Unique Words Research Articles

Evaluating Arabic to English Machine Translation

Online text machine translation systems are widely used throughout the world freely. Most of these systems use statistical machine translation (SMT) that is based on a corpus full with translation examples to learn from them how to translate correctly. Online text machine translation systems differ widely in their effectiveness, and therefore we have to fairly evaluate their effectiveness. Generally the manual (human) evaluation of machine translation (MT) systems is better than the automatic evaluation, but it is not feasible to be used. The distance or similarity of MT candidate output to a set of reference translations are used by many MT evaluation approaches. This study presents a comparison of effectiveness of two free online machine translation systems (Google Translate and Babylon machine translation system) to translate Arabic to English. There are many automatic methods used to evaluate different machine translators, one of these methods; Bilingual Evaluation Understudy (BLEU) method. BLEU is used to evaluate translation quality of two free online machine translation systems under consideration. A corpus consists of more than 1000 Arabic sentences with two reference English translations for each Arabic sentence is used in this study. This corpus of Arabic sentences and their English translations consists of 4169 Arabic words, where the number of unique Arabic words is 2539. This corpus is released online to be used by researchers. These Arabic sentences are distributed among four basic sentence functions (declarative, interrogative, exclamatory, and imperative). The experimental results show that Google machine translation system is better than Babylon machine translation system in terms of precision of translation from Arabic to English.

How Many Visual Concepts?

Visual scenes require complex description and modeling that involves more than a list of words. The visual semantic concept basis is larger than the number of unique words. More efforts are needed to build out the set of visual semantic concepts.

Question Classification Using Extreme Learning Machine on Semantic Features

In statistical machine learning approaches for question classification, efforts based on lexical feature space require high computation power and complex data structures. This is due to the large number of unique words (or high dimensionality). Choosing semantic features instead could significantly reduce the dimensionality of the feature space. This article describes the use of Extreme Learning Machine (ELM) for question classification based on semantic features to improve both the training and testing speeds compared to the benchmark Support Vector Machine (SVM) classifier. Improvements have also been made to the head word extraction and word sense disambiguation processes. These have resulted in a higher accuracy (an increase of 0.2%) for the classification of coarse classes compared to the benchmark. For the fine classes, however, there is a 1.0% decrease in accuracy but is compensated by a significant increase in speed (92.1% on average).

Language Individuation and Marker Words: Shakespeare and His Maxwell's Demon

BackgroundWithin the structural and grammatical bounds of a common language, all authors develop their own distinctive writing styles. Whether the relative occurrence of common words can be measured to produce accurate models of authorship is of particular interest. This work introduces a new score that helps to highlight such variations in word occurrence, and is applied to produce models of authorship of a large group of plays from the Shakespearean era.MethodologyA text corpus containing 55,055 unique words was generated from 168 plays from the Shakespearean era (16th and 17th centuries) of undisputed authorship. A new score, CM1, is introduced to measure variation patterns based on the frequency of occurrence of each word for the authors John Fletcher, Ben Jonson, Thomas Middleton and William Shakespeare, compared to the rest of the authors in the study (which provides a reference of relative word usage at that time). A total of 50 WEKA methods were applied for Fletcher, Jonson and Middleton, to identify those which were able to produce models yielding over 90% classification accuracy. This ensemble of WEKA methods was then applied to model Shakespearean authorship across all 168 plays, yielding a Matthews' correlation coefficient (MCC) performance of over 90%. Furthermore, the best model yielded an MCC of 99%.ConclusionsOur results suggest that different authors, while adhering to the structural and grammatical bounds of a common language, develop measurably distinct styles by the tendency to over-utilise or avoid particular common words and phrasings. Considering language and the potential of words as an abstract chaotic system with a high entropy, similarities can be drawn to the Maxwell's Demon thought experiment; authors subconsciously favour or filter certain words, modifying the probability profile in ways that could reflect their individuality and style.

Self-organization of progress across the century of physics

We make use of information provided in the titles and abstracts of over half a million publications that were published by the American Physical Society during the past 119 years. By identifying all unique words and phrases and determining their monthly usage patterns, we obtain quantifiable insights into the trends of physics discovery from the end of the 19th century to today. We show that the magnitudes of upward and downward trends yield heavy-tailed distributions, and that their emergence is due to the Matthew effect. This indicates that both the rise and fall of scientific paradigms is driven by robust principles of self-organization. Data also confirm that periods of war decelerate scientific progress, and that the later is very much subject to globalisation.

