This work compares two actions' concurrency and co-occurrence employed in temporal modeling languages, one with a PDDL-style action modeling languages used by the AI planning community, exclusion mechanism, and another with an explicit and argue that they explain why MILP or SMT have notion of resources, and investigates their seemed unattractive. Specifically, we observe that PDDL 2.1 implications on constraint-based search. The first [Fox and Long, 2003] induces temporal gaps between consecutive mechanism forces temporal gaps in action schedules interdependent actions, and these gaps often induce and have a high performance penalty. The second twice the number of steps in the plans than what is necessary, mechanism avoids the gaps, with dramatically with strong negative performance implications. The gaps are improved performance.

Person re-identification concerns the matching of pedestrians across disjoint camera views. Due to the changes of viewpoints, lighting conditions and camera features, images of the same person from different views always appear differently, and thus feature representations across disjoint camera views of the same person follow different distributions. In this work, we propose an effective, low cost and easy-to-apply schema called the Mirror Representation, which embeds the view-specific feature transformation and enables alignment of the feature distributions across disjoint views for the same person. The proposed Mirror Representation is also designed to explicitly model the relation between different view-specific transformations and meanwhile control their discrepancy. With our Mirror Representation, we can enhance existing subspace/metric learning models significantly, and we particularly show that kernel marginal fisher analysis significantly outperforms the current state-of-the-art methods through extensive experiments on VIPeR, PRID450S and CUHK01.

When a transition probability matrix is represented as the product of two stochastic matrices, swapping the factors of the multiplication yields another transition matrix that retains some fundamental characteristics of the original. Since the new matrix can be much smaller than its precursor, replacing the former for the latter can lead to significant savings in terms of computational effort. This strategy, dubbed the "stochastic-factorization trick," can be used to compute the stationary distribution of a Markov chain, to determine the fundamental matrix of an absorbing chain, and to compute a decision policy via dynamic programming or reinforcement learning. In this paper we show that the stochastic-factorization trick can also provide benefits in terms of the number of samples needed to estimate a transition matrix. We introduce a probabilistic interpretation of a stochastic factorization and build on the resulting model to develop an algorithm to compute the factorization directly from data. If the transition matrix can be well approximated by a low-order stochastic factorization, estimating its factors instead of the original matrix reduces significantly the number of parameters to be estimated. Thus, when compared to estimating the transition matrix directly via maximum likelihood, the proposed method is able to compute approximations of roughly the same quality using less data. We illustrate the effectiveness of the proposed algorithm by using it to help a reinforcement learning agent learn how to play the game of blackjack.

Recently, a number of papers have shown that for many domains, using multiple heuristics in independent searches performs better than combining them into a single heuristic. Furthermore, using a large number of “weak” heuristics could potentially eliminate the need for the careful design of a few. The standard approach to distribute computation in these multi-heuristic searches is to rotate through the heuristics in a round-robin fashion. However, this strategy can be inefficient especially in the case when only a few of the heuristics are leading to progress. In this paper, we present two principled methods to adaptively distribute computation time among the different searches of the Multi- Heuristic A* algorithm. The first method, Meta-A*, constructs and searches a meta-graph, which represents the problem of finding the best heuristic as the problem of minimizing the total number of expansions. The second treats the scheduling of searches with different heuristics as a multi-armed bandit problem. It applies Dynamic Thompson Sampling (DTS) to keep track of what searches are making progress the most and continuously re-computes the schedule of searches based on this information. We provide a theoretical analysis and compare our new strategies with the round-robin method on a 12-DOF full-body motion planning problem and on sliding tile puzzle problems. In these experiments, we used up to 20 heuristics and observed a several times speedup without loss in solution quality.

Recently, several Web-scale knowledge harvesting systems have been built, each of which is competent at extracting information from certain types of data (e.g., unstructured text, structured tables on the web, etc.). In order to determine the response to a new query posed to such systems (e.g., is sugar a healthy food?), it is useful to integrate opinions from multiple systems. If a response is desired within a specific time budget (e.g., in less than 2 seconds), then maybe only a subset of these resources can be queried. In this paper, we address the problem of knowledge integration for on-demand time-budgeted query answering. We propose a new method, AskWorld, which learns a policy that chooses which queries to send to which resources, by accommodating varying budget constraints that are available only at query (test) time. Through extensive experiments on real world datasets, we demonstrate AskWorld’s capability in selecting most informative resources to query within test-time constraints, resulting in improved performance compared to competitive baselines.

We propose a new approach that exploits the good properties of core-guided and model-guided MaxSAT solvers. In particular, we show how to effectively exploit the structure of unsatisfiable cores in MaxSAT instances. Experimental results on industrial instances show that the proposed approach outperforms both complete and incomplete state-of-the-art MaxSAT solvers at the last international MaxSAT Evaluation in terms of robustness and total number of solved instances.

We describe a novel convolutional neural network architecture with k-max pooling layer that is able to successfully recover the structure of Chinese sentences. This network can capture active features for unseen segments of a sentence to measure how likely the segments are merged to be the constituents. Given an input sentence, after all the scores of possible segments are computed, an efficient dynamic programming parsing algorithm is used to find the globally optimal parse tree. A similar network is then applied to predict syntactic categories for every node in the parse tree. Our networks archived competitive performance to existing benchmark parsers on the CTB-5 dataset without any task-specific feature engineering.

Recommender systems (RSs) provide an effective way of alleviating the information overload problem by selecting personalized choices. Online social networks and user-generated content provide diverse sources for recommendation beyond ratings, which present opportunities as well as challenges for traditional RSs. Although social matrix factorization (Social MF) can integrate ratings with social relations and topic matrix factorization can integrate ratings with item reviews, both of them ignore some useful information. In this paper, we investigate the effective data fusion by combining the two approaches, in two steps. First, we extend Social MF to exploit the graph structure of neighbors. Second, we propose a novel framework MR3 to jointly model these three types of information effectively for rating prediction by aligning latent factors and hidden topics. We achieve more accurate rating prediction on two real-life datasets. Furthermore, we measure the contribution of each data source to the proposed framework.

This work explores the creation of ambiguous images, i.e., images that may induce multistable perception, by evolutionary means. Ambiguous images are created using a general purpose approach, composed of an expression-based evolutionary engine and a set of object detectors, which are trained in advance using Machine Learning techniques. Images are evolved using Genetic Programming and object detectors are used to classify them. The information gathered during classification is used to assign fitness. In a first stage, the system is used to evolve images that resemble a single object. In a second stage, the discovery of ambiguous images is promoted by combining pairs of object detectors. The analysis of the results highlights the ability of the system to evolve ambiguous images and the differences between computational and human ambiguous images.

Fuzzy Description Logics (DLs) are used to represent and reason about vague and imprecise knowledge that is inherent to many application domains. It was recently shown that the complexity of reasoning in finitely valued fuzzy DLs is often not higher than that of the underlying classical DL. We show that this does not hold for fuzzy extensions of the light-weight DL EL, which is used in many biomedical ontologies, under the Lukasiewicz semantics. The complexity of reasoning increases from PTime to ExpTime, even if only one additional truth value is introduced. The same lower bound holds also for infinitely valued Lukasiewicz extensions of EL.