Design of effective health communication systems faces major challenges in terms of accessibility, trust, expert-to-lay knowledge translation, and persuasiveness. It is proposed that some of these challenges can be addressed by use of AI techniques in combination with empirically-based theoretical frameworks from the field of health communication and related areas. This symposium will bring together an interdisciplinary group of scholars to identify possible solutions. AI and health communication topics of interest include communication interventions; games, conversational agents, or dialogue systems for healthy behavior promotion; intelligent interactive monitoring of patient's environment and needs; intelligent interfaces supporting access to healthcare; patient-tailored decision support, explanation for informed consent, and retrieval and summarization of online healthcare information; risk communication and visualization; tailored access to electronic medical records; tailoring health information for low-literacy, low-numeracy, or under-served audiences; virtual healthcare counselors; and virtual patients for training healthcare professionals. Scholars from health communication and related disciplines (sociolinguistics, pragmatics, discourse studies, etc.) will participate in discussion on the following issues as they pertain to the symposium goals: health literacy; healthcare provider-consumer communication, risk communication, including written and visual formats; and use of behavioral, persuasion, and argumentation theories for healthy behavior promotion. By examining these issues, the symposium is expected to lay down conceptual foundations for guiding future advances in AI healthcare systems.

Traditional machine-learned ranking systems for web search are often trained to capture stationary relevance of documents to queries, which has limited ability to track non-stationary user intention in a timely manner. In recency search, for instance, the relevance of documents to a query on breaking news often changes significantly over time, requiring effective adaptation to user intention. In this paper, we focus on recency search and study a number of algorithms to improve ranking results by leveraging user click feedback. Our contributions are three-fold. First, we use real search sessions collected in a random exploration bucket for \emph{reliable} offline evaluation of these algorithms, which provides an unbiased comparison across algorithms without online bucket tests. Second, we propose a re-ranking approach to improve search results for recency queries using user clicks. Third, our empirical comparison of a dozen algorithms on real-life search data suggests importance of a few algorithmic choices in these applications, including generalization across different query-document pairs, specialization to popular queries, and real-time adaptation of user clicks.

We propose AllDiffPrecedence, a new global constraint that combines together an AllDifferent constraint with precedence constraints that strictly order given pairs of variables. We identify a number of applications for this global constraint including instruction scheduling and symmetry breaking. We give an efficient propagation algorithm that enforces bounds consistency on this global constraint. We show how to implement this propagator using a decomposition that extends the bounds consistency enforcing decomposition proposed for the AllDifferent constraint. Finally, we prove that enforcing domain consistency on this global constraint is NP-hard in general.

Recently, several researchers have found that cost-based satisficing search with A* often runs into problems. Although some "work arounds" have been proposed to ameliorate the problem, there has not been any concerted effort to pinpoint its origin. In this paper, we argue that the origins can be traced back to the wide variance in action costs that is observed in most planning domains. We show that such cost variance misleads A* search, and that this is no trifling detail or accidental phenomenon, but a systemic weakness of the very concept of "cost-based evaluation functions + systematic search + combinatorial graphs". We show that satisficing search with sized-based evaluation functions is largely immune to this problem.

We present the extention and application of a new unsupervised statistical learning technique--the Partition Decoupling Method--to gene expression data. Because it has the ability to reveal non-linear and non-convex geometries present in the data, the PDM is an improvement over typical gene expression analysis algorithms, permitting a multi-gene analysis that can reveal phenotypic differences even when the individual genes do not exhibit differential expression. Here, we apply the PDM to publicly-available gene expression data sets, and demonstrate that we are able to identify cell types and treatments with higher accuracy than is obtained through other approaches. By applying it in a pathway-by-pathway fashion, we demonstrate how the PDM may be used to find sets of mechanistically-related genes that discriminate phenotypes.

One of the first step in the realization of an automatic system of check recognition is the extraction of the handwritten area. We propose in this paper an hybrid method to extract these areas. This method is based on digit recognition by Fourier descriptors and different steps of colored image processing . It requires the bank recognition of its code which is located in the check marking band as well as the handwritten color recognition by the method of difference of histograms. The areas extraction is then carried out by the use of some mathematical morphology tools.

this paper presents an enhancement of the medial axis algorithm to be used for finding the optimal shortest path for developed cognitive map. The cognitive map has been developed, based on the architectural blueprint maps. The idea for using the medial-axis is to find main path central pixels; each center pixel represents the center distance between two side boarder pixels. The need for these pixels in the algorithm comes from the need of building a network of nodes for the path, where each node represents a turning in the real world (left, right, critical left, critical right...). The algorithm also ignores from finding the center pixels paths that are too small for intelligent robot navigation. The Idea of this algorithm is to find the possible shortest path between start and end points. The goal of this research is to extract a simple, robust representation of the shape of the cognitive map together with the optimal shortest path between start and end points. The intelligent robot will use this algorithm in order to decrease the time that is needed for sweeping the targeted building.

The identification of the language of the script is an important stage in the process of recognition of the writing. There are several works in this research area, which treat various languages. Most of the used methods are global or statistical. In this present paper, we study the possibility of using the features of scripts to identify the language. The identification of the language of the script by characteristics returns the identification in the case of multilingual documents less difficult. We present by this work, a study on the possibility of using the structural features to identify the Arabic language from an Arabic / Latin text.

Sequential prediction problems such as imitation learning, where future observations depend on previous predictions (actions), violate the common i.i.d. assumptions made in statistical learning. This leads to poor performance in theory and often in practice. Some recent approaches provide stronger guarantees in this setting, but remain somewhat unsatisfactory as they train either non-stationary or stochastic policies and require a large number of iterations. In this paper, we propose a new iterative algorithm, which trains a stationary deterministic policy, that can be seen as a no regret algorithm in an online learning setting. We show that any such no regret algorithm, combined with additional reduction assumptions, must find a policy with good performance under the distribution of observations it induces in such sequential settings. We demonstrate that this new approach outperforms previous approaches on two challenging imitation learning problems and a benchmark sequence labeling problem.

We show that the disagreement coefficient of certain smooth hypothesis classes is $O(m)$, where $m$ is the dimension of the hypothesis space, thereby answering a question posed in \cite{friedman09}.