Search

Testing is a popular way to assess one's competence in a language. The assessment can be conducted by the students for self evaluation or by the teachers in achievement tests. We present two applications of lexical information for assisting the task of test item authoring in this paper. Applying information implicitly contained in a machine readable lexicon, our system offers semantically and lexically similar words to help teachers prepare test items for cloze tests. Employing information about structures and pronunciations of Chinese characters, our system provides characters that are similar in either formation or pronunciation for the task of word correction. Experimental results indicate that our system furnishes quality recommendations for the preparation of test items, in addition to expediting the process.

A collection of 3208 reported errors of Chinese words were analyzed. Among which, 7.2% involved rarely used character, and 98.4% were assigned common classifications of their causes by human subjects. In particular, 80% of the errors observed in writings of middle school students were related to the pronunciations and 30% were related to the compositions of words. Experimental results show that using intuitive Web-based statistics helped us capture only about 75% of these errors. In a related task, the Web-based statistics are useful for recommending incorrect characters for composing test items for "incorrect character identification" tests about 93% of the time.

Introduction: Interleukin 17 (IL-17), correlating with advanced stage or poor prognosis of cancer patients, appears to contribute to tumor progression. Here we demonstrated the presence of IL-17 receptor A (IL-17RA) on the tumor cells directly promoting tumor growth, angiogenesis and metastasis. Methods: B16F10 and its derived IL-17RA deficient melanoma cells were stimulated with recombinant IL-17 (γmIL-17) or subcutaneously implanted into C57BL/6 mice. Cell proliferation, cell invasion, tumor growth, metastasis and angiogensis were monitored and anti-tumor immunity was also assessed. Results: As compared with those in the control B16F10 cells, the γmIL-17 inducing cell proliferation, migration and invasion were significantly reduced in the derived IL-17RA-deficient cells. The tumor growth and metastasis were also apparently reduced in the mice bearing the IL-17RA deficient tumors. In fact, not ony the expression of MMP2 and MMP9 but also the production of VEGF and HGF were reduced both in vitro and in vivo. Additionally, the increased levels of IFNγ+ and Granzyme B+ immune cells as well as the reduced numbers of Tregs and MDSCs appeared in the mice bearing the IL-17RA deficient tumors, indicating the lack of IL-17 enhancing the CTL activity and other anti-tumor immunity. Conclusion: These findings support the hypothesis that IL-17/IL-17RA signaling directly promotes tumor growth as well as recruits MDSCs and Treg cells, resulting in tumor progression. It suggests IL-17RA on tumors as a potential therapeutic target for cancer therapy. Citation Format: Ya-Shan Chen, Chao-Lin Liu, Hui-Shan Chen, Su-Ting Wu, Chia-Rui Shen. Decreased tumor progression by the knockdown of interleukin 17 receoptor A in the B16F10 melanoma model. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 442. doi:10.1158/1538-7445.AM2015-442

Financial distresses in corporations harm both individual investors and financial institutions, and can cause social problems due to the cascading effects. Government agencies, fund managers, and even small investors need to arm themselves with devices for predicting financial distresses in corporations, even when the financial problems are shadowed by malignant window dressing. In previous work, we explored the effectiveness of making predictions based on both financial ratios, including those proposed by Altman for the Z-score models and those used in the common-size analysis, and qualitative indicators, such as corporate governance. In this paper, we report results of our attempt to select the best features from the previously proposed features with genetic algorithms and gain ratio-based methods. Experimental results indicate that the selected features outperform the features used in the Z-score models. Not surprisingly, the genetic algorithms surpass the gain ratio-based methods in the task of feature selection.

This chapter purveys an account of Bayesian networks-related technologies for modeling students in intelligent tutoring systems. Uncertainty exists ubiquitously when we infer students’ internal status, for example, learning needs and emotion, from their external behavior, for example, responses to test items and explorative actions. Bayesian networks offer a mathematically sound mechanism for representing and reasoning about students under uncertainty. This chapter consists of five sections, and commences with a brief overview of intelligent tutoring systems, emphasizing the needs for uncertain reasoning. A succinct survey of Bayesian networks for student modeling is provided in Bayesian Networks, and we go through an example of applying Bayesian networks and mutual IDEA GROUP PUBLISHING This paper appears in the publication, Cognitively Informed Systems: Utilizing Practical Approaches to Enrich Information Presentation and Transfer edited by Eshaa M. Alkhalifa © 2006, Idea Group Inc. 701 E. Chocolate Avenue, Suite 200, Hershey PA 17033-1240, USA Tel: 717/533-8845; Fax 717/533-8661; URL-http://www.idea-group.com ITB12098

