Search

Native Language Identification is one of the growing subfields in Natural Language Processing (NLP). The task of Native Language Identification (NLI) is mainly concerned with predicting the native language of an author’s writing in a second language. In this paper, we investigate the performance of two types of features; content-based features vs. content independent features when they are evaluated on a different corpus (using social media data “Reddit”). In this NLI task, the predefined models are trained on one corpus (TOEFL) and then the trained models are evaluated on a different data using an external corpus (Reddit). Three classifiers are used in this task; the baseline, linear SVM, and Logistic Regression. Results show that content-based features are more accurate and robust than content independent ones when tested within corpus and across corpus.

Пояснювальна записка магістерської дисертації складається з шести розділів, містить 31 таблицю, 7 рисунків, 8 додатків, 20 джерел. Об'єктом дослідження є веб-сервіс аналізу достовірності інформації. Предметом дослідження є алгоритми отримання та обробки текстів, що виступають доказами або запереченнями твердження, що має бути перевіреним. Метою роботи є дослідження, розробка та вдосконалення моделей пошуку уривків текстів та методів відбору речень. Новизною роботи є покращення аналізу відношення тверджень при вирішені задачі NLI (Natural Language Inference) в умовах наявності відволікаючої інформації в контексті тверджень, що перевіряються на істинність. В результаті виконання магістерської дисертації було розроблено програмний продукт, що дозволяє перевірити твердження навіть у великій кількості контекстуальної інформації. The explanatory note of the master's thesis consists of six sections, contains 31 tables, 7 figures, 8 appendices, 20 sources. The object of the study is a web service for analyzing the reliability of information. The subject of research is algorithms for obtaining and processing texts that serve as proofs or refutations of statements that need to be verified. The purpose of the work is to research, develop and improve models for searching text fragments and sentence selection methods. The novelty of the work is the improvement of the analysis of the relation of statements when solving the NLI (Natural Language Inference) problem in the presence of distracting information in the context of statements that are checked for truth. As a result of the master's thesis, a software product was developed that allows you to verify statements even with a large amount of contextual information.

Native language identification (NLI) is a classification task which identifies the mother tongue of a language learner based on spoken or written material. The task gained popularity when it was featured in the 2017 BEA-12-workshop and since then many applications have been successfully found for NLI - ranging from language learning to authorship identification and forensic science. While a considerable amount of research has already been done in this area, we introduce a novel approach of incorporating syntactic information into the implementation of a BERT-based NLI model. In addition, we train separate models to test whether erroneous input sequences perform better than corrected sequences. To answer these questions we carry out both a quantitative and qualitative analysis. In addition, we test our idea of implementing a BERT-based GEC model to supply more training data to our NLI model without the need for manual annotation. Our results suggest that our models do not outperform the SVM baseline, but we attribute this result to the lack of training data in our dataset, as transformer-based architectures like BERT need huge amounts of data to be successfully fine-tuned. In turn, simple linear models like SVM perform well on small amounts of data. We also find that erroneous structures in data come useful when combined with syntactic information but neither boosts the performance of NLI model separately. Furthermore, our implemented GEC system performs well enough to produce more data for our NLI models, as their scores increase after implementing the additional data, resulting from our second experiment. We believe that our proposed architecture is potentially suitable for the NLI task if we incorporate extensions which we suggest in the conclusion section.

Despite great progression within the field of Natural Language Inference, recent studies have revealed that the state-of-the-art neural models suffer from confounding shortcomings. These revelations cast doubt on whether these models possess as deep and robust language understanding as the hype surrounding them suggests. While the field has focused greatly on alleviating these shortcomings through adversarial training, less effort has gone into investigating other, more substantial ways of dealing with the linguistic phenomena that the neural models struggle with. This thesis will therefore propose and explore a novel hybrid approach to NLI, which seeks to combine the strengths of the state-of-the-art neural models with the preciseness of rule-based systems. The method of hybridizing is based on a principle of transitive reasoning, and it is designed to be independent of model choice in order to develop in tandem with an ever-progressing state-of-the-art. Our results indicate that this hybrid approach, dubbed T-NLI, may indeed improve results on datasets that are challenging for typical neural models. The ambition of this project is that these results will incite more optimism regarding further research of hybrid approaches to NLI.

