Computer Science

We present our preliminary work on extracting fine-grained user opinions from game review texts. In sentiment analysis, user-generated texts such as blogs, comments and reviews are usually represented by the words which appeared in the texts. However for complex multi-faceted objects such as games, single words are not sufficient to represent opinions on individual aspects of the object. We propose to represent such an object by pairs of aspect and its quality/value, for example "greatgraphics". We used a large adjective-context co-occurrence matrix extracted from user reviews posted at a game site, and applied co-clustering to reduce the dimensions of the matrix. The derived co-clusters are pairs of row clusters × column clusters. By examining the derived co-clusters, we were able to discover the aspects and their qualities which the users care about strongly in games.

New Zealand Obese (NZO) male mice develop a polygenic juvenile-onset obesity and maturity-onset hyperinsulinemia and hyperglycemia (diabesity). Here we report on metabolic and molecular changes associated with the antidiabesity action of CL316,243 (CL), a ␤ 3 -adrenergic receptor agonist. Dietary CL treatment initiated at weaning reduced the peripubertal rise in body weight and adiposity while promoting growth without suppressing hyperphagia. The changes in adiposity, in turn, suppressed development of hyperinsulinemia, hyperleptinemia, hyperlipidemia, and hyperglycemia. These CL-induced alterations were reflected by decreased adipose tissue mass, increased expression of transcripts for uncoupling protein-1 (UCP-1), peroxisome proliferator-activated receptor alpha (PPAR␣), peroxisome proliferater-activated receptor coactivator-1 (PGC-1), and robust development of brown adipocyte function in white fat. Increased drug-mediated energy dissipation elicited a 1.5°C increase in whole body temperature under conditions of increased food intake but with no change in physical activity. Indirect calorimetry of mice treated with CL showed both increased energy expenditure and a restoration of a prominent diurnal pattern in the respiratory exchange ratio suggesting improved nutrient sensing. Our data suggest that CL promotes increased energy dissipation in white and brown fat depots by augmenting thermogenesis and by metabolic re-partitioning of energy in a diabesity-protective fashion. This is the first report demonstrating the effects of dietary ␤ 3 -agonist in preventing the onset of diabesity in a polygenic rodent model of type 2 diabetes.

New Zealand Obese (NZO) male mice develop a polygenic juvenile-onset obesity and maturity-onset hyperinsulinemia and hyperglycemia (diabesity). Here we report on metabolic and molecular changes associated with the antidiabesity action of CL316,243 (CL), a ␤ 3 -adrenergic receptor agonist. Dietary CL treatment initiated at weaning reduced the peripubertal rise in body weight and adiposity while promoting growth without suppressing hyperphagia. The changes in adiposity, in turn, suppressed development of hyperinsulinemia, hyperleptinemia, hyperlipidemia, and hyperglycemia. These CL-induced alterations were reflected by decreased adipose tissue mass, increased expression of transcripts for uncoupling protein-1 (UCP-1), peroxisome proliferator-activated receptor alpha (PPAR␣), peroxisome proliferater-activated receptor coactivator-1 (PGC-1), and robust development of brown adipocyte function in white fat. Increased drug-mediated energy dissipation elicited a 1.5°C increase in whole body temperature under conditions of increased food intake but with no change in physical activity. Indirect calorimetry of mice treated with CL showed both increased energy expenditure and a restoration of a prominent diurnal pattern in the respiratory exchange ratio suggesting improved nutrient sensing. Our data suggest that CL promotes increased energy dissipation in white and brown fat depots by augmenting thermogenesis and by metabolic re-partitioning of energy in a diabesity-protective fashion. This is the first report demonstrating the effects of dietary ␤ 3 -agonist in preventing the onset of diabesity in a polygenic rodent model of type 2 diabetes.

