A Robust Parser of Czech

Mr. Kubon's project was inspired by the growing need for an automatic, syntactic analyser (parser) of Czech, which could be used in the syntactic processing of large amounts of texts. Mr. Kubon notes that such a tool would be very useful, especially in the field of corpus linguistics, where creating a large-scale "tree bank" (a collection of syntactic representations of natural language sentences) is a very important step towards the investigation of the properties of a given language. The work involved in syntactically parsing a whole corpus in order to get a representative set of syntactic structures would be almost inconceivable without the help of some kind of robust (semi)automatic parser. The need for the automatic natural language parser to be robust increases with the size of the linguistic data in the corpus or in any other kind of text which is going to be parsed. Practical experience shows that apart from syntactically correct sentences, there are many sentences which contain a "real" grammatical error. These sentences may be corrected in small-scale texts, but not generally in the whole corpus. In order to be able to complete the overall project, it was necessary to address a number of smaller problems. These were; 1. the adaptation of a suitable formalism able to describe the formal grammar of the system; 2. the definition of the structure of the system's dictionary containing all relevant lexico-syntactic information, and the development of a formal grammar able to robustly parse Czech sentences from the test suite; 3. filling the syntactic dictionary with sample data allowing the system to be tested and debugged during its development (about 1000 words); 4. the development of a set of sample sentences containing a reasonable amount of grammatical and ungrammatical phenomena covering some of the most typical syntactic constructions being used in Czech. Number 3, building a formal grammar, was the main task of the project. The grammar is of course far from complete (Mr. Kubon notes that it is debatable whether any formal grammar describing a natural language may ever be complete), but it covers the most frequent syntactic phenomena, allowing for the representation of a syntactic structure of simple clauses and also the structure of certain types of complex sentences. The stress was not so much on building a wide coverage grammar, but on the description and demonstration of a method. This method uses a similar approach as that of grammar-based grammar checking. The problem of reconstructing the "correct" form of the syntactic representation of a sentence is closely related to the problem of localisation and identification of syntactic errors. Without a precise knowledge of the nature and location of syntactic errors it is not possible to build a reliable estimation of a "correct" syntactic tree. The incremental way of building the grammar used in this project is also an important methodological issue. Experience from previous projects showed that building a grammar by creating a huge block of metarules is more complicated than the incremental method, which begins with the metarules covering most common syntactic phenomena first, and adds less important ones later, especially from the point of view of testing and debugging the grammar. The sample of the syntactic dictionary containing lexico-syntactical information (task 4) now has slightly more than 1000 lexical items representing all classes of words. During the creation of the dictionary it turned out that the task of assigning complete and correct lexico-syntactic information to verbs is a very complicated and time-consuming process which would itself be worth a separate project. The final task undertaken in this project was the development of a method allowing effective testing and debugging of the grammar during the process of its development. The problem of the consistency of new and modified rules of the formal grammar with the rules already existing is one of the crucial problems of every project aiming at the development of a large-scale formal grammar of a natural language. This method allows for the detection of any discrepancy or inconsistency of the grammar with respect to a test-bed of sentences containing all syntactic phenomena covered by the grammar. This is not only the first robust parser of Czech, but also one of the first robust parsers of a Slavic language. Since Slavic languages display a wide range of common features, it is reasonable to claim that this system may serve as a pattern for similar systems in other languages. To transfer the system into any other language it is only necessary to revise the grammar and to change the data contained in the dictionary (but not necessarily the structure of primary lexico-syntactic information). The formalism and methods used in this project can be used in other Slavic languages without substantial changes.

Towards an Emergent Semantic Web

In his in uential article about the evolution of the Web, Berners-Lee [1] envisions a Semantic Web in which humans and computers alike are capable of understanding and processing information. This vision is yet to materialize. The main obstacle for the Semantic Web vision is that in today's Web meaning is rooted most often not in formal semantics, but in natural language and, in the sense of semiology, emerges not before interpretation and processing. Yet, an automated form of interpretation and processing can be tackled by precisiating raw natural language. To do that, Web agents extract fuzzy grassroots ontologies through induction from existing Web content. Inductive fuzzy grassroots ontologies thus constitute organically evolved knowledge bases that resemble automated gradual thesauri, which allow precisiating natural language [2]. The Web agents' underlying dynamic, self-organizing, and best-effort induction, enable a sub-syntactical bottom up learning of semiotic associations. Thus, knowledge is induced from the users' natural use of language in mutual Web interactions, and stored in a gradual, thesauri-like lexical-world knowledge database as a top-level ontology, eventually allowing a form of computing with words [3]. Since when computing with words the objects of computation are words, phrases and propositions drawn from natural languages, it proves to be a practical notion to yield emergent semantics for the Semantic Web. In the end, an improved understanding by computers on the one hand should upgrade human- computer interaction on the Web, and, on the other hand allow an initial version of human- intelligence amplification through the Web.