On the Selection of the Redundant Carrier Positions in UW-OFDM

Unique-word (UW) OFDM is a new multicarrier technique that was recently proposed in [M. Huemer, C. Hofbauer, and J. B. Huber, “The Potential of Unique Words in OFDM, Proc. InOWo, Hamburg, Germany, Sep. 2010, pp. 140-144]. In this technique, the guard interval that contains a priori known samples is part of the FFT interval. To be able to construct the UW-OFDM signal, some carriers must be sacrificed to transmit redundant information-these carriers are called the redundant carriers. It turns out that the average redundant energy needed strongly depends on the positions of the redundant carriers. In this paper, we look for the redundant carrier placement that minimizes the average redundant energy. However, this optimization is a NP-hard problem. Therefore, we reformulate the optimization problem and implement a branch and bound (B&B) algorithm to find the optimum. The results of the B&B search algorithm are compared to a theoretical lower bound, and with the results for two low cost redundant carrier placements, i.e., the quasi-uniform (QU) distribution [H. Steendam, “The Quasi-Uniform Redundant Carrier Placement for UW-OFDM,” Vehicular Technology Conf., Fall 2012, Quebec, Canada, Sep. 2012], for the case where no guard bands are present, and the maximum distance (MD) distribution, in case forbidden guard bands are present. We show in this paper that when the full band is available for redundant carriers, both the redundant carrier placement found with the B&B algorithm and the QU distribution result in a redundant energy that is very close to the theoretical lower bound. Although the B&B slightly outperforms the QU distribution, this implies that the QU distribution is a good low cost solution to the redundant carrier placement problem in the case of no guard bands. However, when there are forbidden guard bands in which the redundant carriers cannot be placed, the QU distribution can no longer be used. In that case, the results of the B&B algorithm are compared to that of the MD distribution. We show in this paper that the proposed B&B algorithm outperforms the MD distribution in the sense that it needs noticeably lower redundant energy.

Online Lexical Competition During Spoken Word Recognition and Word Learning in Children and Adults

Lexical competition that occurs as speech unfolds is a hallmark of adult oral language comprehension crucial to rapid incremental speech processing. This study used pause detection to examine whether lexical competition operates similarly at 7-8years and tested variables that influence "online" lexical activity in adults. Children (n=20) and adults (n=17) were slower to detect pauses in familiar words with later uniqueness points. Faster latencies were obtained for words with late uniqueness points in constraining compared with neutral sentences; no such effect was observed for early unique words. Following exposure to novel competitors ("biscal"), children (n=18) and adults (n=18) showed competition for existing words with early uniqueness points ("biscuit") after 24hr. Thus, online lexical competition effects are remarkably similar across development.

HyDEn: A Hybrid Steganocryptographic Approach for Data Encryption Using Randomized Error-Correcting DNA Codes

This paper presents a novel hybrid DNA encryption (HyDEn) approach that uses randomized assignments of unique error-correcting DNA Hamming code words for single characters in the extended ASCII set. HyDEn relies on custom-built quaternary codes and a private key used in the randomized assignment of code words and the cyclic permutations applied on the encoded message. Along with its ability to detect and correct errors, HyDEn equals or outperforms existing cryptographic methods and represents a promising in silico DNA steganographic approach.

Biometrics and Biosecurity 2013

Biometrics and Biosecurity 2013

Frequency of high-quality communication behaviors used by primary care providers of heterozygous infants after newborn screening

ObjectiveTo examine the quality of communication likely to be experienced by parents when being first informed about how newborn screening identified heterozygous “carrier” status for cystic fibrosis or sickle cell disease. MethodsPrimary care providers (PCPs) of infants found to have carrier status were telephoned over a 48-month period, and asked to rehearse with a standardized patient how they would inform the infants’ parent(s). 214 rehearsal transcripts were abstracted using explicit criteria methods to measure the frequency of five categories of high-quality communication behaviors. ResultsOverall, PCPs used large amounts of jargon and failed to use high quality communication behaviors. On average, PCPs used 18.6 total jargon words (8.7 unique words), but explained 2.4 jargon words. The most frequent assessment of understanding was the close-ended version, although it was only seen in 129 of 214 transcripts. The most common organizing behavior was importance emphasis (121/214). Precautionary empathy was rare; the most frequent behavior was “instruction about emotion” (33/214). ConclusionThe limited use of high-quality communication behaviors in rehearsals raises concern about parental understanding, decision-making, and psychosocial outcomes after newborn screening. Practice implicationsMeasurement of specific behaviors may help PCPs to improve communication, and thereby improve the patient experience.