GADD45 has been suggested to coordinate cell cycle regulation with the repair of DNA damage following ionizing radiation (IR). Although the GADD45 gene is transcriptionally up-regulated in response to IR, alterations in in vivo transcription factor (TF) binding or chromatin structure associated with up-regulation have not been defined. To understand how chromatin structure might influence TF binding and GADD45 up-regulation, key regulatory regions of the gene were identified by in vivo DNase I hypersensitivity (HS) analysis. Chromatin structure and in vivo TF binding in these regions were subsequently monitored in both non-irradiated and irradiated human ML-1 cells. In non-irradiated cells expressing basal levels of GADD45, the gene exhibited a highly organized chromatin structure with distinctly positioned nucleosomes. Also identified in non-irradiated cells were DNA-protein interactions at octamer binding motifs and a CCAAT box in the promoter and at consensus binding sites for AP-1 and p53 within intron 3. Upon irradiation and a subsequent 15-fold increase in GADD45 mRNA levels, neither the chromatin structure nor the pattern of TF binding in key regulatory regions was altered. These results suggest that the GADD45 gene is poised for upregulation and can be rapidly induced independent of gross changes in chromatin structure or TF binding.

Tumor-associated aberrant silencing of CpG island-containing genes has been correlated with increased cytosine methylation, a "closed" chromatin structure, and exclusion of transcription factor binding in the CpG island/promoter regions of affected genes. Given the lack of understanding of what constitutes a closed chromatin structure in CpG islands, however, it has been difficult to assess the relationship among cytosine methylation, chromatin structure, and inappropriate gene silencing. In this study, nuclease accessibility analysis was used to more clearly define the chromatin structure in the CpG island of the human O 6 -methylguanine DNA methyltransferase (MGMT) gene. Chromatin structure was then related to in vivo DNA-protein interactions and cytosine methylation status of the MGMT CpG island in human glioma cells varying in MGMT expression. The results of these studies indicated that the "open" chromatin structure associated with the MGMT CpG island in MGMT ؉ cells consisted of an approximately 250-bp transcription factor-binding, nuclease-accessible, nucleosome-free region of DNA, whose formation was associated with at least four flanking, precisely positioned nucleosome-like structures. In MGMT ؊ cells, this precise nucleosomal array was lost and was replaced by randomly positioned nucleosomes (i.e., the closed chromatin structure), regardless of whether methylation of the CpG island was spread over the entire island or limited to regions outside the transcription factor binding region. These results suggest that CpG islands facilitate the expression of housekeeping genes by facilitating nucleosomal positioning and that the conditions that alter the formation of this array (such as perhaps methylation) may indirectly affect CpG island-containing gene expression.

War features prominently in history, that is, in the records of literate societies. This paper addresses the issue of prehistoric warfare, that is, whether human societies without writing fight wars, and if so how often, under what circumstances, what tactics and technology they use, and how their warfare differs from that of literate societies. As background, the paper reviews theoretical perspectives on prehistoric warfare and considers categories of data and problems with their interpretation in the absence of primary documents. In particular, researchers continue to invoke the perspectives of the philosophers Thomas Hobbes and Jean-Jacques Rousseau as touchstones of their thoughts on the topic of prehistoric warfare. The author has no particular training or expertise in anthropology or archeology, but from reviewing the literature suggests that available data supports neither the Hobbesian nor the Rousseauian extreme. Societies without writing sometimes fight and sometimes do no...

Simulation-based training (SBT) programs are commonly employed by organizations to train individuals and teams for e ective workplace cognitive and psychomotor skills in a broad range of applications. Distributed cognition has become a popular cognitive framework for the design and evaluation of these SBT environments, with structured methodologies such as Distributed Cognition for Teamwork (DiCoT) used for analysis. However, the analysis and evaluations generated by such distributed cognition frameworks require extensive domain-knowledge and manual coding and interpretation, and the analysis is primarily qualitative. In this work, we propose and develop the application of multimodal learning analysis techniques to SBT scenarios. Using these analysis methods, we can use the rich multimodal data collected in SBT environments to generate more automated interpretations of trainee performance that supplement and extend traditional DiCoT analysis. To demonstrate the use of these methods, we present a case study of nurses training in a mixed-reality manikin-based (MRMB) training environment. We show how the combined analysis of the video, speech, and eye-tracking data collected as the nurses train in the MRMB environment supports and enhances traditional qualitative DiCoT analysis. By applying such quantitative data-driven analysis methods, we can better analyze trainee activities online in SBT and MRMB environments. With continued development, these analysis methods could be used to provide targeted feedback to learners, a detailed review of training performance to the instructors, and data-driven evidence for improving the environment to simulation designers.

