President Barack Obama hasn't had to drive himself around for years. And he may not have to after he leaves office either. Obama Monday threw his weight behind self-driving cars -- as long as they're safe. His op-ed in the Pittsburgh Post-Gazette, which will appear in Tuesday's print editions, lauds American innovation and comes just a day after Lyft CEO John Zimmer predicted the end of individual vehicle ownership and a week after Uber began an autonomous vehicle experiment in Pittsburgh. "The progress we've seen in automated vehicles over the past several years shows what our country is capable of when our engineers and entrepreneurs, our scientists and our students – backed by federal and private investment – pour their best work and brightest ideas toward a big, bold goal," Obama said.

A self-driving Ford Fusion Hybrid Uber is displayed last week at the companies' Advanced Technologies Center in Pittsburgh. On Tuesday, the Department of Transportation will announce a set of policy proposals designed to ensure the safety of autonomous and automated vehicles. A self-driving Ford Fusion Hybrid Uber is displayed last week at the companies' Advanced Technologies Center in Pittsburgh. On Tuesday, the Department of Transportation will announce a set of policy proposals designed to ensure the safety of autonomous and automated vehicles. The White House wants you to know that the era of self-driving cars is closer than you might expect.

The Obama administration said companies developing driverless cars should adopt a series of government recommendations to certify their vehicles are ready for U.S. roads, a policy aimed at front-running possible conflicting local rules and potentially reducing traffic fatalities. The U.S. Transportation Department, in guidelines to be rolled out Tuesday, urges that car makers and other developers submit to a 15-point "safety assessment" outlining how driverless cars are tested, safeguards should systems fail and how...

Any doubt that driverless cars, trucks and buses are on a fast track to join their human counterparts on the nation's highways may be knocked aside Tuesday, when the Department of Transportation releases long-awaited guidelines for the development of autonomous cars. The guidelines, which specify safety criteria and promise to eliminate red tape, will help "bring lifesaving technologies to the roads safely while providing innovators the space they need to develop new solutions," said the U.S. Department of Transportation in an early summary released Monday. Automakers and tech companies have been barreling ahead in their race to churn out driverless vehicles, prompting the government to play catch-up when it comes to how to regulate the technology. Under the guidelines, car manufacturers and researchers will be required to submit to a "15-point safety assessment" for driverless cars, including how the vehicles respond to system failure, whether they make data available for crash reconstruction and even whether their artificial intelligence software takes driving ethics into account. The Transportation Department will hasten approval or rejection of special exemptions to regulations for driverless cars, with an upper limit of six months.

SAN FRANCISCO -- Federal regulators, faced with a growing number of self-driving car tests on roads across the U.S., plan to issue a flurry of new guidelines Tuesday aimed at automakers and tech companies. The U.S. Department of Transportation will require any new tech to meet a 15-point safety assessment, consider new powers to allow administrators to limit the deployment of experimental vehicles, and will issue a model for state self-driving car policies aimed at developing a cohesive set of national regulations. Officials will solicit public comments on the topic of self-driving car regulations for the next 60 days on the Transportation Department website and plan to update self-driving car policies annually. "We're laying it out there, what we care about, and inviting the industry to show us how they meet those standards," Department of Transportation Secretary Anthony Foxx told reporters during a briefing late Monday. "Some companies haven't dealt with us, but they'll learn quickly we can go really deep on these topics. We want the public to be safe."

The prescribed viewpoint appears to be good enough to accommodate most of the existing spectral graph techniques as a consequence of just one single formalism and algorithm. Keywords and phrases: Nonparametric spectral graph analysis; Graph correlation density field (GraField); Empirical and smoothed spectral graph analysis; High-dimensional discrete data smoothing.

Darren Holly steers coils of steel through Pentaflex, a manufacturer of parts for heavy trucks, in Springfield, Ohio. Darren Holly steers coils of steel through Pentaflex, a manufacturer of parts for heavy trucks, in Springfield, Ohio. Americans who endured the brutal 2007-2009 recession and slow recovery now are seeing an economic sunrise: Wages are up, jobs are growing and more families are lifting themselves up out of poverty. And yet, dark clouds are still hanging over millions of Americans. No set of sunny statistics can help an unemployed coal miner in Kentucky pay the mortgage.

Quid, a data analytics firm, uses proprietary software to search, visualize and analyze text. Quid, a data analytics firm, uses proprietary software to search, visualize and analyze text. Predictions are for psychics -- and in this very unpredictable political season they might do a better job than the pundits. I set out to see how well it could predict which controversies around the candidates were likely to re-emerge over the course of a month. And two human pundits have agreed to compete against the machine.

Most of us have wondered what we might do if we didn't need to work – if we woke up one morning to discover we had won the lottery, say. We entertain ourselves with visions of multiple homes, trips around the world or the players we would sign after buying Arsenal. For many of us, the most tantalising aspect of such visions is the freedom it would bring: to do what one wants, when one wants and how one wants. But imagine how that vision might change if such freedom were extended to everyone. Some day, probably not in our lifetimes but perhaps not long after, machines will be able to do most of the tasks that people can.

Since blogging that I Need an AI BS-Meter a number of people have sent me pointers to a subset of AI I loosely think of as Result Explainers -- everything from pending government regulations (EU's Global Data Protection Regulations -- GDPR) to the latest in academic research (Local Interpretable Model-agnostic Explanations -- LIME). As the authors of the EU's GDPR state, widespread adoption of AI cannot occur until vendors are able to communicate results in a "concise, intelligible and easily accessible form, using clear and plain language." This got me thinking, "What should Result Explainers look like?" Should they generate trust scores, a series of Google-Maps like directions that get you from data to results, a series of diagrams? And as my colleague Patrick at Lab41 has pointed out, "Why should we trust a Result Explainer if we don't trust AI to begin with? As you might expect there isn't one right answer. That said, recent advances in recommenders, digital assistants, user interface design and initiatives like DARPA's recently announced Explainable Artificial Intelligence (XAI) grand challenge suggest we may be on the brink of a few breakthroughs. Again, as the authors of the EU's General Data Protection Regulations note, while the resulting classifiers, models, predictors, etc. can be very powerful they also frequently confound explanation -- e.g., the output of SVMs and Gaussian processes can be difficult to render, ensemble methods hide information as a result of aggregation and averaging, neural nets create high data dimensionality, and so on. End users care a lot more about results than they do about models. Unfortunately assessing result quality takes us right back to the models, as nonparametric models are only as good as the data used to train them (along with the type of model structure and associated parameters that were selected). But these models frequently hide information. Part of the magic of AI is that it finds stuff based on features that previously may not have been well understood. Unfortunately, the features models train on are frequently unclear. Assigning labels to pre-trained models can help mitigate some of this ambiguity -- e.g., "This model was trained with over 100,000 high-res color images of cats." These labels may be misleading though, as the model may contain feature biases that are not well understood -- e.g., "the training data is dominated by images of "well-fed, indoor cats from Japan."