Phrases provide a better foundation for indexing and retrieving documents than individual words. Constituents of phrases make other component words in the phrase less ambiguous than when the words appear separately. Intuitively, classifiers that employ phrases for indexing should perform better than those that use words. Although pioneers have explored the possibility of indexing English documents decades ago, there are relatively fewer similar attempts for Chinese documents, partially because segmenting Chinese text into words correctly is not easy already. We build a domain dependent word list with the help of Chien's PAT tree-based method and HowNet, and use the resulting word list for defining relevant phrases for classifying Chinese judicial documents. Experimental results indicate that using phrases for indexing indeed allows us to classify judicial documents that are closely similar to each other. With a relatively more efficient algorithm, our classifier offers better performances than those reported in related works.

We take advantages of abundant text resource on the Internet and information about English phonetics for assisting human teachers to prepare test items for listening and dictation in English. In this preliminary exploration, we built an environment in which teachers choose words that they want to have test items for, and teachers compose the final test items based on the test items that are algorithmically generated by our system. The output of the current system indicates that computers can play active roles in assisting the composition of test items, though we have not done a field test over the usability issues.

Along with the increasing accessibility of the Web, services available on the Web have expanded into several aspects of our lives. As a relatively new comer, Web-based legal services have become another front in the arena. For instance, through such services, people obtain rough pictures of the lawsuits of interest before they decide whether and how to approach their counselors. As a step toward offering online legal services, we propose algorithms for classifying criminal charges in Chinese based on the alleged criminal violations that the district attorney used to prosecute the defendants. The classification is not trivial because some prosecutions involve violations of multiple articles in the criminal law, and we wish to classify lawsuits as precisely as possible. We employ techniques of instance-based classification and introspective learning for the classification task, and achieve satisfactory initial results for classifying charges of larceny and gambling.

We report experience in applying techniques for natural language processing to algorithmically generating test items for both reading and listening cloze items. We propose a word sense disambiguation-based method for locating sentences in which designated words carry specific senses, and apply a collocation-based method for selecting distractors that are necessary for multiple-choice cloze items. Experimental results indicate that our system was able to produce a usable item for every 1.6 items it returned. We also attempt to measure distance between sounds of words by considering phonetic features of the words. With the help of voice synthesizers, we were able to assist the task of composing listening cloze items. By providing both reading and listening cloze items, we would like to offer a somewhat adaptive system for assisting Taiwanese children in learning English vocabulary.

The author examines methods for a special class of path planning problems in which the routes are constrained. General search algorithms assume that we can move around in the traffic network freely, so they extend the partial paths from the very last location to each of its neighbors to form more partial paths. The best partial paths are then selected to expand, unless die selected partial path happens to he a solution. Without proper guidance, this strategy way lead to inefficient planning algorithms when the way one can move around in the networks Is constrained. This scenario could happen in public transportation system where passengers cannot order drivers to change the routes of public buses to meet individual travel needs. A few recently proposed path-planning algorithms for public transportation systems capture the route constraints by matrices. Although they work for some applications, they are not perfect for cooperating with traditional algorithms; for best-path planning. Applying special properties of matrix multiplication, the author also employs matrices for capturing the route constraints. The author improves previous designs, and comes up with the so-called Q matrices that serve well in the A* algorithm for best-path planning under route constraints.

We consider stochastic networks in which link travel times are dependent, discrete random variables. We present methods for computing bounds on path travel times using stochastic dominance relationships among link travel times, and discuss techniques for controlling tightness of the bounds. We apply these methods to shortest-path problems, show that the proposed algorithm can provide bounds on the recommended path, and elaborate on extensions of the algorithm for demonstrating the anytime property.

We investigate the effectiveness of machine-generated criteria for classification problems related to criminal summary judgments. Our system utilizes documents of closed lawsuits as training data for generating keyword-based and case-based classification criteria, and applies these machine-generated criteria for the classification tasks. To construct databases of the classification criteria, we employ different levels of lexical knowledge in extracting information from legal documents in Chinese, and build a case instance for each closed lawsuit. Experimental results indicate that case-based classification outperforms keyword-based classification, and that machine-generated cases may offer performance accuracy that is about 7% below that of human-provided cases. Hoping to boost inference efficiency of our classifiers, we also design methods that merge the machine-generated criteria. Empirical results show that our methods can maintain the classification quality within 20% of the quality achieved by human-provided cases, even when we aggressively reduce the number of previously machine-generated cases by about seventy percents.

