Science Magazine's report on Friday that an artificial intelligence system was caught stealing banking customers' money may have made you rethink vesting your funds in the burgeoning technology. But have no fear – the article was an April Fool's joke. While there are still valid and plenty of concerns about AI, the threat that the so called "Deep Learning Interface for Accounting (DELIA)" may embezzle your funds is not one of them. In fact, AI systems are a remarkably valuable resource in finance. Established banks and startups alike implement AI to detect fraud.

Awareness of Artificial Intelligence (AI), Machine Learning (ML) and Neuro-linguistic Programming (NLP) is ever-increasing, with AI being one of the most talked about technologies across multiple industries. But exactly who is using it? And what are they using it for? The Pistoia Alliance is a not-for-profit organization that aims to promote collaboration between pharmaceutical, technology, and life science professionals to lower barriers to R&D innovation by addressing challenges related to data aggregation and accessibility. One of their current projects focuses on promoting the use of AI in R&D.

PowerVision, a world leader in AI robotics and technology unveils the PowerRay, an innovative underwater robot with groundbreaking features to revolutionize recreational fishing enthusiasts and marine photography. As a financial hub and trusted gateway to the emerging Asian markets, PowerVision has strategically selected Singapore to lead the Southeast Asian release following the global launch of PowerRay in San Francisco (USA), Munich (Germany), Beijing (China) and Tokyo (Japan). Singapore is the fifth city to launch the world's first underwater drone, followed by Hong Kong in March, as well as Malaysia, Indonesia and Thailand in April. PowerRay is the first consumer-level lifestyle underwater robot from PowerVision and was first unveiled at CES 2017 as one of a series of marine and aerial "Power" drones including the PowerEye, PowerEgg, and PowerDolphin. They have been well received by the media and consumers for its advanced product concept and functional advantages.

We consider online detection strategies for identifying a change point in a stream of quantum particles allegedly prepared in identical states. We show that the identification of the change point can be done without error via sequential local measurements while attaining the optimal performance bound set by quantum mechanics. In this way, we establish the task of exactly identifying a quantum change point as an instance where local protocols are as powerful as global ones. The optimal online detection strategy requires only one bit of memory between subsequent measurements, and it is amenable to experimental realization with current technology.

We propose an architecture for VQA which utilizes recurrent layers to generate visual and textual attention. The memory characteristic of the proposed recurrent attention units offers a rich joint embedding of visual and textual features and enables the model to reason relations between several parts of the image and question. Our single model outperforms the first place winner on the VQA 1.0 dataset, performs within margin to the current state-of-the-art ensemble model. We also experiment with replacing attention mechanisms in other state-of-the-art models with our implementation and show increased accuracy. In both cases, our recurrent attention mechanism improves performance in tasks requiring sequential or relational reasoning on the VQA dataset.

Many practical machine learning tasks employ very deep convolutional neural networks. Such large depths pose formidable computational challenges in training and operating the network. It is therefore important to understand how fast the energy contained in the propagated signals (a.k.a. feature maps) decays across layers. In addition, it is desirable that the feature extractor generated by the network be informative in the sense of the only signal mapping to the all-zeros feature vector being the zero input signal. This "trivial null-set" property can be accomplished by asking for "energy conservation" in the sense of the energy in the feature vector being proportional to that of the corresponding input signal. This paper establishes conditions for energy conservation (and thus for a trivial null-set) for a wide class of deep convolutional neural network-based feature extractors and characterizes corresponding feature map energy decay rates. Specifically, we consider general scattering networks employing the modulus non-linearity and we find that under mild analyticity and high-pass conditions on the filters (which encompass, inter alia, various constructions of Weyl-Heisenberg filters, wavelets, ridgelets, ($\alpha$)-curvelets, and shearlets) the feature map energy decays at least polynomially fast. For broad families of wavelets and Weyl-Heisenberg filters, the guaranteed decay rate is shown to be exponential. Moreover, we provide handy estimates of the number of layers needed to have at least $((1-\varepsilon)\cdot 100)\%$ of the input signal energy be contained in the feature vector.

FRAMOS, a specialist in industrial image processing, has announced that it is developing state-of-the-art methods, such as 3D cameras, depth modules, and intelligent algorithms, for the IRiS artificial intelligence project for the reliable classification of packing scenarios and the analysis of container contents by intelligent robots. The IRiS project (Interactive Robotic System for Unloading of Sea Containers), being undertaken by FRAMOS, the Bremen Institute for Production und Logistics, and partners BLG Handelslogistik and Schulz Systemtechnik, is conducting research on the automated unloading of standard 40-foot containers. In the future, intelligent robots will carry out this difficult and predominantly manual task automatically. Germany's Federal Ministry of Transport and Digital Infrastructure (BMVI) is funding the three-year project with 2.2 million euros; and, TÜV Rheinland is on board as sponsor of the project. The majority of all sea containers shipped worldwide are unloaded and discharged in the port itself. These containers, with a capacity of 65 cubic meters (2295 cft) and a payload of 26 metric tons, can hold up to 1,800 parcels weighing up to 35 kg each.

A new report shows the finding of researchers attempting to teach killer whales mimic human sounds, and yes, even speak. You have to hear it to believe it. Scientists have demonstrated for the first time that orcas, also known as killer whales, can mimic human words, including "Amy," "Bye-Bye" and "One-Two-Three." Wikie, a 14-year-old female killer whale housed at Marineland Aquarium in Antibes, France, was tested by researchers including José Z. Abramson to get her to speak. Wikie had previously participated in an action imitation study, so she already knew the "copy" command, giving her a leg (or a fin) up when it came to "speaking."

Researchers from the University of Valladolid (Spain) have created a computer model based on neural networks that calculates the probability in Spanish provinces of corruption, as well as the conditions that favor it. This alert system confirms that the probabilities increase when the same party stays in government for more years. The study, published in Social Indicators Research, does not mention the provinces most prone to corruption so as not to generate controversy, explains one of the authors, Ivan Pastor, who says, "A greater propensity or high probability does not imply corruption will actually happen." The data indicate that the real estate tax, the exaggerated increase in the price of housing, the opening of bank branches and the creation of new companies are some of the variables that seem to induce public corruption, and when they are added together in a region, more rigorous control of public accounts might be warranted. "In addition, as might be expected, our model confirms that an increase in the number of years of government by the same political party increases the chances of corruption, regardless of whether or not the party governs with majority," says Pastor.

The prime minister says she wants the UK to lead the world in deciding how artificial intelligence can be deployed in a safe and ethical manner. In a speech at the World Economic Forum in Davos, Theresa May said a new advisory body, previously announced in the Autumn Budget, will co-ordinate efforts with other countries. In addition, she confirmed that the UK would join the Davos forum's own council on artificial intelligence. But others may have stronger claims. Earlier this week, Google picked France as the base for a new research centre dedicated to exploring how AI can be applied to health and the environment.