Figure 1: A snapshot of the temporal fully-weighted graphs of a simulation nurse triad taken at ≃ 760 seconds, indicated by the red vertical line. At the top of the figure, the participating nurses (P1 to P3) have their own graph built based on their gaze data that has been mapped to the mixed-reality nursing scenario by matching gaze to relevant objects. These graphs model gaze evolution and decay by altering the nodes' and edges' weight over time. Node weight represents accumulated fixation duration and edge weight indicates the accumulated number of transitions between two objects. The nurses' gaze over the chosen time interval, presented at the bottom of the figure, illustrates the temporal change in gaze fixation for the three nurses across the objects.

Simulation-based training (SBT) programs are commonly employed by organizations to train individuals and teams for e ective workplace cognitive and psychomotor skills in a broad range of applications. Distributed cognition has become a popular cognitive framework for the design and evaluation of these SBT environments, with structured methodologies such as Distributed Cognition for Teamwork (DiCoT) used for analysis. However, the analysis and evaluations generated by such distributed cognition frameworks require extensive domain-knowledge and manual coding and interpretation, and the analysis is primarily qualitative. In this work, we propose and develop the application of multimodal learning analysis techniques to SBT scenarios. Using these analysis methods, we can use the rich multimodal data collected in SBT environments to generate more automated interpretations of trainee performance that supplement and extend traditional DiCoT analysis. To demonstrate the use of these methods, we present a case study of nurses training in a mixed-reality manikin-based (MRMB) training environment. We show how the combined analysis of the video, speech, and eye-tracking data collected as the nurses train in the MRMB environment supports and enhances traditional qualitative DiCoT analysis. By applying such quantitative data-driven analysis methods, we can better analyze trainee activities online in SBT and MRMB environments. With continued development, these analysis methods could be used to provide targeted feedback to learners, a detailed review of training performance to the instructors, and data-driven evidence for improving the environment to simulation designers.

Bidirectional Encoder Representations from Transformers (BERT) [Devlin et al., 2018] has been shown to be effective at modeling a multitude of datasets across a wide variety of Natural Language Processing (NLP) tasks; however, little research has been done regarding BERT's effectiveness at modeling domain-specific datasets. Specifically, scientific and medical datasets present a particularly difficult challenge in NLP, as these types of corpora are often rife with technical jargon that is largely absent from the canonical corpora that BERT and other transfer learning models were originally trained on. This thesis is a Systematic Literature Review (SLR) of twenty-seven studies that were selected to address the various methods of implementation when applying BERT to scientific and medical datasets. These studies show that despite the datasets' esoteric subject matter, BERT can be effective at a wide range of tasks when applied to domain-specific datasets. Furthermore, these studies show that the addition of domain-specific pretraining, either through additional pretraining or the utilization of domain-specific BERT derivatives such as BioBERT [Lee et al., 2019], can further augment BERT's performance on scientific and medical texts.

Figure 1: A snapshot of the temporal fully-weighted graphs of a simulation nurse triad taken at ≃ 760 seconds, indicated by the red vertical line. At the top of the figure, the participating nurses (P1 to P3) have their own graph built based on their gaze data that has been mapped to the mixed-reality nursing scenario by matching gaze to relevant objects. These graphs model gaze evolution and decay by altering the nodes' and edges' weight over time. Node weight represents accumulated fixation duration and edge weight indicates the accumulated number of transitions between two objects. The nurses' gaze over the chosen time interval, presented at the bottom of the figure, illustrates the temporal change in gaze fixation for the three nurses across the objects.

This paper explores the use of large language models (LLMs) to score and explain short-answer assessments in K-12 science. While existing methods can score more structured math and computer science assessments, they often do not provide explanations for the scores. Our study focuses on employing GPT-4 for automated assessment in middle school Earth Science, combining few-shot and active learning with chainof-thought reasoning. Using a human-in-the-loop approach, we successfully score and provide meaningful explanations for formative assessment responses. A systematic analysis of our method's pros and cons sheds light on the potential for human-in-the-loop techniques to enhance automated grading for open-ended science assessments.

This paper explores the use of large language models (LLMs) to score and explain short-answer assessments in K-12 science. While existing methods can score more structured math and computer science assessments, they often do not provide explanations for the scores. Our study focuses on employing GPT-4 for automated assessment in middle school Earth Science, combining few-shot and active learning with chainof-thought reasoning. Using a human-in-the-loop approach, we successfully score and provide meaningful explanations for formative assessment responses. A systematic analysis of our method's pros and cons sheds light on the potential for human-in-the-loop techniques to enhance automated grading for open-ended science assessments.