Variational Auto-encoders (VAEs) have been very successful as methods for forming compressed latent representations of complex, often high-dimensional, data. In this paper, we derive an alternative variational lower bound from the one common in VAEs, which aims to minimize aggregate information loss. Using our lower bound as the objective function for an auto-encoder enables us to place a prior on the bulk statistics, corresponding to an aggregate posterior of all latent codes, as opposed to a single code posterior as in the original VAE. This alternative form of prior constraint allows individual posteriors more flexibility to preserve necessary information for good reconstruction quality. We further derive an analytic approximation to our lower bound, leading to our proposed model - WiSE-VAE. Through various examples, we demonstrate that WiSE-VAE can reach excellent reconstruction quality in comparison to other state-of-the-art VAE models, while still retaining the ability to learn a smooth, compact representation.

Deep reinforcement learning (DRL) has gained a lot of attention in recent years, and has been proven to be able to play Atari games and Go at or above human levels. However, those games are assumed to have a small fixed number of actions and could be trained with a simple CNN network. In this paper, we study a special class of Asian popular card games called Dou Di Zhu, in which two adversarial groups of agents must consider numerous card combinations at each time step, leading to huge number of actions. We propose a novel method to handle combinatorial actions, which we call combinational Q-learning (CQL). We employ a two-stage network to reduce action space and also leverage order-invariant max-pooling operations to extract relationships between primitive actions. Results show that our method prevails over state-of-the art methods like naive Q-learning and A3C. We develop an easy-to-use card game environments and train all agents adversarially from sractch, with only knowledge of game rules and verify that our agents are comparative to humans. Our code to reproduce all reported results will be available online.

Despite being popularly used in many application domains, neural network models have been found to be vulnerable to adversarial examples, examples formed by applying imperceptible perturbation on legitimate examples from the datasets. Adversarial examples can pose potential risks on safety and security of real-world applications. However, existing defense approaches are still vulnerable to adversarial examples, especially in a white-box attack scenario. To address this problem, we propose a new defense approach, named MULDEF, based on robustness diversity. Our approach consists of (1) a general defense framework based on multiple models and (2) a technique for generating these multiple models to achieve high defense capability. In particular, given a target model to defend, our framework includes multiple models (constructed from the target model) to form a model family. The model family is designed to achieve robustness diversity (i.e., an adversarial example successfully attacking one model cannot succeed in attacking other models in the family). At runtime, a model is randomly selected from the family to be applied on each input example. Our general framework can inspire rich future research to construct a desirable model family achieving higher robustness diversity. Our evaluation results show that MULDEF (with only up to 5 models in the family) can already substantially improve the target model's accuracy on adversarial examples by 35-74% in a white-box attack scenario, while maintaining similar accuracy on legitimate examples as the target model.

For manufacturers, managing the supply chain from beginning to end has been like traveling two superhighways interrupted by a long stretch of dirt road. Manufacturers have benefited from increasingly powerful tools for demand planning and logistics management – the first and last parts of their supply chains – but tracking the performance of manufacturing production across the supply chain has remained stuck in the era of clipboards, whiteboards, spreadsheets and manually assembled reports. For most companies, understanding machine capacity, throughput, efficiency, and quality across the supply chain remains a black box. Companies that rely heavily on contract manufacturers have even less visibility – challenged by partners with different systems, processes, and levels of willingness to collaborate. Today's supply chain monitoring systems lack the ability to look at machine and part/batch-level data across the supply chain, limiting a global manufacturer's ability to manage their supplier base as an integrated platform.

Apple's 2019 line-up might have been revealed in a new report by a respected Apple analyst. Just about everything in Apple's line-up will be getting updated, according to the report. Not only will long-awaited new products arrive – such as the AirPower charging mat – there will be multiple new Mac computers and iPhones, it suggests. Three of the products will include entirely new designs, claimed Ming-Chi Kuo in the new report. Each of those are Macs: there will be a large Mac Pro that could have up to a 16.5-inch display, a big new monitor and an all new Mac Pro to go alongside it.

On the heels of President Trump's State of the Union address making a case for increased border security and other initiatives, federal agencies have been moving quickly on artificial intelligence, Internet of Things and robocall spoofing. Meanwhile, the EU and Japan have entered into an agreement on cross-border data flows, and India is taking action on investment in e-commerce. Want our Strategic Policy Advisory team to take a look at other topics? Let us know in the comments! Earlier this week, President Trump signed the American AI Initiative, an Executive Order directing federal agencies to develop new AI R&D budgets, share resources with academia and industry, create educational programs to improve the AI talent pipeline, and develop regulatory guidances for AI implementation that balance innovation with civil liberties.

Champion debater Harish Natarajan argues against IBM Debater, represented by a screen with a blue oval, in a competition at the IBM Think conference. Champion debater Harish Natarajan argues against IBM Debater, represented by a screen with a blue oval, in a competition at the IBM Think conference. Champion debater Harish Natarajan argues against IBM Debater, represented by a screen with a blue oval, in a competition at the IBM Think conference. Champion debater Harish Natarajan argues against IBM Debater, represented by a screen with a blue oval, in a competition at the IBM Think conference. IBM fell short in its latest attempt to prove machines can triumph over man.

China steps up plans for using artificial intelligence to strengthen its military; Bill Hemmer reports. China wants to become the world leader in artificial intelligence (AI) by 2030, and President Donald Trump is taking the necessary steps to ensure that won't happen. Last week, the president signed an executive order, the American AI Initiative, that will accelerate America's pursuit of the sophisticated AI capabilities necessary to maintain our military and economic dominance. The Chinese government exercises significant control over its domestic technology firms and frequently steals intellectual property from foreign companies that operate in the country. In addition, it has poured sizable resources into its AI programs in a deliberate effort to eclipse America in the race to perfect this world-changing technology.

The panel of speakers was Alessandro Annoni, JRC, Setting the scene, Ashley Fox, MEP and the European Parliament rapporteur on a Comprehensive European industrial policy on artificial intelligence and robotics, Catelijne Muller, Chair of the Study group on artificial intelligence, and Mady Delvaux, MEP and European Parliament rapporteur on Civil Rules on robotics. Alessandro Annoni addressed the conference with a general discussion of the "strong political tensions" that were arising from Artificial Intelligence. Defining the tech, Annoni said: "Artificial intelligence should not be considered a simple technology…it is a collection of technologies. It is a new paradigm that is aiming to give more power to the machine. It's a technology that will replace humans in some cases."

Much more rapidly than anyone originally thought possible, facial recognition technology has become part of the cultural mainstream. Facebook, for example, now uses AI-powered facial recognition software as part of its core social networking platform to identify people, while law enforcement agencies around the world have experimented with facial recognition surveillance cameras to reduce crime and improve public safety. But now it looks like society is finally starting to wake up to the immense privacy implications of real-time facial recognition surveillance. For example, San Francisco is now considering an outright ban on facial recognition surveillance. If pending legislation known as "Stop Secret Surveillance" passes, this would make San Francisco the first city ever to ban (and not just regulate) facial recognition technology.