PhiSpy: a novel algorithm for finding prophages in bacterial genomes that combines similarity- and composition-based strategies

Prophages are phages in lysogeny that are integrated into, and replicated as part of, the host bacterial genome. These mobile elements can have tremendous impact on their bacterial hosts’ genomes and phenotypes, which may lead to strain emergence and diversification, increased virulence or antibiotic resistance. However, finding prophages in microbial genomes remains a problem with no definitive solution. The majority of existing tools rely on detecting genomic regions enriched in protein-coding genes with known phage homologs, which hinders the de novo discovery of phage regions. In this study, a weighted phage detection algorithm, PhiSpy was developed based on seven distinctive characteristics of prophages, i.e. protein length, transcription strand directionality, customized AT and GC skew, the abundance of unique phage words, phage insertion points and the similarity of phage proteins. The first five characteristics are capable of identifying prophages without any sequence similarity with known phage genes. PhiSpy locates prophages by ranking genomic regions enriched in distinctive phage traits, which leads to the successful prediction of 94% of prophages in 50 complete bacterial genomes with a 6% false-negative rate and a 0.66% false-positive rate.

Document Summarization Using Positive Pointwise Mutual Information

The degree of success in document summarization processes depends on the performance of the method used in identifying significant sentences in the documents. The collection of unique words characterizes the major signature of the document, and forms the basis for Term-Sentence-Matrix (TSM). The Positive Pointwise Mutual Information, which works well for measuring semantic similarity in the Term-Sentence-Matrix, is used in our method to assign weights for each entry in the Term-Sentence-Matrix. The Sentence-Rank-Matrix generated from this weighted TSM, is then used to extract a summary from the document. Our experiments show that such a method would outperform most of the existing methods in producing summaries from large documents.

Detection of Fraudulent Emails by Authorship Extraction

Fraudulent emails can be detected by extraction of authorship information from the contents of emails. This paper presents information extraction based on unique words from the emails. These unique words will be used as representative features to train Radial Basis function (RBF). Final weights are obtained and subsequently used for testing. The percentage of identification of email authorship depends upon number of RBF centers and the type of functional words used for training RBF. One hundred and fifty authors with over one hundred files from the sent folder of Enron email dataset are considered. A total of 300 unique words of number of characters in each word ranging from three to seven are considered. Training and testing of RBF are done by taking different lengths of words. Our simulation shows the effectiveness of the proposed RBF network for email authorship identification. The accuracy of authorship identification ranges from 95% to 97%.

New Sentence Recognition Materials Developed Using a Basic Non-Native English Lexicon

The objective of this project was to develop new sentence test materials drawing on a basic non-native English lexicon that could be used to test speech recognition for various listener populations. These materials have been designed to provide a test tool that is less linguistically biased, compared with materials that are currently available, for sentence recognition for non-native as well as native speakers of English. One hundred non-native speakers of English were interviewed on a range of 20 conversational topics. Over 26 hr of recorded non-native English speech were transcribed. These transcriptions were used to create a lexicon of over 4,000 unique words. The words from this lexicon were used to create the new materials based on a simple syntactic sentence structure frame. Twenty lists of 25 sentences were developed. Each sentence has 4 keywords, providing 100 keywords per list. Lists were equated for rate of occurrence of keywords in lexicon, high-frequency count (total number of affricates and fricatives), number of syllables, and distribution of syntactic structure. Listening-in-noise results for native-English-speaking, normal-hearing listeners indicated similar performance across lists. The Basic English Lexicon materials provide a large set of sentences for native and non-native English speech-recognition testing.