Recent results on analytical modeling of non-linear interference due to signal propagation in multi-span optical systems are reviewed, including high-symbol rate and ultra-wideband scenarios. Latest advances in nonlinearity mitigation, based on the use of machine learning techniques and nonlinearity tolerant modulation formats, are also discussed.

The accuracy of both GN and Enhanced GN (EGN) models are investigated for a legacy dispersion-managed ultra-long-haul optical link by comparing its predictions with split-step simulations. The EGN model shows good accuracy in system performance prediction, when signal power depletion and ASE contribution to nonlinear interference (NLI) are considered.

We derived an approximate non-linear interference (NLI) closed-form model (CFM), capable of handling a very broad range of optical WDM system scenarios. We tested the CFM over 8500 randomized C-band WDM systems, of which 6250 were fully-loaded and 2250 were partially loaded. The systems had highly diversified channel formats, symbol rates, fibers, as well as other parameters. We improved the CFM accuracy by augmenting the formula with simple machine-learning factors, optimized by leveraging the system test-set. We further improvedthe CFM by adding a term which models special situations where NLI has high self-coherence. In the end, we obtained a very good match with the results found using the numerically-integrated Enhanced GN-model (or EGN-model). We also checked the CFM accuracy by comparing its predictions with full-C-Band split-step simulations of 300 randomized systems. The combined high accuracy and very fast computation time (milliseconds) of the CFM potentially make it an effective tool for real-time physical-layer-aware optical network management and control., Author's version of the manuscript accepted for publication on the Journal of Lightwave Technology, May 2020

To effectively operate multivendor disaggregated networks, the performance of the physical layer needs to be assessed by a quality-of transmission estimator (QoT-E) delivering quick results with a given reliability range. Current state-of-the-art wavelength-division-multiplexing channels are based on multilevel modulation formats relying on DSP-operated coherent receivers, propagating on uncompensated and amplified optical links. In this transmission scenario, beside amplified spontaneous emission noise accumulation, nonlinear propagation impairments are well summarized by the accumulation of a Gaussian-distributed disturbance: the nonlinear interference (NLI). When exploiting a transmission bandwidth exceeding the C-band, the interaction of NLI generation with the stimulated Raman scattering (SRS) must be properly considered. We present the derivation of the generalized Gaussian noise (GGN) model for NLI generation, including the SRS and, in general, a spectral and spatial variation of gain/loss. We validate its accuracy by comparing performances predicted by a QoT-E based on the GGN model with measurements on a testbed exploiting commercial equipment, including 100 Gb/s transponders. Considering that operational parameters of the commercial equipment are known with a large range of uncertainty, an excellent agreement with errors within 0.5 dB on the generalized Signal-to-noise ratio ( $\text{SNR}$ ) is shown, demonstrating that the GGN-model can be used for the QoT-E in multivendor network scenarios. Moreover, the GGN model has shown the capability to predict the spectral tilting due to SRS in $\text{SNR}$ performances, enabling its application to evaluate the impact of linear pretilting for SRS precompensation and NLI generation.

This chapter aims to provide a comprehensive picture of the impact of fiber nonlinear effects on modern coherent wavelength division multiplexing () systems' performance. First, the main nonlinearity models currently available are introduced and discussed in depth. Then, various specific aspects are addressed, such as the interplay of polarization mode dispersion ()/polarization dependent loss () and nonlinearity, or the dependence of nonlinear effects on modulation format. The important topic of nonlinear effects mitigation is then dealt with. Finally, system performance metrics and capacity are discussed extensively, as to how they are fundamentally influenced and limited by fiber nonlinearity.

