In many classification problems a classifier should be robust to small variations in the input vector. This is a desired property not only for particular transformations, such as translation and rotation in image classification problems, but also for all others for which the change is small enough to retain the object perceptually indistinguishable. We propose two extensions of the backpropagation algorithm that train a neural network to be robust to variations in the feature vector. While the first of them enforces robustness of the loss function to all variations, the second method trains the predictions to be robust to a particular variation which changes the loss function the most. The second methods demonstrates better results, but is slightly slower. We analytically compare the proposed algorithm with two the most similar approaches (Tangent BP and Adversarial Training), and propose their fast versions. In the experimental part we perform comparison of all algorithms in terms of classification accuracy and robustness to noise on MNIST and CIFAR-10 datasets. Additionally we analyze how the performance of the proposed algorithm depends on the dataset size and data augmentation.

Artificial intelligence (AI) research has explored a variety of problems and approaches since its inception, but for the last 20 years or so has been focused on the problems surrounding the construction of intelligent agents — systems that perceive and act in some environment. In this context, "intelligence" is related to statistical and economic notions of rationality — colloquially, the ability to make good decisions, plans, or inferences. The adoption of probabilistic and decision-theoretic representations and statistical learning methods has led to a large degree of integration and cross-fertilization among AI, machine learning, statistics, control theory, neuroscience, and other fields. The establishment of shared theoretical frameworks, combined with the availability of data and processing power, has yielded remarkable successes in various component tasks such as speech recognition, image classification, autonomous vehicles, machine translation, legged locomotion, and question-answering systems. As capabilities in these areas and others cross the threshold from laboratory research to economically valuable technologies, a virtuous cycle takes hold whereby even small improvements in performance are worth large sums of money, prompting greater investments in research. There is now a broad consensus that AI research is progressing steadily, and that its impact on society is likely to increase. The potential benefits are huge, since everything that civilization has to offer is a product of human intelligence; we cannot predict what we might achieve when this intelligence is magnified by the tools AI may provide, but the eradication of disease and poverty are not unfathomable. Because of the great potential of AI, it is important to research how to reap its benefits while avoiding potential pitfalls. The progress in AI research makes it timely to focus research not only on making AI more capable, but also on maximizing the societal benefit of AI. Such considerations motivated the AAAI 2008–09 Presidential Panel on Long-Term AI Futures and other projects on AI impacts, and constitute a significant expansion of the field of AI itself, which up to now has focused largely on techniques that are neutral with respect to purpose. We recommend expanded research aimed at ensuring that increasingly capable AI systems are robust and beneficial: our AI systems must do what we want them to do. The attached research priorities document [see page X] gives many examples of such research directions that can help maximize the societal benefit of AI. This research is by necessity interdisciplinary, because it involves both society and AI. It ranges from economics, law and philosophy to computer security, formal methods and, of course, various branches of AI itself. In summary, we believe that research on how to make AI systems robust and beneficial is both important and timely, and that there are concrete research directions that can be pursued today.

We consider an adversarial formulation of the problem ofpredicting a time series with square loss. The aim is to predictan arbitrary sequence of vectors almost as well as the bestsmooth comparator sequence in retrospect. Our approach allowsnatural measures of smoothness such as the squared norm ofincrements. More generally, we consider a linear time seriesmodel and penalize the comparator sequence through the energy ofthe implied driving noise terms. We derive the minimax strategyfor all problems of this type and show that it can be implementedefficiently. The optimal predictions are linear in the previousobservations. We obtain an explicit expression for the regret interms of the parameters defining the problem. For typical,simple definitions of smoothness, the computation of the optimalpredictions involves only sparse matrices. In the case ofnorm-constrained data, where the smoothness is defined in termsof the squared norm of the comparator's increments, we show thatthe regret grows as $T/\sqrt{\lambda_T}$, where $T$ is the lengthof the game and $\lambda_T$ is an increasing limit on comparatorsmoothness.

Hamiltonian Monte Carlo (HMC) is a successful approach for sampling from continuous densities. However, it has difficulty simulating Hamiltonian dynamics with non-smooth functions, leading to poor performance. This paper is motivated by the behavior of Hamiltonian dynamics in physical systems like optics. We introduce a modification of the Leapfrog discretization of Hamiltonian dynamics on piecewise continuous energies, where intersections of the trajectory with discontinuities are detected, and the momentum is reflected or refracted to compensate for the change in energy. We prove that this method preserves the correct stationary distribution when boundaries are affine. Experiments show that by reducing the number of rejected samples, this method improves on traditional HMC.

