Translation-based Approach Research Articles

Advancements in Neural Machine Translation: Methodological Innovations and Empirical Insights for Cross-Linguistic Discourse Preservation

Abstract: Natural Language Processing (NLP), in the form recognizable today, really began to take hold in the 1980s, when machine learning helped propel it to soaring heights. However, due to a lack of processing power, machine learning, and to an extent, NLP, started slowing down in innovation and ideas and had almost ground to a relative halt, until the last decade, when a sudden increase in both productivity and interest in the machine learning helped increase the amount of knowledge in the space itself. This review provides several different case studies using different methodologies. The first paper was a deep analysis on how researchers were able to use Tesseract and Google Vision in tandem with automatic data mining methods to enrich the Cherokee language database in order to preserve it from extinction. The second paper takes a query translation-based approach toward translating English to Indian languages and utilizes a Multilingual Cross-Language Information Retrieval (MLCIR) system with tools such as Part of Speech Tagger (POST), Stop-Word, and Porter Stemmer. The third paper presents CoVe, which transfers knowledge from machine translation to improve performance on NLP tasks like sentiment analysis and question answering by using contextualized word vectors along with word embeddings, achieving new state-of-the-art results on some datasets. The fourth paper aims to translate English to Pakistan Sign Language (PSL) and also uses POST and goes through dependency analysis, sentence classification, and PSL using PLS trees. The fifth paper uses a Multilingual Neural Machine Translation (NMT) system for LowResource languages and incorporates two main models: a recurrent NMT and a transformer NMT. The sixth paper analyzes how a fine-tuned transformer model seems to work better than transformer models trained from scratch on high-resource languages, while vice-versa seems to occur for low-resource languages. The seventh paper adds to this by talking about how multilingual translation seems to work better than a back-translation model. Given the diverse array of approaches that could be used, we aim to identify the most efficient and correct methodology for future researchers to use in their work, based on the papers in this literature review.

Multilingual hope speech detection: A Robust framework using transfer learning of fine-tuning RoBERTa model

Hope Speech Detection (HSD) from social media is a new direction for promoting and supporting positive content to encourage harmony and positivity in society. As users of social media belong to different linguistic communities, hope speech detection is rarely studied as a multilingual task considering low-resource languages. Moreover, prior studies explored only monolingual techniques, and the Russian language is not addressed. This study tackles the issue of Multi-lingual Hope Speech Detection (MHSD) in English and Russian languages using the transfer learning paradigm with fine-tuning approach. We explore joint multi-lingual and translation-based approaches to tackle the task of multilingualism, where the latter approach adopts the translation mechanism to transform all content into one language and then classify them. The joint multi-lingual method handles it by designing a universal classifier for various languages. We explore the strengths of the Robustly Optimized BERT Pre-Training Approach (RoBERTa) that showed a benchmark in capturing the semantics and contextual information within the content. The proposed framework consists of several stages: 1) data preprocessing, 2) representation of data using RoBERTa models, 3) fine-tuning phase, and 4) classification of hope speech into two labels. A new Russian corpus for hope speech detection is built, containing YouTube comments. Several experiments are conducted in English and Russian languages by using semi-supervised bilingual English and Russian datasets. The findings show that the proposed framework demonstrated benchmark performance and outperformed the baselines. Furthermore, the translation-based approach (Russian-RoBERTa) offered the best performance by achieving 94% accuracy and 80.24% f1-score.

Complex Knowledge Graph Embeddings Based on Convolution and Translation

Link prediction involves the use of entities and relations that already exist in a knowledge graph to reason about missing entities or relations. Different approaches have been proposed to date for performing this task. This paper proposes a combined use of the translation-based approach with the Convolutional Neural Network (CNN)-based approach, resulting in a novel model, called ConCMH. In the proposed model, first, entities and relations are embedded into the complex space, followed by a vector multiplication of entity embeddings and relational embeddings and taking the real part of the results to generate a feature matrix of their interaction. Next, a 2D convolution is used to extract features from this matrix and generate feature maps. Finally, the feature vectors are transformed into predicted entity embeddings by obtaining the inner product of the feature mapping and the entity embedding matrix. The proposed ConCMH model is compared against state-of-the-art models on the four most commonly used benchmark datasets and the obtained experimental results confirm its superiority in the majority of cases.

Fighting hate speech from bilingual hinglish speaker's perspective, a transformer- and translation-based approach.

Many people have begun to use social media platforms due to the increased use of the Internet over the previous decade. It has a lot of benefits, but it also comes with a lot of risks and drawbacks, such as Hate speech. People in multilingual societies, such as India, frequently mix their native language with English while speaking, so detecting hate content in such bilingual code-mixed data has drawn the larger interest of the research community. The majority of previous work focuses on high-resource language such as English, but very few researchers have concentrated on the mixed bilingual data like Hinglish. In this study, we investigated the performance of transformer models like IndicBERT and multilingual Bidirectional Encoder Representation(mBERT), as well as transfer learning from pre-trained language models like ULMFiT and Bidirectional encoder Representation(BERT), to find hateful content in Hinglish. Also, Transformer-based Interpreter and Feature extraction model on Deep Neural Network (TIF-DNN), is proposed in this work. The experimental results found that our proposed model outperforms existing state-of-art methods for Hate speech identification in Hinglish language with an accuracy of 73%.

