Controllers for dynamical systems that operate in safety-critical settings must account for stochastic disturbances. Such disturbances are often modeled as process noise in a dynamical system, and common assumptions are that the underlying distributions are known and/or Gaussian. In practice, however, these assumptions may be unrealistic and can lead to poor approximations of the true noise distribution. We present a novel controller synthesis method that does not rely on any explicit representation of the noise distributions. In particular, we address the problem of computing a controller that provides probabilistic guarantees on safely reaching a target, while also avoiding unsafe regions of the state space. First, we abstract the continuous control system into a finite-state model that captures noise by probabilistic transitions between discrete states. As a key contribution, we adapt tools from the scenario approach to compute probably approximately correct (PAC) bounds on these transition probabilities, based on a finite number of samples of the noise. We capture these bounds in the transition probability intervals of a so-called interval Markov decision process (iMDP). This iMDP is, with a user-specified confidence probability, robust against uncertainty in the transition probabilities, and the tightness of the probability intervals can be controlled through the number of samples. We use state-of-the-art verification techniques to provide guarantees on the iMDP and compute a controller for which these guarantees carry over to the original control system. In addition, we develop a tailored computational scheme that reduces the complexity of the synthesis of these guarantees on the iMDP. Benchmarks on realistic control systems show the practical applicability of our method, even when the iMDP has hundreds of millions of transitions.

We introduce the framework of qualitative optimization problems (or, simply, optimization problems) to represent preference theories. The formalism uses separate modules to describe the space of outcomes to be compared (the generator) and the preferences on outcomes (the selector). We consider two types of optimization problems. They differ in the way the generator, which we model by a propositional theory, is interpreted: by the standard propositional logic semantics, and by the equilibrium-model (answer-set) semantics. Under the latter interpretation of generators, optimization problems directly generalize answer-set optimization programs proposed previously. We study strong equivalence of optimization problems, which guarantees their interchangeability within any larger context. We characterize several versions of strong equivalence obtained by restricting the class of optimization problems that can be used as extensions and establish the complexity of associated reasoning tasks. Understanding strong equivalence is essential for modular representation of optimization problems and rewriting techniques to simplify them without changing their inherent properties.

The problem of fair division of indivisible goods is a fundamental problem of resource allocation in multi-agent systems, also studied extensively in social choice. Recently, the problem was generalized to the case when goods form a graph and the goal is to allocate goods to agents so that each agent's bundle forms a connected subgraph. For the maximin share fairness criterion, researchers proved that if goods form a tree, an allocation offering each agent a bundle of at least her maximin share value always exists. Moreover, it can be found in polynomial time. In this paper we consider the problem of maximin share allocations of goods on a cycle. Despite the simplicity of the graph, the problem turns out to be significantly harder than its tree version. We present cases when maximin share allocations of goods on cycles exist and provide in this case results on allocations guaranteeing each agent a certain fraction of her maximin share. We also study algorithms for computing maximin share allocations of goods on cycles.

Most matching algorithms are centralized in that a single agent determines how other agents are matched together. This is contrary to how humans form matches in the real world. In this work, we propose three decentralized approaches for finding matchings that are inspired by three techniques that humans use to find matches. The first is to have individuals wander a grid environment, interacting and deciding preferences over potential partners. The second uses affiliation networks where agencies recommend potential partners. The third is based on small-world social networks, where we assume that individuals probabilistically introduce their friends to one another. we introduce a heuristic algorithm that can be used in each of these environments. We also explore how this algorithm can scale to a large number of agents.

Reinforcement Learning has the limitation that problems become too large very quickly. Dividing the problem into a hierarchy of subtasks allows for a strategy of divide and conquer, which is what makes Hierarchical Reinforcement Learning (HRL) algorithms often more efficient at finding solutions quicker than more naive approaches. One of the biggest challenges with HRL is the construction of a hierarchy to be used by the algorithm. Hierarchies are often designed by a person using their own knowledge of the problem. We propose method for automatically discovering task hierarchies based on a data mining technique, Association Rule Learning (ARL). These hierarchies can then be applied to Semi-Markov Decision Process (SMDP) problems using the options technique

There is an increasing demand for content filtering and flagging on social media in relation to cybersecurity and social media conduct monitoring. This task is challenging and there is a large body of recent work that addresses it within the Natural Language and Video Processing communities. In this work, we propose two novel perspectives on this problem and provide preliminary evidence for their potential success. First, for text-based data, we utilize the current state of the art topic-based summarization algorithms and provide an interactive topic-conditioning approach to enable multiple summarizations based on different highlighted topics. Second, due to the interactivity aspect, we are able to characterize how this approach can be integrated within the process of a human analyst to improve both the quality of filtered data and the effort.

Spectral clustering has found extensive use in many areas. Most traditional spectral clustering algorithms work in three separate steps: similarity graph construction; continuous labels learning; discretizing the learned labels by k-means clustering. Such common practice has two potential flaws, which may lead to severe information loss and performance degradation. First, predefined similarity graph might not be optimal for subsequent clustering. It is well-accepted that similarity graph highly affects the clustering results. To this end, we propose to automatically learn similarity information from data and simultaneously consider the constraint that the similarity matrix has exact c connected components if there are c clusters. Second, the discrete solution may deviate from the spectral solution since k-means method is well-known as sensitive to the initialization of cluster centers. In this work, we transform the candidate solution into a new one that better approximates the discrete one. Finally, those three subtasks are integrated into a unified framework, with each subtask iteratively boosted by using the results of the others towards an overall optimal solution. It is known that the performance of a kernel method is largely determined by the choice of kernels. To tackle this practical problem of how to select the most suitable kernel for a particular data set, we further extend our model to incorporate multiple kernel learning ability. Extensive experiments demonstrate the superiority of our proposed method as compared to existing clustering approaches.

Ethical Considerations in Artificial Intelligence Courses

The recent surge in interest in ethics in artificial intelligence may leave many educators wondering how to address moral, ethical, and philosophical issues in their AI courses. As instructors we want to develop curriculum that not only prepares students to be artificial intelligence practitioners, but also to understand the moral, ethical, and philosophical impacts that artificial intelligence will have on society. In this article we provide practical case studies and links to resources for use by AI educators. We also provide concrete suggestions on how to integrate AI ethics into a general artificial intelligence course and how to teach a stand-alone artificial intelligence ethics course.

Conditional preference networks (CP-nets) are an intuitive and expressive representation for qualitative preferences. Such models must somehow be acquired. Psychologists argue that direct elicitation is suspect. On the other hand, learning general CP-nets from pairwise comparisons is NP-hard, and — for some notions of learning — this extends even to the simplest forms of CP-nets. We introduce a novel, concise encoding of binary-valued, tree-structured CP-nets that supports the first local-search-based CP-net learning algorithms. While exact learning of binary-valued, tree-structured CP-nets — for a strict, entailment-based notion of learning — is already in P, our algorithm is the first space-efficient learning algorithm that gracefully handles noisy (i.e., realistic) comparison sets.