In 2015, car-and-rocket man Elon Musk joined with influential startup backer Sam Altman to put artificial intelligence on a new, more open course. They cofounded a research institute called OpenAI to make new AI discoveries and give them away for the common good. Now, the institute's researchers are sufficiently worried by something they built that they won't release it to the public. The AI system that gave its creators pause was designed to learn the patterns of language. It does that very well--scoring better on some reading-comprehension tests than any other automated system.

When it comes to exploring far off planets, robots need to be able to tackle all sorts of challenges, so NASA has been working on a series of climbing robots to take on different tasks in inhospitable environments. First up is LEMUR (Limbed Excursion Mechanical Utility Robot) which can climb rock walls using hundreds of fishhooks in its fingers. It uses A.I. to navigate around obstacles that it cannot climb, and is one of NASA's first generation of climbing robots. It was developed to perform repair tasks aboard the International Space Station, and below you can see it in a field test in Death Valley, California. Then there's the somewhat terrifying-looking Ice Worm, which was adapted from one of LEMUR's limbs.

The goal of this paper is to develop a machine learning based approach that utilizes phase space alone to separate the Gaia DR2 stars into two categories: those accreted onto the Milky Way from in situ stars that were born within the Galaxy. Traditional selection methods that have been used to identify accreted stars typically rely on full 3D velocity and/or metallicity information, which significantly reduces the number of classifiable stars. The approach advocated here is applicable to a much larger fraction of Gaia DR2. A method known as transfer learning is shown to be effective through extensive testing on a set of mock Gaia catalogs that are based on the FIRE cosmological zoom-in hydrodynamic simulations of Milky Way-mass galaxies. The machine is first trained on simulated data using only 5D kinematics as inputs, and is then further trained on a cross-matched Gaia/RAVE data set, which improves sensitivity to properties of the real Milky Way. The result is a catalog that identifies ~650,000 accreted stars within Gaia DR2. This catalog can yield empirical insights into the merger history of the Milky Way, and could be used to infer properties of the dark matter distribution.

Each year, around 6 million car accidents occur in the U.S. on average. Road safety features (e.g., concrete barriers, metal crash barriers, rumble strips) play an important role in preventing or mitigating vehicle crashes. Accurate maps of road safety features is an important component of safety management systems for federal or state transportation agencies, helping traffic engineers identify locations to invest on safety infrastructure. In current practice, mapping road safety features is largely done manually (e.g., observations on the road or visual interpretation of streetview imagery), which is both expensive and time consuming. In this paper, we propose a deep learning approach to automatically map road safety features from streetview imagery. Unlike existing Convolutional Neural Networks (CNNs) that classify each image individually, we propose to further add Recurrent Neural Network (Long Short Term Memory) to capture geographic context of images (spatial autocorrelation effect along linear road network paths). Evaluations on real world streetview imagery show that our proposed model outperforms several baseline methods.

Improving the accuracy and robustness of deep neural nets (DNNs) and adapting them to small training data are primary tasks in deep learning research. In this paper, we replace the output activation function of DNNs, typically the data-agnostic softmax function, with a graph Laplacian-based high dimensional interpolating function which, in the continuum limit, converges to the solution of a Laplace-Beltrami equation on a high dimensional manifold. Furthermore, we propose end-to-end training and testing algorithms for this new architecture. The proposed DNN with graph interpolating activation integrates the advantages of both deep learning and manifold learning. Compared to the conventional DNNs with the softmax function as output activation, the new framework demonstrates the following major advantages: First, it is better applicable to data-efficient learning in which we train high capacity DNNs without using a large number of training data. Second, it remarkably improves both natural accuracy on the clean images and robust accuracy on the adversarial images crafted by both white-box and black-box adversarial attacks. Third, it is a natural choice for semi-supervised learning.

We offer a graphical interpretation of unfairness in a dataset as the presence of an unfair causal path in the causal Bayesian network representing the data-generation mechanism. We use this viewpoint to revisit the recent debate surrounding the COMPAS pretrial risk assessment tool and, more generally, to point out that fairness evaluation on a model requires careful considerations on the patterns of unfairness underlying the training data. We show that causal Bayesian networks provide us with a powerful tool to measure unfairness in a dataset and to design fair models in complex unfairness scenarios.

All major powers are currently focusing on the development of autonomous weapon systems. Artificial Intelligence (AI) is a technological breakthrough that would render the world unrecognisable as we know it today. Though the idea that machines would possess human-like cognitive capabilities might have sounded like science fiction in the past century, it has now transitioned to reality. Since its inception, Artificial Intelligence has drawn attention from a diverse number of fields and the concerned "researches and developments" are moving at a staggering pace. AI-powered smart assistants to advanced training simulations and even self-driving vehicles are a reality now.

Roundup Hello, here's a few announcements from the world of machine learning beyond what we've already covered this week. AlphaStar is coming out to play: AlphaStar, the StarCraft II-playing bot built by DeepMind researchers, will be facing human players in a series of 1v1 games online. StarCraft II players can enter the open competition league set up by Blizzard Entertainment, the creators of the popular battle strategy game, and opt-in to play against AlphaStar. But nobody will know if they're facing the bot, however, because it'll be entering the matches anonymously. Characters in the StarCraft II are from three species: Terran, Zerg or Protoss.

We already knew humans could make biased AIs -- but the United Nations says the reverse is true as well. Millions of people talk to AI voice assistants, such as Apple's Siri and Amazon's Alexa. When those assistants talk back, they do so in female-sounding voices, and a new UN report argues that those voices and the words they're programmed to say amplify gender biases and encourage users to be sexist -- but it's not too late to change course. The report is the work of the United Nations Educational, Scientific, and Cultural Organization (UNESCO), and its title -- "I'd blush if I could" -- is the response Siri was programmed in 2011 to give if a user called her a "bitch." According to UNESCO, that programming exemplifies the problems with today's AI assistants.

One of China's fastest growing robotic startups is ramping up its international expansion, even as Trump's tariffs force the company to dial back on its US ambitions. Geek Plus Robotics automates supply chains by replacing warehouse workers with bots. The Beijing-based company recently raised more than 100 million dollars and expects to more than double that at its next funding round. China correspondent Tom Mackenzie spoke exclusively to the company's founder and CEO in Beijing about the firm' plans.