Power continues to be a driving force in central processing units (CPU) design. Most of the advanced breakthroughs in power have been in a realm that is applicable to workstation CPUs. Advanced power management systems will manage temperature, dynamic voltage scaling and dynamic frequency scaling in a CPU. The use of power management systems for microcontrollers and embedded CPUs has been modest, and mostly focuses on very large scale integration (VLSI) level optimizations compared to system level optimizations. In this paper, a dynamic frequency controlling (DFC) technique is introduced, to lay the foundation of a system level power management system for commercial microcontrollers. The DFC technique allows a commercial microcontroller to have minor modifications on both the hardware and software side, to allow the clock frequency to change to save power; results in this study show a 10% savings. By adding an additional layer of software abstraction at the interrupt level, the microcontroller can operate without having knowledge of the current clock frequency, and this can be accomplished without having to use an embedded operating system.

In dynamic Web service environments, the performance of the Internet is unpredictable; the reliability and effectiveness of remote Web services are also unclear. Therefore, it can hardly be guaranteed that the quality of Web service (QoWS) attributes of Web services do not fluctuate with the dynamic Web service environments. When a composite service is planned in the context of dynamic service environments, there is another aspect which has not been taken into account by existing works, namely, concurrency - the fact that multiple requests to a composite service may arrive at the same time. Considering the dynamics of Web service environments and concurrency of requests, we propose in this paper a service selection scheme which adopts top-k dominating queries to generate a composition solution rather than only select the best composition solution for a given request. The experimental results have investigated the efficiency and effectiveness of our approach and shown that it outperforms baseline and traditional methods for service selection.

We propose two new approaches that automatically parallelize Java programs at runtime. These approaches, which rely on run-time trace information collected during program execution, dynamically recompile Java byte code that can be executed in parallel. One approach utilizes trace information to improve traditional loop parallelization, and the other parallelizes traces instead of loop iterations. We also describe a cost/benefit model that makes intelligent parallelization decisions, as well as a parallel execution environment to execute parallelized programs. These techniques are based on Jikes RVM. Our approach is evaluated by parallelizing sequential Java programs, and its performance is compared to that of the manually parallelized code. According to the experimental results, our approach has low overheads and achieves competitive speedups compared to the manually parallelizing code. Moreover, trace parallelization can exploit parallelism beyond loop iterations.

Machine-to-machine communications for healthcare is emerging for the benefit of humans. In addition to novel medium access, we provide a systematic view to look for ways to develop technology to accomplish this goal, and a thorough vision toward effective system and network design.

In this paper, we focus on efficient construction of tightest matched subtree (TMSubtree) results, for keyword queries on extensible markup language (XML) data, based on smallest lowest common ancestor (SLCA) semantics. Here, "matched" means that all nodes in a returned subtree satisfy the constraint that the set of distinct keywords of the subtree rooted at each node is not subsumed by that of any of its sibling nodes, while "tightest" means that no two subtrees rooted at two sibling nodes can contain the same set of keywords. Assume that d is the depth of a given TMSubtree, m is the number of keywords of a given query Q. We proved that if d ${\leq}$ m, a matched subtree result has at most 2m! nodes; otherwise, the size of a matched subtree result is bounded by (d - m + 2)m!. Based on this theoretical result, we propose a pipelined algorithm to construct TMSubtree results without rescanning all node labels. Experiments verify the benefits of our algorithm in aiding keyword search over XML data.

Computational techniques for topic classification can support qualitative research by automatically applying labels in preparation for qualitative analyses. This paper presents an evaluation of supervised learning techniques applied to one such use case, namely, that of labeling emotions, instructions and information in suicide notes. We train a collection of one-versus-all binary support vector machine classifiers, using cost-sensitive learning to deal with class imbalance. The features investigated range from a simple bag-of-words and n-grams over stems, to information drawn from syntactic dependency analysis and WordNet synonym sets. The experimental results are complemented by an analysis of systematic errors in both the output of our system and the gold-standard annotations.

We present a methodology that automatically selects indexing algorithms for each heading in Medical Subject Headings (MeSH), National Library of Medicine's vocabulary for indexing MEDLINE. While manually comparing indexing methods is manageable with a limited number of MeSH headings, a large number of them make automation of this selection desirable. Results show that this process can be automated, based on previously indexed MEDLINE citations. We find that AdaBoostM1 is better suited to index a group of MeSH hedings named Check Tags, and helps improve the micro F-measure from 0.5385 to 0.7157, and the macro F-measure from 0.4123 to 0.5387 (both p < 0.01).

Query reformulation is an interactive process of revising user queries according to the query results. To assist biomedical researchers in this process, we present novel methods for automatically generating query reformulation suggestions. While previous work on query reformulation focused on addition of words to user queries, our method can deal with three types of query reformulation (i.e., addition, removal and replacement). The accuracy of the method for the addition type is ten times better than PubMed's "Also try", while the execution time is short enough for practical use.

The search for relevant and actionable information is a key to achieving clinical and research goals in biomedicine. Biomedical information exists in different forms: as text and illustrations in journal articles and other documents, in images stored in databases, and as patients' cases in electronic health records. This paper presents ways to move beyond conventional text-based searching of these resources, by combining text and visual features in search queries and document representation. A combination of techniques and tools from the fields of natural language processing, information retrieval, and content-based image retrieval allows the development of building blocks for advanced information services. Such services enable searching by textual as well as visual queries, and retrieving documents enriched by relevant images, charts, and other illustrations from the journal literature, patient records and image databases.