The BusPlus project aims at improving the off-peak hours public transit service in Canberra, Australia. To address the difficulty of covering a large geographic area, BusPlus proposes a hub and shuttle model consisting of a combination of a few high-frequency bus routes between key hubs and a large number of shuttles that bring passengers from their origin to the closest hub and take them from their last bus stop to their destination. This paper focuses on the design of bus network and proposes an efficient solving method to this multimodal network design problem based on the Benders decomposition method. Starting from a MIP formulation of the problem, the paper presents a Benders decomposition approach using dedicated solution techniques for solving independent sub-problems, Pareto optimal cuts, cut bundling, and core point update. Computational results on real-world data from Canberra's public transit system justify the design choices and show that the approach outperforms the MIP formulation by two orders of magnitude. Moreover, the results show that the hub and shuttle model may decrease transit time by a factor of 2, while staying within the costs of the existing transit system.

We introduce a novel approach to Compressed Path Databases, space efficient oracles used to very quickly identify the first edge on a shortest path. Our algorithm achieves query running times on the 100 nanosecond scale, being significantly faster than state-of-the-art first-move oracles from the literature. Space consumption is competitive, due to a compression approach that rearranges rows and columns in a first-move matrix and then performs run length encoding (RLE) on the contents of the matrix. One variant of our implemented system was, by a convincing margin, the fastest entry in the 2014 Grid-Based Path Planning Competition. We give a first tractability analysis for the compression scheme used by our algorithm. We study the complexity of computing a database of minimum size for general directed and undirected graphs. We find that in both cases the problem is NP-complete. We also show that, for graphs which can be decomposed along articulation points, the problem can be decomposed into independent parts, with a corresponding reduction in its level of difficulty. In particular, this leads to simple and tractable algorithms with linear running time which yield optimal compression results for trees.

We study the problem of computing possible and necessary winners for partially specified weighted and unweighted tournaments. This problem arises naturally in elections with incompletely specified votes, partially completed sports competitions, and more generally in any scenario where the outcome of some pairwise comparisons is not yet fully known. We specifically consider a number of well-known solution concepts---including the uncovered set, Borda, ranked pairs, and maximin---and show that for most of them, possible and necessary winners can be identified in polynomial time. These positive algorithmic results stand in sharp contrast to earlier results concerning possible and necessary winners given partially specified preference profiles.

The paper proposes a fresh look at the concept of goal and advances that motivational attitudes like desire, goal and intention are just facets of the broader notion of (acceptable) outcome. We propose to encode the preferences of an agent as sequences of "alternative acceptable outcomes". We then study how the agent's beliefs and norms can be used to filter the mental attitudes out of the sequences of alternative acceptable outcomes. Finally, we formalise such intuitions in a novel Modal Defeasible Logic and we prove that the resulting formalisation is computationally feasible.

Finding high quality plans for large planning problems is hard. Although some current anytime planners are often able to improve plans quickly, they tend to reach a limit at which the plans produced are still very far from the best possible, but these planners fail to find any further improvement, even when given several hours of runtime. We present an approach to continuing plan quality optimisation at larger time scales, and its implementation in a system called BDPO2. Key to this approach is a decomposition into subproblems of improving parts of the current best plan. The decomposition is based on block deordering, a form of plan deordering which identifies hierarchical plan structure. BDPO2 can be seen as an application of the large neighbourhood search (LNS) local search strategy to planning, where the neighbourhood of a plan is defined by replacing one or more subplans with improved subplans. On-line learning is also used to adapt the strategy for selecting subplans and subplanners over the course of plan optimisation. Even starting from the best plans found by other means, BDPO2 is able to continue improving plan quality, often producing better plans than other anytime planners when all are given enough runtime. The best results, however, are achieved by a combination of different techniques working together.

Answering conjunctive queries over ontology-enriched datasets is a core reasoning task for many applications. Query answering is, however, computationally very expensive, which has led to the development of query answering procedures that sacrifice either expressive power of the ontology language, or the completeness of query answers in order to improve scalability. In this paper, we describe a hybrid approach to query answering over OWL 2 ontologies that combines a datalog reasoner with a fully-fledged OWL 2 reasoner in order to provide scalable `pay-as-you-go' performance. The key feature of our approach is that it delegates the bulk of the computation to the datalog reasoner and resorts to expensive OWL 2 reasoning only as necessary to fully answer the query. Furthermore, although our main goal is to efficiently answer queries over OWL 2 ontologies and data, our technical results are very general and our approach is applicable to first-order knowledge representation languages that can be captured by rules allowing for existential quantification and disjunction in the head; our only assumption is the availability of a datalog reasoner and a fully-fledged reasoner for the language of interest, both of which are used as `black boxes'. We have implemented our techniques in the PAGOdA system, which combines the datalog reasoner RDFox and the OWL 2 reasoner HermiT. Our extensive evaluation shows that PAGOdA succeeds in providing scalable pay-as-you-go query answering for a wide range of OWL 2 ontologies, datasets and queries.