We tested the accuracy of a fully-closed-form approximate GN-model formula over 3000 different $C$ -band fully-loaded WDM-QAM system scenarios and 1200 partially-loaded ones. By leveraging the large system test-set, we modified the formula to obtain a closed-form formula approximating the enhanced GN-model (EGN-model). The combined high accuracy and very fast computation time (ms) of such formula potentially make it an effective tool for real-time physical-layer-aware optical network management and control.

Is it possible to infer the native language of an author from a non-native text? Can we perform this task fully automatically? The interest in answers to these questions led to the emergence of a research field called Native Language Identification (NLI) in the first decade of this century. The requirement to automatically identify a particular property based on some language data situates the task in the intersection between computer science and linguistics, or in the context of computational linguistics, which combines both disciplines. This thesis targets several relevant research questions in the context of NLI. In particular, what is the role of surface features and more abstract linguistic cues? How to combine different sets of features, and how to optimize the resulting large models? Do the findings generalize across different data sets? Can we benefit from considering the task in the light of the language variation theory? In order to approach these questions, we conduct a range of quantitative and qualitative explorations, employing different machine learning techniques. We show how linguistic insight can advance technology, and how technology can advance linguistic insight, constituting a fruitful and promising interplay.

This report explores of the feasibility of using pattern matching for implementing a robust Natural Language Interface (NLI) over a limited Application Programming Interface (API). Because APIs are used to such a great extent today and often in mobile applications, it becomes more important to find simple ways of making them accessible to end users. A very intuitive way to access information via an API is using natural language. Therefore, this study first explores the possibility of building a corpus of the most common phrases used for a particular API. It is then explored how those phrases adhere to patterns, and how these patterns can be used to extract meaning from a phrase. Finally it evaluates an implementation of an NLI using pattern matching system based on the patterns. The result of the building of the corpus shows that although the amount of unique phrases used with our API seems to increase quite steadily, the amount of patterns those phrases follow converges to a constant quickly. This implies that it is possible to use these patterns to create an NLI that is robust enough to query an API effectively. The evaluation of the pattern matching system indicates that this technique can be used to successfully extract information from a phrase if its pattern is known by the system. Den här rapporten utforskar hur genomförbart det är att använda mönstermatchning för att implementera ett robust användargränssnitt för styrning med naturligt språk (Natural Language Interface, NLI) över en begränsad Application Programming Interface (API). Eftersom APIer används i stor utsträckning idag, ofta i mobila applikationer, har det blivit allt mer viktigt att hitta sätt att göra dem ännu mer tillgängliga för slutanvändare. Ett mycket intuitivt sätt att komma åt information är med hjälp av naturligt språk via en API. I den här rapporten redogörs först för möjligheten att bygga ett korpus för en viss API and att skapa mönster för mönstermatchning på det korpuset. Därefter utvärderas en implementation av ett NLI som bygger på mönstermatchning med hjälp av korpuset. Resultatet av korpusuppbyggnaden visar att trots att antalet unika fraser som används för vårt API ökar ganska stadigt, så konvergerar antalat mönster på de fraserna relativt snabbt mot en konstant. Detta antyder att det är mycket möjligt att använda desssa mönster för att skapa en NLI som är robust nog för en API. Utvärderingen av implementationen av mönstermatchingssystemet antyder att tekniken kan användas för att framgångsrikt extrahera information från fraser om mönstret frasen följer finns i systemet.