Computing travel plans for desired trips in public transportation systems is not exactly the same as finding a shortest driving path in a given area. Path planning in the context of public transportation systems must consider the route constraint that public vehicles serve on particular paths and that passengers cannot order the drivers to change the bus routes. Explicit representation of the route constraint helps us to design efficient algorithms that focus on viable routes for computing travel plans of interest. This paper presents two strategies for capturing the route constraint. The first strategy employs connectivity matrices, and applies special properties of matrices for quickly identifying feasible travel plans for the desired trips. The second strategy uses hubs where many service routes concentrate for computing travel plans. Our algorithms perform very well in field tests.

Agglutinin protein purified from the seeds of Abrus precatorius has a high antitumour activity and was crystallized at room temperature with polyethylene glycol 8000 as the precipitant. The agglutinin crystal diffracted to 3.45 A and belongs to one of two possible tetragonal space groups, P4(1)2(1)2 or P4(3)2(1)2, with unit-cell parameters a = b = 141.91, c = 105.63 A. The asymmetric unit contains a heterotetrameric protein molecule of molecular weight 134 kDa and has a solvent content of approximately 38%.

Penicillins are effective for treating bacterial infections in many clinical situations, but in certain cases, they cause allergic reactions varying from mild skin rashes to anaphylactic shock and death. These symptoms often resemble those seen in immediate or delayed hypersensitivity reactions, and may be associated with drug specific IgE and/or drug specific T cells. Currently, the laboratory tests include isotope based serum IgE antibody measurement as well as skin test. However, radioisotope allergen specific tests (RAST) are not convenient, and the standardized skin-testing reagents and protocol do not usually exist. In fact, the skin test is often avoided because of the high risks. Therefore, in the current study, we have developed a non-isotope based penicillin-reactive IgE screening platform with high sensitivity and specificity. Using this screening platform to detect IgE of a mouse model sensitized with penicillin reveals that high levels of penicillin specific IgE were found in the penicillin sensitized animals but not in OVA sensitized or untreated control animals. Penicillin specific T cells produce relatively higher levels of IL-4 and IL-5 in the penicillin sensitized group, indicating a Th2-cytokine dominated response of T-cell dependent IgE production against penicillin. This observation and the development of such convenient platform to measure penicillin specific IgE may benefit further study of penicillin allergy.

We examine methods for a special class of path-planning problems in which the routes are constrained. The scenario could happen, for instance, in transit systems where passengers cannot order drivers to change the routes of public buses to meet individual travel needs. General search algorithms are applicable to this class of problems, but may not find the desired solution as efficiently as possible. This paper reports three different strategies that capture the route constraints for improving efficiency of path planning algorithms. The first strategy applies hierarchical planning, and the rest employs matrices for encoding route constraints. We propose and prove that the Q matrix is instrumental for capturing route constraints and measuring quality of service of the transportation network. Moreover, we discuss how we may apply the Q matrix in designing admissible heuristic functions that are crucial for applying the A* algorithm for best-path planning under route constraints.

Targeted inhibition of multidrug ABCG2 transporter is believed to improve cancer therapeutics. However, the consequences of ABCG2 inhibition have not been systematically evaluated since ABCG2 is expressed in several organs including the liver. Here, we demonstrate that ABCG2-deficient hepatocytes have increased amounts of fragmental mitochondria accompanied by disruption of mitochondrial dynamics and functions. This disruption was due to ABCG2 knockout elevating intracellular protoporphyrin IX, which led to upregulation of DRP-1-mediated mitochondrial fission. The finding that ABCG2 deficiency can generate dysfunctional mitochondria in hepatocytes raises concerns regarding the systematic use of ABCG2 inhibitor in cancer patients.

We describe technology for robust traffic monitoring and automated vehicle control using decision-theoretic and probabilistic reasoning methods. In this work, we have designed and implemented probabilistic models for deriving high-level descriptions of traffic conditions, as well as the maneuvers and intentions of individual vehicles, from visual observation of a traffic scene. Enhancements to standard probabilistic modeling and inference techniques have improved the performance of uncertain reasoning over time with continuous variables. We have demonstrated our models and algorithms in real-time analysis of traffic images as well as control of simulated vehicles.