When language of instruction and language of application differ: Cognitive costs of bilingual mathematics learning

Bilingual education programs implicitly assume that the acquired knowledge is represented in a language-independent way. This assumption, however, stands in strong contrast to research findings showing that information may be represented in a way closely tied to the specific language of instruction and learning. The present study aims to examine whether and to which extent cognitive costs appear during arithmetic learning when language of instruction and language of retrieving differ. Thirty-nine high school students participating in a bilingual education program underwent a four-day training on multiplication and subtraction problems in one language (German or French), followed by a test session in which they had to solve trained as well as untrained problems in both languages. We found that cognitive costs related to language switching appeared for both arithmetic operations. Implications of our findings are discussed with respect to bilingual education as well as to cognitive mechanisms underlying different arithmetic operations.

Willy Loman and the "Soul of a New Machine": Technology and the Common Man

As Death of a Salesman opens, Willy Loman returns home “tired to the death” (p. 13). Lost in reveries about the beautiful countryside and the past, he's been driving off the road; and now he wants a cheese sandwich. But Linda's suggestion that he try a new American-type cheese — “It's whipped” (p. 16) — irritates Willy: “Why do you get American when I like Swiss?” (p. 17). His anger at being contradicted unleashes an indictment of modern industrialized America: The street is lined with cars. There's not a breath of fresh air in the neighborhood. The grass don't grow any more, you can't raise a carrot in the back yard. (p. 17). In the old days, “This time of year it was lilac and wisteria.” Now: “Smell the stink from that apartment house! And another one on the other side…” (pp. 17–18). But just as Willy defines the conflict between nature and industry, he pauses and simply wonders: “How can they whip cheese?” (p. 18). The clash between the old agrarian ideal and capitalistic enterprise is well documented in the literature on Death of a Salesman, as is the spiritual shift from Thomas Jefferson to Andrew Carnegie to Dale Carnegie that the play reflects. The son of a pioneer inventor and the slave to broken machines, Willy Loman seems to epitomize the victim of modern technology.

Mining free-text medical records for companion animal enteric syndrome surveillance.

Large amounts of animal health care data are present in veterinary electronic medical records (EMR) and they present an opportunity for companion animal disease surveillance. Veterinary patient records are largely in free-text without clinical coding or fixed vocabulary. Text-mining, a computer and information technology application, is needed to identify cases of interest and to add structure to the otherwise unstructured data. In this study EMR's were extracted from veterinary management programs of 12 participating veterinary practices and stored in a data warehouse. Using commercially available text-mining software (WordStat™), we developed a categorization dictionary that could be used to automatically classify and extract enteric syndrome cases from the warehoused electronic medical records. The diagnostic accuracy of the text-miner for retrieving cases of enteric syndrome was measured against human reviewers who independently categorized a random sample of 2500 cases as enteric syndrome positive or negative. Compared to the reviewers, the text-miner retrieved cases with enteric signs with a sensitivity of 87.6% (95%CI, 80.4-92.9%) and a specificity of 99.3% (95%CI, 98.9-99.6%). Automatic and accurate detection of enteric syndrome cases provides an opportunity for community surveillance of enteric pathogens in companion animals.