This report explores of the feasibility of using pattern matching for implementing a robust Natural Language Interface (NLI) over a limited Application Programming Interface (API). Because APIs are used to such a great extent today and often in mobile applications, it becomes more important to find simple ways of making them accessible to end users. A very intuitive way to access information via an API is using natural language. Therefore, this study first explores the possibility of building a corpus of the most common phrases used for a particular API. It is then explored how those phrases adhere to patterns, and how these patterns can be used to extract meaning from a phrase. Finally it evaluates an implementation of an NLI using pattern matching system based on the patterns. The result of the building of the corpus shows that although the amount of unique phrases used with our API seems to increase quite steadily, the amount of patterns those phrases follow converges to a constant quickly. This implies that it is possible to use these patterns to create an NLI that is robust enough to query an API effectively. The evaluation of the pattern matching system indicates that this technique can be used to successfully extract information from a phrase if its pattern is known by the system. Den här rapporten utforskar hur genomförbart det är att använda mönstermatchning för att implementera ett robust användargränssnitt för styrning med naturligt språk (Natural Language Interface, NLI) över en begränsad Application Programming Interface (API). Eftersom APIer används i stor utsträckning idag, ofta i mobila applikationer, har det blivit allt mer viktigt att hitta sätt att göra dem ännu mer tillgängliga för slutanvändare. Ett mycket intuitivt sätt att komma åt information är med hjälp av naturligt språk via en API. I den här rapporten redogörs först för möjligheten att bygga ett korpus för en viss API and att skapa mönster för mönstermatchning på det korpuset. Därefter utvärderas en implementation av ett NLI som bygger på mönstermatchning med hjälp av korpuset. Resultatet av korpusuppbyggnaden visar att trots att antalet unika fraser som används för vårt API ökar ganska stadigt, så konvergerar antalat mönster på de fraserna relativt snabbt mot en konstant. Detta antyder att det är mycket möjligt att använda desssa mönster för att skapa en NLI som är robust nog för en API. Utvärderingen av implementationen av mönstermatchingssystemet antyder att tekniken kan användas för att framgångsrikt extrahera information från fraser om mönstret frasen följer finns i systemet.

While microRNAs have emerged as an important component of gene regulatory networks, it remains unclear how microRNAs collaborate with transcription factors in the gene networks that determines neuronal cell fate. Here we show that in the developing spinal cord, the expression of miR-218 is directly upregulated by the Isl1–Lhx3 complex, which drives motor neuron fate. Inhibition of miR-218 suppresses the generation of motor neurons in both chick neural tube and mouse embryonic stem cells, suggesting that miR-218 plays a crucial role in motor neuron differentiation. Results from unbiased RISC-trap screens, in vivo reporter assays and overexpression studies indicated that miR-218 directly represses transcripts that promote developmental programs for interneurons. In addition, we found that miR-218 activity is required for Isl1–Lhx3 to effectively induce motor neurons and suppress interneuron fates. Together our results reveal an essential role of miR-218 as a downstream effector of the Isl1–Lhx3 complex in establishing motor neuron identity. microRNAs have emerged as important components of numerous gene regulatory networks. Here the authors demonstrate that miR-218 is an essential component of the gene regulatory network that controls motor neuron fate specification in the developing spinal cord.

Presentation provided for NLI (Networked Learning Initiatives) on Friday, October 24th, 2014. As VT Life Sciences & Health Sciences faculty, researchers, and graduate students, your work is geared towards addressing questions, problems, and discoveries that are important to Virginia communities, and to our global society. Join your colleagues and Ginny Pannabecker, Life Science and Scholarly Communication Librarian, for a discussion of author rights in Life Sciences & Health Sciences. Share your journal selection and publishing experiences. Find out about options to keep the copyrights you need to use your published works in teaching, to make your published works freely available for public interest research, and to provide long-term access to supplementary data sets - in Virginia and beyond.

In this paper, we have implemented an adaptive noise canceller (ANC) for ECG signals with the help of Modified Particle Swarm Optimization (MPSO). Implementing MPSO on ANC provides better performance than any other optimization technique used to enhance the ECG signal. In this work, the various modes of MPSO for finding fidelity parameters like signal to noise ratio (SNR), peak reconstruction error (PRE) and mean square error (MSE) have been evaluated. Our simulation results show that 19% improvement in SNR, 91% decrease in peak reconstruction error, and 99% reduction in mean square error can be achieved using proposed algorithm over the conventional PSO technique.

We propose an accurate nonlinearity closed-form model (CFM) for forward Raman-amplified WDM links which enables fast system optimization. We show a detailed study of a single-span link SNR maximization and flattening.