The Words-in-Noise Test (WIN), List 3: A Practice List

The Words-in-Noise Test (WIN) was developed as an instrument to quantify the ability of listeners to understand monosyllabic words in background noise using multitalker babble (Wilson, 2003). The 50% point, which is calculated with the Spearman-Kärber equation (Finney, 1952), is used as the evaluative metric with the WIN materials. Initially, the WIN was designed as a 70-word instrument that presented ten unique words at each of seven signal-to-noise ratios from 24 to 0 dB in 4 dB decrements. Subsequently, the 70-word list was parsed into two 35-word lists that achieved equivalent recognition performances (Wilson and Burks, 2005). This report involves the development of a third list (WIN List 3) that was developed to serve as a practice list to familiarize the participant with listening to words presented in background babble. To determine-on young listeners with normal hearing and on older listeners with sensorineural hearing loss-the psychometric properties of the WIN List 3 materials. A quasi-experimental, repeated-measures design was used. Twenty-four young adult listeners (M = 21.6 yr) with normal pure-tone thresholds (≤ 20 dB HL at 250 to 8000 Hz) and 24 older listeners (M = 65.9 yr) with sensorineural hearing loss participated. The level of the babble was fixed at 80 dB SPL with the level of the words varied from 104 to 80 dB SPL in 4 dB decrements. For listeners with normal hearing, the 50% points for Lists 1 and 2 were similar (4.3 and 5.1 dB S/N, respectively), both of which were lower than the 50% point for List 3 (7.4 dB S/N). A similar relation was observed with the listeners with hearing loss, 50% points for Lists 1 and 2 of 12.2 and 12.4 dB S/N, respectively, compared to 15.8 dB S/N for List 3. The differences between Lists 1 and 2 and List 3 were significant. The relations among the psychometric functions and the relations among the individual data both reflected these differences. The significant ∼3 dB difference between performances on WIN Lists 1 and 2 and on WIN List 3 by the listeners with normal hearing and the listeners with hearing loss dictates caution with the use of List 3. The use of WIN List 3 should be reserved for ancillary purposes in which equivalent recognition performances are not required, for example, as a practice list or a stand alone measure.

A method for determining the number of documents needed for a gold standard corpus

The unstructured narratives in medicine have been increasingly targeted for content extraction using the techniques of natural language processing (NLP). In most cases, these efforts are facilitated by creating a manually annotated set of narratives containing the ground truth; commonly referred to as a gold standard corpus. This corpus is used for modeling, fine-tuning, and testing NLP software as well as providing the basis for training in machine learning. Determining the number of annotated documents (size) for this corpus is important, but rarely described; rather, the factors of cost and time appear to dominate decision-making about corpus size. In this report, a method is outlined to determine gold standard size based on the capture probabilities for the unique words within a target corpus. To demonstrate this method, a corpus of dictation letters from the Michigan Pain Consultant (MPC) clinics for pain management are described and analyzed. A well-formed working corpus of 10,000 dictations was first constructed to provide a representative subset of the total, with no more than one dictation letter per patient. Each dictation was divided into words and common words were removed. The Poisson function was used to determine probabilities of word capture within samples taken from the working corpus, and then integrated over word length to give a single capture probability as a function of sample size. For these MPC dictations, a sample size of 500 documents is predicted to give a capture probability of approximately 0.95. Continuing the demonstration of sample selection, a provisional gold standard corpus of 500 documents was selected and examined for its similarity to the MPC structured coding and demographic data available for each patient. It is shown that a representative sample, of justifiable size, can be selected for use as a gold standard.