Data Accuracy in Medical Record Abstraction

Clinical Research Data Quality Literature Review and Pooled Analysis We present a literature review and secondary analysis of data accuracy in clinical research and related secondary data uses. A total of 93 papers meeting our inclusion criteria were categorized according to the data processing methods. Quantitative data accuracy information was abstracted from the articles and pooled. Our analysis demonstrates that the accuracy associated with data processing methods varies widely, with error rates ranging from 2 errors per 10,000 files to 5019 errors per 10,000 fields. Medical record abstraction was associated with the highest error rates (70–5019 errors per 10,000 fields). Data entered and processed at healthcare facilities had comparable error rates to data processed at central data processing centers. Error rates for data processed with single entry in the presence of on-screen checks were comparable to double entered data. While data processing and cleaning methods may explain a significant amount of the variability in data accuracy, additional factors not resolvable here likely exist. Defining Data Quality for Clinical Research: A Concept Analysis Despite notable previous attempts by experts to define data quality, the concept remains ambiguous and subject to the vagaries of natural language. This current lack of clarity continues to hamper research related to data quality issues. We present a formal concept analysis of data quality, which builds on and synthesizes previously published work. We further posit that discipline-level specificity may be required to achieve the desired definitional clarity. To this end, we combine work from the clinical research domain with findings from the general data quality literature to produce a discipline-specific definition and operationalization for data quality in clinical research. While the results are helpful to clinical research, the methodology of concept analysis may be useful in other fields to clarify data quality attributes and to achieve operational definitions. Medical Record Abstractor’s Perceptions of Factors Impacting the Accuracy of Abstracted Data Medical record abstraction (MRA) is known to be a significant source of data errors in secondary data uses. Factors impacting the accuracy of abstracted data are not reported consistently in the literature. Two Delphi processes were conducted with experienced medical record abstractors to assess abstractor’s perceptions about the factors. The Delphi process identified 9 factors that were not found in the literature, and differed with the literature by 5 factors in the top 25%. The Delphi results refuted seven factors reported in the literature as impacting the quality of abstracted data. The results provide insight into and indicate content validity of a significant number of the factors reported in the literature. Further, the results indicate general consistency between the perceptions of clinical research medical record abstractors and registry and quality improvement abstractors. Distributed Cognition Artifacts on Clinical Research Data Collection Forms Medical record abstraction, a primary mode of data collection in secondary data use, is associated with high error rates. Distributed cognition in medical record abstraction has not been studied as a possible explanation for abstraction errors. We employed the theory of distributed representation and representational analysis to systematically evaluate cognitive demands in medical record abstraction and the extent of external cognitive support employed in a sample of clinical research data collection forms. We show that the cognitive load required for abstraction in 61% of the sampled data elements was high, exceedingly so in 9%. Further, the data collection forms did not support external cognition for the most complex data elements. High working memory demands are a possible explanation for the association of data errors with data elements requiring abstractor interpretation, comparison, mapping or calculation. The representational analysis used here can be used to identify data elements with high cognitive demands.

Two way clustering of Microarray Data using a Hybrid Approach

The Microarray technique is rather powerful, as it allows to test up thousands of genes at a time, but this produces an overwhelming set of data files containing huge amounts of data, which is quite difficult to pre-process, separate, classify and correlate for interesting conclusions to be extracted. Modern machine learning, data mining and clustering techniques based on information theory, are needed to read and interpret the information contents buried in those large data sets. Independent Component Analysis method can be used to correct the data affected by corruption processes or to filter the uncorrectable one and then clustering methods can group similar genes or classify samples. In this paper a hybrid approach is used to obtain a two way unsupervised clustering for a corrected microarray data.

A Satisfaction-based Model for Affect Recognition from Conversational Features in Spoken Dialog Systems

Detecting user affect automatically during real-time conversation is the main challenge towards our greater aim of infusing social intelligence into a natural-language mixed-initiative High-Fidelity (Hi-Fi) audio control spoken dialog agent. In recent years, studies on affect detection from voice have moved on to using realistic, non-acted data, which is subtler. However, it is more challenging to perceive subtler emotions and this is demonstrated in tasks such as labelling and machine prediction. This paper attempts to address part of this challenge by considering the role of user satisfaction ratings and also conversational/dialog features in discriminating contentment and frustration, two types of emotions that are known to be prevalent within spoken human-computer interaction. However, given the laboratory constraints, users might be positively biased when rating the system, indirectly making the reliability of the satisfaction data questionable. Machine learning experiments were conducted on two datasets, users and annotators, which were then compared in order to assess the reliability of these datasets. Our results indicated that standard classifiers were significantly more successful in discriminating the abovementioned emotions and their intensities (reflected by user satisfaction ratings) from annotator data than from user data. These results corroborated that: first, satisfaction data could be used directly as an alternative target variable to model affect, and that they could be predicted exclusively by dialog features. Second, these were only true when trying to predict the abovementioned emotions using annotator?s data, suggesting that user bias does exist in a laboratory-led evaluation.

