A group of Google employees and shareholders are calling for the search giant to link executive compensation with its ability to meet diversity goals. They will present a proposal for that policy at Google parent company Alphabet's annual shareholder meeting on Wednesday. All this despite the fact that Alphabet has already voiced its opposition to the proposal, saying it's not in the best of the company or its shareholders. The move caps off several months of controversy at Google, largely surrounding the firm's decision to take part in a partnership with the Pentagon, where its TensorFlow software would be used in its AI drone program, called Project Maven. Ultimately, Google announced it wouldn't renew its contract with the Pentagon when it expires in 2019, after thousands of employees wrote an open letter urging it to do so, while others resigned over the firm's involvement.

Asynchronous momentum stochastic gradient descent algorithms (Async-MSGD) have been widely used in distributed machine learning, e.g., training large collaborative filtering systems and deep neural networks. Due to current technical limit, however, establishing convergence properties of Async-MSGD for these highly complicated nonoconvex problems is generally infeasible. Therefore, we propose to analyze the algorithm through a simpler but nontrivial nonconvex problem - streaming PCA. This allows us to make progress toward understanding Aync-MSGD and gaining new insights for more general problems. Specifically, by exploiting the diffusion approximation of stochastic optimization, we establish the asymptotic rate of convergence of Async-MSGD for streaming PCA. Our results indicate a fundamental tradeoff between asynchrony and momentum: To ensure convergence and acceleration through asynchrony, we have to reduce the momentum (compared with Sync-MSGD). To the best of our knowledge, this is the first theoretical attempt on understanding Async-MSGD for distributed nonconvex stochastic optimization. Numerical experiments on both streaming PCA and training deep neural networks are provided to support our findings for Async-MSGD.

In recent years, neural networks have demonstrated outstanding effectiveness in a large amount of applications. However, recent works have shown that neural networks are susceptible to adversarial examples, indicating possible flaws intrinsic to the network structures. To address this problem and improve the robustness of neural networks, we investigate the fundamental mechanisms behind adversarial examples and propose a novel robust training method via regulating adversarial gradients. The regulation effectively squeezes the adversarial gradients of neural networks and significantly increases the difficulty of adversarial example generation. Without any adversarial example involved, the robust training method could generate naturally robust networks, which are near-immune to various types of adversarial examples. Experiments show the naturally robust networks can achieve optimal accuracy against Fast Gradient Sign Method (FGSM) and C&W attacks on MNIST, Cifar10, and Google Speech Command dataset. Moreover, our proposed method also provides neural networks with consistent robustness against transferable attacks.

Trains have been a prominent mode of long-distance travel for decades, especially in the countries with a significant land area and large population. India, with a population of 1.324 billion people in 2016, has a railway system of network route length of 66, 687 kilometers, with 11, 122 locomotives, 7, 216 stations, that served 8.107 billion ridership in 2016 [7]. The Indian railway system is fourth largest in the world in terms of network size. However its trains are plagued with endemic delays that can be credited to (a) obsolete technology, e.g., dated rail engines, (b) size, e.g., large network structure and high railway traffic, (c) weather, e.g., fog in winter months in north India and rains during summer monsoons countrywide. In this paper, we take the initial steps in understanding and predicting train delays.

Deep learning on graph structures has shown exciting results in various applications. However, few attentions have been paid to the robustness of such models, in contrast to numerous research work for image or text adversarial attack and defense. In this paper, we focus on the adversarial attacks that fool the model by modifying the combinatorial structure of data. We first propose a reinforcement learning based attack method that learns the generalizable attack policy, while only requiring prediction labels from the target classifier. Also, variants of genetic algorithms and gradient methods are presented in the scenario where prediction confidence or gradients are available. We use both synthetic and real-world data to show that, a family of Graph Neural Network models are vulnerable to these attacks, in both graph-level and node-level classification tasks. We also show such attacks can be used to diagnose the learned classifiers.

The chest X-ray (CXR) is by far the most commonly performed radiological examination for screening and diagnosis of many cardiac and pulmonary diseases. There is an immense world-wide shortage of physicians capable of providing rapid and accurate interpretation of this study. A radiologist-driven analysis of over two million CXR reports generated an ontology including the 40 most prevalent pathologies on CXR. By manually tagging a relatively small set of sentences, we were able to construct a training set of 959k studies. A deep learning model was trained to predict the findings given the patient frontal and lateral scans. For 12 of the findings we compare the model performance against a team of radiologists and show that in most cases the radiologists agree on average more with the algorithm than with each other.

H.H. Sheikh Abdullah bin Zayed Al Nahyan, Minister of Foreign Affairs and International Cooperation, has visited the headquarters of'Element AI', a Canadian artificial intelligence company, within the framework of his official visit to Canada.

The brand new iPhone software can now be downloaded. When Apple unveiled iOS 12 – the new operating system, which brings features including animoji changes and new ways of limiting how long users spend on their phones – it gave it to developers so they could start getting their apps ready for the new software. It won't be fully available to the public until this autumn, probably alongside new iPhones. However, that developer version of iOS 12 is now out in the wild and so it is possible for anyone to download it to their device should they want to. Doing so involves a number of risks, however – all of them not worth taking.

MIT researchers developed a virtual reality-based system, "Flight Goggles", that allows UAV navigate rooms while avoiding virtual obstacles. "We think this is a game-changer in the development of drone technology, for drones that go fast," said Sertac Karaman, Associate Professor of Aeronautics and Astronautics at MIT. "In the upcoming years, we want to enter a drone racing competition with an autonomous drone, and beat the best human player," added Karaman. The virtual images collected by the UAV reportedly occur at 90 frames per second, which is triple, the speed that the human eye can capture. "The drone will be flying in an empty room, but will be'hallucinating' a completely different environment, and will learn in that environment," explained Karaman.

Alibaba Group is powering ahead with a range of AI research and initiatives in a bid to realise its vision: To make it easy to do business everywhere and anywhere. That's according to an Alibaba Group chief scientist and associate dean of machine intelligence and technology, Xiaofeng Ren, who spoke at CeBIT about how to develop AI applications that power the future of business. "Alibaba has changed the everyday life of the Chinese in China. Looking forward, our visionary leader, Jack Ma, wants us to be able to reach two billion consumers and to help 10 million businesses around the world. That's a very big call, but we already have half of the platforms in place."