Current approaches in approximate inference for Bayesian neural networks minimise the Kullback-Leibler divergence to approximate the true posterior over the weights. However, this approximation is without knowledge of the final application, and therefore cannot guarantee optimal predictions for a given task. To make more suitable task-specific approximations, we introduce a new loss-calibrated evidence lower bound for Bayesian neural networks in the context of supervised learning, informed by Bayesian decision theory. By introducing a lower bound that depends on a utility function, we ensure that our approximation achieves higher utility than traditional methods for applications that have asymmetric utility functions. Furthermore, in using dropout inference, we highlight that our new objective is identical to that of standard dropout neural networks, with an additional utility-dependent penalty term. We demonstrate our new loss-calibrated model with an illustrative medical example and a restricted model capacity experiment, and highlight failure modes of the comparable weighted cross entropy approach. Lastly, we demonstrate the scalability of our method to real world applications with per-pixel semantic segmentation on an autonomous driving data set.

We report on our first experiments to train deep neural networks that automatically translate informalized $\LaTeX{}$-written Mizar texts into the formal Mizar language. Using Luong et al.'s neural machine translation model (NMT), we tested our aligned informal-formal corpora against various hyperparameters and evaluated their results. Our experiments show that NMT is able to generate correct Mizar statements on more than 60 percent of the inference data, indicating that formalization through artificial neural network is a promising approach for automated formalization of mathematics. We present several case studies to illustrate our results.

Coordinating agents to complete a set of tasks with intercoupled temporal and resource constraints is computationally challenging, yet human domain experts can solve these difficult scheduling problems using paradigms learned through years of apprenticeship. A process for manually codifying this domain knowledge within a computational framework is necessary to scale beyond the ``single-expert, single-trainee" apprenticeship model. However, human domain experts often have difficulty describing their decision-making processes, causing the codification of this knowledge to become laborious. We propose a new approach for capturing domain-expert heuristics through a pairwise ranking formulation. Our approach is model-free and does not require enumerating or iterating through a large state space. We empirically demonstrate that this approach accurately learns multifaceted heuristics on a synthetic data set incorporating job-shop scheduling and vehicle routing problems, as well as on two real-world data sets consisting of demonstrations of experts solving a weapon-to-target assignment problem and a hospital resource allocation problem. We also demonstrate that policies learned from human scheduling demonstration via apprenticeship learning can substantially improve the efficiency of a branch-and-bound search for an optimal schedule. We employ this human-machine collaborative optimization technique on a variant of the weapon-to-target assignment problem. We demonstrate that this technique generates solutions substantially superior to those produced by human domain experts at a rate up to 9.5 times faster than an optimization approach and can be applied to optimally solve problems twice as complex as those solved by a human demonstrator.

Supervised learning algorithms are nowadays successfully scaling up to datasets that are very large in volume, leveraging the potential of in-memory cluster-computing Big Data frameworks. Still, massive datasets with a number of large-domain categorical features are a difficult challenge for any classifier. Most off-the-shelf solutions cannot cope with this problem. In this work we introduce DAC, a Distributed Associative Classifier. DAC exploits ensemble learning to distribute the training of an associative classifier among parallel workers and improve the final quality of the model. Furthermore, it adopts several novel techniques to reach high scalability without sacrificing quality, among which a preventive pruning of classification rules in the extraction phase based on Gini impurity. We ran experiments on Apache Spark, on a real large-scale dataset with more than 4 billion records and 800 million distinct categories. The results showed that DAC improves on a state-of-the-art solution in both prediction quality and execution time. Since the generated model is human-readable, it can not only classify new records, but also allow understanding both the logic behind the prediction and the properties of the model, becoming a useful aid for decision makers.

Department of Philosophy, Peking University In standard epistemic logic, agent names are usually assumed to be common knowledge. This is unreasonable for various applications. Inspired by term modal logic and assignment operators in dynamic logic, we introduce a lightweight modal predicate logic whose names are not rigid. The language can handle various de dicto /de re distinctions in a natural way. We show the decidability of the logic over arbitrary models and give a complete axiomatisation over S5 models.

After the successful completion of the production of the material-optimised concrete façade mullions, Fabio Scotto and Ena Lloret-Frischti of the Gramazio Kohler Research Group at ETH Zurich and the Chair for Physical Chemistry of Building Materials, ETH Zurich take a look back at the experiments and prototypes which were necessary in the development of a final robotic fabrication process. The integration of Smart Dynamic Casting (SDC) for the production of the façade mullions for the first floor of DFAB HOUSE has led us to the development of an adaptive robotic setup which allows us to produce custom-made reinforced concrete structures. Until the final development of a robust robotic process, we had to overcome several challenges during the experimental and prototypical phase. Scaling down the production system and minimizing the friction forces Our first main task was to scale down the production system to realise structures with a minimal cross section of 100 70 mm. This had a direct impact on the formwork system we were working with previously.

Video games these days are contingent on their realistic, detailed, immersive worlds. But building that takes lots of programmers many hours; Vice City, after all, was not built in a day. What if an AI could do some of the heavy lifting (so to speak) instead? Computer scientists at Italy's Politecnico di Milano found a way to let AI handle all the difficult parts of level design so that designers and gamers alike can get to the fun parts more quickly. The 1993 first-person shooter DOOM was a natural place to start. People run the thing on as much technology as possible -- there are people out there who will figure out how to run the game on anything with an operating system, from a thermostat to a calculator.

When industrial factories were introduced in Europe in the nineteenth century, even staunch critics of capitalism such as Friedrich Engels acknowledged that mass production necessitated centralized authority, regardless of whether the economic system was capitalist or socialist. In the twentieth century, theorists such as Langdon Winner extended this line of thinking to other technologies. He thought that the atom bomb, for example, should be considered an "inherently political artifact," because its "lethal properties demand that it be controlled by a centralized, rigidly hierarchical chain of command."

What will be the ethical limit in the development of Artificial Intelligence? What is the role of man in the face of automated systems? There are many questions that still arise when developing a strategy of Artificial Intelligence in companies and within society itself. What technologies are necessary to achieve standardization and better use of the opportunities that Artificial Intelligence opens to businesses? What role does security have in an AI environment?

Does the Voice Activated Assistant have a place in the business? There are lots of voice activated assistants available to help us with our daily tasks. Some are built into mobile devices and some are purchased as specific items. The key is that they all respond to commands and can interact with other devices in our lives. These devices are already within most business today, but may not be used to their full potential.