student-t distribution. However, the use of a non-Gaussian likelihood would incur the need for a computationally expensive Bayesian approximate computation in the posterior inferences. The proposed approach models an outlier as a noisy and biased observation of an unknown regression function, and accordingly, the likelihood contains bias terms to explain the degree of deviations from the regression function. We entail how the biases can be estimated accurately with other hyperparameters by a regularized maximum likelihood estimation. Conditioned on the bias estimates, the robust GP regression can be reduced to a standard GP regression problem with analytical forms of the predictive mean and variance estimates. Therefore, the proposed approach is simple and very computationally attractive. It also gives a very robust and accurate GP estimate for many tested scenarios. For the numerical evaluation, we perform a comprehensive simulation study to evaluate the proposed approach with the comparison to the existing robust GP approaches under various simulated scenarios of different outlier proportions and different noise levels. The approach is applied to data from two measurement systems, where the predictors are based on robust environmental parameter measurements and the response variables utilize more complex chemical sensing methods that contain a certain percentage of outliers. The utility of the measurement systems and value of the environmental data are improved through the computationally efficient GP regression and bias model.

To support human decision making with machine learning models, we often need to elucidate patterns embedded in the models that are unsalient, unknown, or counterintuitive to humans. While existing approaches focus on explaining machine predictions with real-time assistance, we explore model-driven tutorials to help humans understand these patterns in a training phase. We consider both tutorials with guidelines from scientific papers, analogous to current practices of science communication, and automatically selected examples from training data with explanations. We use deceptive review detection as a testbed and conduct large-scale, randomized human-subject experiments to examine the effectiveness of such tutorials. We find that tutorials indeed improve human performance, with and without real-time assistance. In particular, although deep learning provides superior predictive performance than simple models, tutorials and explanations from simple models are more useful to humans. Our work suggests future directions for human-centered tutorials and explanations towards a synergy between humans and AI.

We present a unifying framework encompassing many social choice settings. Viewing each social choice setting as voting in a suitable metric space, we consider a general model of social choice over metric spaces, in which---similarly to the spatial model of elections---each voter specifies an ideal element of the metric space. The ideal element functions as a vote, where each voter prefers elements that are closer to her ideal element. But it also functions as a proposal, thus making all participants equal not only as voters but also as proposers. We consider Condorcet aggregation and a continuum of solution concepts, ranging from minimizing the sum of distances to minimizing the maximum distance. We study applications of the abstract model to various social choice settings, including single-winner elections, committee elections, participatory budgeting, and participatory legislation. For each setting, we compare each solution concept to known voting rules and study various properties of the resulting voting rules. Our framework provides expressive aggregation for a broad range of social choice settings while remaining simple for voters, and may enable a unified and integrated implementation for all these settings, as well as unified extensions such as sybil-resiliency, proxy voting, and deliberative decision making.

Climate-driven change in the Gulf of Maine is raising new threats that "red tides" will become more frequent and prolonged. But at the same time, powerful new data collection techniques and artificial intelligence are providing more precise ways to predict where and when toxic algae will bloom. One of those new machine learning prediction models has been developed by a former intern at Bigelow Labs in East Boothbay. In a busy shed on a Portland wharf, workers for Bangs Island Mussels sort and clean shellfish hauled from Casco Bay that morning. Wholesaler George Parr has come to pay a visit.

Now a team of scientists has repurposed living cells -- scraped from frog embryos -- and assembled them into entirely new life-forms. These millimeter-wide "xenobots" can move toward a target, perhaps pick up a payload (like a medicine that needs to be carried to a specific place inside a patient) -- and heal themselves after being cut. "These are novel living machines," says Joshua Bongard, a computer scientist and robotics expert at the University of Vermont who co-led the new research. "They're neither a traditional robot nor a known species of animal. The new creatures were designed on a supercomputer at UVM -- and then assembled and tested by biologists at Tufts University. "We can imagine many useful applications of these living robots that other machines can't do," says co-leader Michael Levin who directs the Center for Regenerative and Developmental Biology at Tufts, "like searching out nasty compounds or radioactive contamination, gathering microplastic in the oceans, ...