An approach to automatic learning assessment based on the computational theory of perceptions

E-learning systems output a huge quantity of data on a learning process. However, it takes a lot of specialist human resources to manually process these data and generate an assessment report. Additionally, for formative assessment, the report should state the attainment level of the learning goals defined by the instructor. This paper describes the use of the granular linguistic model of a phenomenon (GLMP) to model the assessment of the learning process and implement the automated generation of an assessment report. GLMP is based on fuzzy logic and the computational theory of perceptions. This technique is useful for implementing complex assessment criteria using inference systems based on linguistic rules. Apart from the grade, the model also generates a detailed natural language progress report on the achieved proficiency level, based exclusively on the objective data gathered from correct and incorrect responses. This is illustrated by applying the model to the assessment of Dijkstra’s algorithm learning using a visual simulation-based graph algorithm learning environment, called GRAPHs

Design, development and field evaluation of a Spanish into sign language translation system

This paper describes the design, development and field evaluation of a machine translation system from Spanish to Spanish Sign Language (LSE: Lengua de Signos Española). The developed system focuses on helping Deaf people when they want to renew their Driver’s License. The system is made up of a speech recognizer (for decoding the spoken utterance into a word sequence), a natural language translator (for converting a word sequence into a sequence of signs belonging to the sign language), and a 3D avatar animation module (for playing back the signs). For the natural language translator, three technological approaches have been implemented and evaluated: an example-based strategy, a rule-based translation method and a statistical translator. For the final version, the implemented language translator combines all the alternatives into a hierarchical structure. This paper includes a detailed description of the field evaluation. This evaluation was carried out in the Local Traffic Office in Toledo involving real government employees and Deaf people. The evaluation includes objective measurements from the system and subjective information from questionnaires. The paper details the main problems found and a discussion on how to solve them (some of them specific for LSE).

Exploring cognition processes in second language acquisition: the case of cognates and false-friends in EST

This article explores one aspect of the processing perspective in L2 learning in an EST context: the processing of new content words, in English, of the type ‘cognates’ and ‘false friends’, by Spanish speaking engineering students. The paper does not try to offer a comprehensive overview of language acquisition mechanisms, but rather it is intended to review more narrowly how our conceptual systems, governed by intricately linked networks of neural connections in the brain, make language development possible, creating, at the same time, some L2 processing problems. The case of ‘cognates and false friends’ in specialised contexts is brought here to illustrate some of the processing problems that the L2 learner has to confront, and how mappings in the visual, phonological and semantic (conceptual) brain structures function in second language processing of new vocabulary. Resumen Este artículo pretende reflexionar sobre un aspecto de la perspectiva del procesamiento de segundas lenguas (L2) en el contexto del ICT: el procesamiento de palabras nuevas, en inglés, conocidas como “cognados” y “falsos amigos”, por parte de estudiantes de ingeniería españoles. No se pretende ofrecer una visión completa de los mecanismos de adquisición del lenguaje, más bien se intenta mostrar cómo nuestro sistema conceptual, gobernado por una complicada red de conexiones neuronales en el cerebro, hace posible el desarrollo del lenguaje, aunque ello conlleve ciertas dificultades en el procesamiento de segundas lenguas. El caso de los “cognados” y los “falsos amigos”, en los lenguajes de especialidad, se trae para ilustrar algunos de los problemas de procesamiento que el estudiante de una lengua extranjera tiene que afrontar y el funcionamiento de las correspondencias entre las estructuras visuales, fonológicas y semánticas (conceptuales) del cerebro en el procesamiento de nuevo vocabulario.

Translating bus information into sign language for deaf people

This paper describes the application of language translation technologies for generating bus information in Spanish Sign Language (LSE: Lengua de Signos Española). In this work, two main systems have been developed: the first for translating text messages from information panels and the second for translating spoken Spanish into natural conversations at the information point of the bus company. Both systems are made up of a natural language translator (for converting a word sentence into a sequence of LSE signs), and a 3D avatar animation module (for playing back the signs). For the natural language translator, two technological approaches have been analyzed and integrated: an example-based strategy and a statistical translator. When translating spoken utterances, it is also necessary to incorporate a speech recognizer for decoding the spoken utterance into a word sequence, prior to the language translation module. This paper includes a detailed description of the field evaluation carried out in this domain. This evaluation has been carried out at the customer information office in Madrid involving both real bus company employees and deaf people. The evaluation includes objective measurements from the system and information from questionnaires. In the field evaluation, the whole translation presents an SER (Sign Error Rate) of less than 10% and a BLEU greater than 90%.

Learning 3D structure from 2D images using LBP features

An automatic machine learning strategy for computing the 3D structure of monocular images from a single image query using Local Binary Patterns is presented. The 3D structure is inferred through a training set composed by a repository of color and depth images, assuming that images with similar structure present similar depth maps. Local Binary Patterns are used to characterize the structure of the color images. The depth maps of those color images with a similar structure to the query image are adaptively combined and filtered to estimate the final depth map. Using public databases, promising results have been obtained outperforming other state-of-the-art algorithms and with a computational cost similar to the most efficient 2D-to-3D algorithms.