Symbolic connectionism in natural language disambiguation

Natural language understanding involves the simultaneous consideration of a large number of different sources of information. Traditional methods employed in language analysis have focused on developing powerful formalisms to represent syntactic or semantic structures along with rules for transforming language into these formalisms. However, they make use of only small subsets of knowledge. This article describes how to use the whole range of information through a neurosymbolic architecture which is a hybridization of a symbolic network and subsymbol vectors generated from a connectionist network. Besides initializing the symbolic network with prior knowledge, the subsymbol vectors are used to enhance the system's capability in disambiguation and provide flexibility in sentence understanding. The model captures a diversity of information including word associations, syntactic restrictions, case-role expectations, semantic rules and context. It attains highly interactive processing by representing knowledge in an associative network on which actual semantic inferences are performed. An integrated use of previously analyzed sentences in understanding is another important feature of our model. The model dynamically selects one hypothesis among multiple hypotheses. This notion is supported by three simulations which show the degree of disambiguation relies both on the amount of linguistic rules and the semantic-associative information available to support the inference processes in natural language understanding. Unlike many similar systems, our hybrid system is more sophisticated in tackling language disambiguation problems by using linguistic clues from disparate sources as well as modeling context effects into the sentence analysis. It is potentially more powerful than any systems relying on one processing paradigm     

A general compiler framework for speculative multithreaded processors

Speculative multithreading (SpMT) promises to be an effective mechanism for parallelizing nonnumeric programs, which tend to have irregular and pointer-intensive data structures and complex flows of control. Proper thread formation is crucial for obtaining good speedup in an SpMT system. This paper presents a compiler framework for partitioning a sequential program into multiple threads for parallel execution in an SpMT system. This framework is very general and supports speculative threads, nonspeculative threads, loop-centric threads, and out-of-order thread spawning. It is therefore useful for compiling for a wide variety of SpMT architectures. For effective partitioning of programs, the compiler uses profiling, interprocedural pointer analysis, data dependence information, and control dependence information. The compiler is implemented on the SUIF-MachSUIF platform. A simulation-based evaluation of the generated threads shows that the use of nonspeculative threads and nonloop speculative threads provides a significant increase in speedup for nonnumeric programs.     

Solution of a Well-Field Design Problem with Implicit Filtering

Problems involving the management of groundwater resources occur routinely, and management decisions based upon optimization approaches offer the potential to save substantial amounts of money. However, this class of application is notoriously difficult to solve due to non-convex objective functions with multiple local minima and both nonlinear models and nonlinear constraints. We solve a subset of community test problems from this application field using MODFLOW, a standard groundwater flow model, and IFFCO, an implicit filtering algorithm that was designed to solve problems similar to those of focus in this work. While sampling methods have received only scant attention in the groundwater optimization literature, we show encouraging results that suggest they are deserving of more widespread consideration for this class of problems. In keeping with our objectives for the community problems, we have packaged the approaches used in this work to facilitate additional work on these problems by others and the application of implicit filtering to other problems in this field. We provide the data for our formulation and solution on the web.     

Overt and Relational Aggression and Perceived Popularity: Developmental Differences in Concurrent and Prospective Relations.

Relations of overt and relational aggression with perceived popularity among children and early adolescents were examined in 2 studies (Ns = 607 and 1,049). Among older youths, positive concurrent relations found between overt aggression and perceived popularity became nonsignificant when relational aggression was controlled, whereas positive associations found between relational aggression and perceived popularity held when overt aggression was controlled. Aggression and perceived popularity were not positively related for the younger participants. The 2nd study also examined the temporal ordering of these relations over 6 months. For older girls, positive relations between relational aggression and perceived popularity were bidirectional. For older boys, relational aggression did not predict increased perceived popularity, but perceived popularity predicted increased relational aggression. Implications for intervention are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The performance characteristics of a novel multi-electrode electromagnetic flowmeter

The design of a novel multi-electrode electromagnetic flowmeter, which provides nine conventional electromagnetic flowmeters housed in a single body, has been used in adverse flow conditions immediately downstream from a header tank. The results show that careful initial calibration, in accord with the pipeline conditions recommended in international standards, enables errors of under 1% to be achieved. The results show that the two electrode planes closest to the header provide results with the greatest errors. The performance with the electrodes mounted in the vertical plane is generally better than that achieved with the electrodes mounted in the horizontal plane. Pressure measurements together with laser Doppler anemometry (LDA) and particle image velocimetry (PIV) techniques have been used to provide details of the magnitude of the disturbance caused to the flow field by the header and outflow combination.     

Pharmacogenetics Biomarkers and Their Specific Role in Neoadjuvant Chemoradiotherapy Treatments: An Exploratory Study on Rectal Cancer Patients

Background: Pathological complete response (pCR) to neoadjuvant chemoradiotherapy (CRT) in locally advanced rectal cancer (LARC) is still ascribed to a minority of patients. A pathway based-approach could highlight the predictive role of germline single nucleotide polymorphisms (SNPs). The primary aim of this study was to define new predictive biomarkers considering treatment specificities. Secondary aim was to determine new potential predictive biomarkers independent from radiotherapy (RT) dosage and cotreatment with oxaliplatin. Methods: Thirty germ-line SNPs in twenty-one genes were selected according to a pathway-based approach. Genetic analyses were performed on 280 LARC patients who underwent fluoropyrimidine-based CRT. The potential predictive role of these SNPs in determining pathological tumor response was tested in Group 1 (94 patients undergoing also oxaliplatin), Group 2 (73 patients treated with high RT dosage), Group 3 (113 patients treated with standard RT dosage), and in the pooled population (280 patients). Results: Nine new predictive biomarkers were identified in the three groups. The most promising one was rs3136228-MSH6 (p = 0.004) arising from Group 3. In the pooled population, rs1801133-MTHFR showed only a trend (p = 0.073). Conclusion: This exploratory study highlighted new potential predictive biomarkers of neoadjuvant CRT and underlined the importance to strictly define treatment peculiarities in pharmacogenetic analyses.     

Availability-adjusted X-factor

The notion of extended X-factor contribution has been proposed and evaluated recently as a tool to identify system capacity constraints based on machine group utilization and raw processing time. This provides a great convenience for factory-floor managers examining system capacity issues by using only local- or machine-level information. The purpose of this study is twofold: to study fundamental properties of the extended X-factor contribution measure augmented with availability, and to use this new measure to investigate resource allocation for optimizing mean cycle time. The availability-adjusted X-factor was first introduced in the context of local cycle time and its relation to allocation of personnel in individual machine groups. The availability-adjusted X-factor contribution measure developed and evaluated in the current study differs from the previous measure by its ability to identify capacity-constraining machines in the entire system. The measure presented herein is a more accurate indicator of capacity constraints than the extended X-factor contribution measures. With an objective to minimize cycle time or maximize throughput by properly allocating available resources, the results presented herein clearly demonstrate the effectiveness of this new measure to identify capacity-constraining machines. This study also uses availability as a decision variable for mean cycle time optimization.