Have you seen Barack Obama call Donald Trump a "complete dipshit", or Mark Zuckerberg brag about having "total control of billions of people's stolen data", or witnessed Jon Snow's moving apology for the dismal ending to Game of Thrones? Answer yes and you've seen a deepfake. The 21st century's answer to Photoshopping, deepfakes use a form of artificial intelligence called deep learning to make images of fake events, hence the name deepfake. Want to put new words in a politician's mouth, star in your favourite movie, or dance like a pro? Then it's time to make a deepfake.

DUBAI, UNITED ARAB EMIRATES – A flurry of diplomatic visits and meetings crisscrossing the Persian Gulf have driven urgent efforts in recent days to defuse the possibility of all-out war after the U.S. killed Iran's top military commander. Global leaders and top diplomats are repeating the mantra of "de-escalation" and "dialog," yet none has publicly laid out a path to achieving either. The United States and Iran have said they do not want war, but fears have grown that the crisis could spin out of Tehran's or Washington's control. Tensions have careened from one crisis to another since President Donald Trump withdrew the U.S. from Iran's nuclear deal with world powers. The U.S. drone strike that killed Revolutionary Guard Gen. Qassem Soleimani and a senior Iraqi militia leader in Baghdad on Jan. 3 was seen as a major provocation.

WASHINGTON – President Donald Trump on Monday morning defended his decision to kill Iranian commander Qassem Soleimani, contending Soleimani posed an impending threat to the United States but also saying that was not important, given the military leader's history. "The Fake News Media and their Democrat Partners are working hard to determine whether or not the future attack by terrorist Soleimani was'imminent' or not, & was my team in agreement." "The answer to both is a strong YES., but it doesn't really matter because of his horrible past!" Since confirming that Iranian military leader Qassem Soleimani had been killed by a U.S. airstrike in Baghdad, administration officials have claimed they acted because of an imminent risk of attacks on American diplomats and service members in Iraq and throughout the region. Democrats and a few Republicans in Congress have questioned the justification of the attacks and said they have not been given adequate, detailed briefings. Last week Trump posited in an interview that Iran had been poised to attack four American embassies before Soleimani was killed in a U.S. drone strike on Jan. 3. But on Sunday U.S. Defense Secretary Mark Esper said he did not see specific evidence that Iran was planning an attack.

Artificial intelligence (AI) represents a powerful but double-edged sword as nations confront global warming, poverty and issues of peace and justice. An international team of scientists this week released a first-ever study of how AI can help--as well as hinder--sustainable development worldwide. Published today in Nature Communications, the analysis focuses on how AI impacts the 17 goals for sustainable development adopted by the United Nations in 2015. The study was co-authored by a diverse group of researchers led by Ricardo Vinuesa and Francesco Fuso Nerini, assistant professors at KTH Royal Institute of Technology. They were joined by Max Tegmark, professor at Massachusetts Institute of Technology (MIT) and author of the bestselling book Life 3.0, as well as Virginia Dignum, professor of AI Ethics at Umeå University, among other authors.

Aidoc, the leading provider of AI solutions for radiologists, today announced that the US Food and Drug Administration (FDA) has cleared its AI solution for flagging Large-Vessel Occlusion (LVO) in head CTA scans, marking Aidoc's fourth FDA-cleared AI package. Combined with Aidoc's previously-cleared AI module for flagging and prioritizing intracranial hemorrhage, together they provide a comprehensive AI package for the identification and triage of both ischemic and hemorrhagic stroke in CTs, speeding time to treatment when every minute counts. "Stroke is the ultimate time-critical condition," said Dr. Marcel Maya, Co-chair Department of Imaging, Cedars-Sinai Medical Center. "The faster we can identify, diagnose and treat it, the better the outcome for patients. Aidoc's comprehensive stroke package flags both large vessel occlusion and hemorrhages inside our existing workflows, ensuring we can diagnose stroke faster and decide on the best course of treatment. We're already seeing how this has a positive impact on department efficiency and patient length of stay."