Non-Systematic Complex Number RS Coded OFDM by Unique Word Prefix

In this paper, we expand our recently introduced concept of unique word orthogonal frequency division multiplexing (UW-OFDM). In UW-OFDM the cyclic prefixes (CPs) are replaced by deterministic sequences, the so-called unique words (UWs). The UWs are generated by appropriately loading a set of redundant subcarriers. By that a systematic complex number Reed-Solomon (RS) code construction is introduced in a quite natural way, because an RS code may be defined as the set of vectors, for which a block of successive zeros occurs in the other domain w.r.t. a discrete Fourier transform. (For a fixed block different to zero, i.e., a UW, a coset code of an RS code is generated.) A remaining problem in the original systematic coded UW-OFDM concept is the fact that the redundant subcarrier symbols disproportionately contribute to the mean OFDM symbol energy. In this paper we introduce the concept of non-systematic coded UW-OFDM, where the redundancy is no longer allocated to dedicated subcarriers, but distributed over all subcarriers. We derive optimum complex valued code generator matrices matched to the best linear unbiased estimator (BLUE) and to the linear minimum mean square error (LMMSE) data estimator, respectively. With the help of simulations we highlight the advantageous spectral properties and the superior bit error ratio (BER) performance of non-systematic coded UW-OFDM compared to systematic coded UW-OFDM and to CP-OFDM in additive white Gaussian noise (AWGN ) as well as in frequency selective environments.

Toward enhanced Arabic speech recognition using part of speech tagging

One major source of suboptimal performance in automatic continuous speech recognition systems is misrecognition of small words. In general, errors resulting from small words are much more than errors resulting from long words. Therefore, compounding some words (small or long) to produce longer words is welcome by speech recognition decoders. In this paper, we present a novel approach to artificially generate compound words using part of speech tagging. For this purpose, we consider two cases in Arabic speech where two words are pronounced without a silence period in between: a noun followed by an adjective, and a preposition followed by any word. To collect the candidate compound words, we use Stanford Arabic tagger to tag all words in our baseline transcription corpus. Then, compound words are generated whenever any of the two cases occur in a sequence of two words. The unique compound words are then added to the expanded pronunciation dictionary, as well as to the language model. Using Sphinx 3, we test the proposed method for a 5.4 hours speech corpus of modern standard Arabic. The results show a significant improvement, as the word error rate is reduced by 2.39%.

Efficacy of Forensic Statement Analysis in Distinguishing Truthful from Deceptive Eyewitness Accounts of Highly Stressful Events*

Laboratory-based detecting deception research suggests that truthful statements differ from those of deceptive statements. This nonlaboratory study tested whether forensic statement analysis (FSA) methods would distinguish genuine from false eyewitness accounts about exposure to a highly stressful event. A total of 35 military participants were assigned to truthful or deceptive eyewitness conditions. Genuine eyewitness reported truthfully about exposure to interrogation stress. Deceptive eyewitnesses studied transcripts of genuine eyewitnesses for 24 h and falsely claimed they had been interrogated. Cognitive Interviews were recorded, transcribed, and assessed by FSA raters blind to the status of participants. Genuine accounts contained more unique words, external and contextual referents, and a greater total word count than did deceptive statements. The type-token ratio was lower in genuine statements. The classification accuracy using FSA techniques was 82%. FSA methods may be effective in real-world circumstances and have relevance to professionals in law enforcement, security, and criminal justice.

Search terms and a validated brief search filter to retrieve publications on health-related values in Medline: a word frequency analysis study.

Healthcare debates and policy developments are increasingly concerned with a broad range of values-related areas. These include not only ethical, moral, religious, and other types of values 'proper', but also beliefs, preferences, experiences, choices, satisfaction, quality of life, etc. Research on such issues may be difficult to retrieve. This study used word frequency analysis to generate a broad pool of search terms and a brief filter to facilitate relevant searches in bibliographic databases. Word frequency analysis for 'values terms' was performed on citations on diabetes, obesity, dementia, and schizophrenia (Medline; 2004-2006; 4440 citations; 1,110,291 words). Concordance® and SPSS 14.0 were used. Text words and MeSH terms of high frequency and precision were compiled into a search filter. It was validated on datasets of citations on dentistry and food hypersensitivity. 144 unique text words and 124 unique MeSH terms of moderate and high frequency (≥ 20) and very high precision (≥ 90%) were identified. Of these, 19 text words and seven MeSH terms were compiled into a 'brief values filter'. In the derivation dataset, it had a sensitivity of 76.8% and precision of 86.8%. In the validation datasets, its sensitivity and precision were, respectively, 70.1% and 63.6% (food hypersensitivity) and 47.1% and 82.6% (dentistry). This study provided a varied pool of search terms and a simple and highly effective tool for retrieving publications on health-related values. Further work is required to facilitate access to such research and enhance its chances of being translated into practice, policy, and service improvements.