When Pairs Meet Triplets: Improving Low-Resource Captioning via Multi-Objective Optimization

Image captioning for low-resource languages has attracted much attention recently. Researchers propose to augment the low-resource caption dataset into (image, rich-resource language, and low-resource language) triplets and develop the dual attention mechanism to exploit the existence of triplets in training to improve the performance. However, datasets in triplet form are usually small due to their high collecting cost. On the other hand, there are already many large-scale datasets, which contain one pair from the triplet, such as caption datasets in the rich-resource language and translation datasets from the rich-resource language to the low-resource language. In this article, we revisit the caption-translation pipeline of the translation-based approach to utilize not only the triplet dataset but also large-scale paired datasets in training. The caption-translation pipeline is composed of two models, one caption model of the rich-resource language and one translation model from the rich-resource language to the low-resource language. Unfortunately, it is not trivial to fully benefit from incorporating both the triplet dataset and paired datasets into the pipeline, due to the gap between the training and testing phases and the instability in the training process. We propose to jointly optimize the two models of the pipeline in an end-to-end manner to bridge the training and testing gap, and introduce two auxiliary training objectives to stabilize the training process. Experimental results show that the proposed method improves significantly over the state-of-the-art methods.

The End of Originalism? Comparing-Cultures and Text Exegesis

When it comes to interpreting the Constitution of the United States, constitutional scholars have long ignored the fact that the United States were a multilingual society at their inception. This essay will examine the multi-lingual situation of the eighteenth century United States (with a special focus on the German and Dutch minorities) and its relevance for the interpretation of the Constitution today, addressing theories as original intent originalism, original common meaning originalism and “living constitution.” Using a translation-based approach, it will suggest that a current paradigm in constitutional interpretation, New Originalism, is based on the dangerous illusion that the text of the Constitution had a single, discernible meaning for the Founding-Era’s general public, a public that was in fact heterogenous not only in terms of culture and race, but also of language, rendering any historical plain meaning approach a precarious proceeding.

Translation from layout-based to visual android test scripts: An empirical evaluation

Mobile GUI tests can be classified as layout-based – i.e. using GUI properties as locators – or Visual – i.e. using widgets’ screen captures as locators –. Visual test scripts require significant maintenance efforts to be kept aligned with the tested application as it evolves or it is ported to different devices.This work aims to conceptualize a translation-based approach to automatically derive Visual tests from existing layout-based counterparts or repair them when graphical changes occur, and to develop a tool that implements and validates the approach.We present TOGGLE, a tool that translates Espresso layout-based tests for Android apps to Visual tests that conform to either SikuliX, EyeAutomate, or a combination of the two tools’ syntax. An experiment is conducted to measure the precision of the translation approach, which is evaluated on maintenance tasks triggered by graphical changes due to device diversity.Our results demonstrate the feasibility of a translation-based approach, show that script portability to different devices is improved (from 32% to 93%), and indicate that translation can repair up to 90% of Visual locators in failing tests.GUI test translation mitigates challenges with Visual tests like maintenance effort and portability, enabling their wider use in industrial practice.

SentiFars

There is no doubt about the usefulness of public opinion toward different issues in social media and the World Wide Web. Extracting the feelings of people about an issue from text is not straightforward. Polarity lexicons that assign polarity tags or scores to words and phrases play an important role in sentiment analysis systems. As English is the richest language in this area, getting benefits from existing English resources in order to build new ones has attracted the interest of many researchers in recent years. In this article, we propose a new translation-based approach for building polarity resources in resource-lean languages such as Persian. The results of empirical evaluation of the proposed approach prove its effectiveness. The generated resource is the largest publicly available polarity lexicon for Persian.

Effectiveness of force dynamic explanations of English causative verbs and the role of imagery

AbstractThe current study examines the effectiveness of a CL-based force dynamic (FD) lesson relative to a more conventional approach that seeks to convey the target semantics through corresponding L1 forms. Exp. 1 (N=67) examined Japanese EFL students’ acquisition of the English verbsforce, get, have, help, let, make, andprevent, comparing the effectiveness of force dynamic explanations with dynamic diagrams and a translation-based approach. Both groups showed significant learning on both the posttest and delayed posttest, but the FD group’s greater gains over the conventional instruction on both the posttest and delayed posttest did not reach significance. Exp. 2 (N=97) replicated most of the Exp. 1 conditions but with a slightly altered instructional format to ensure participant’s focus on force dynamic relationships. The FD group’s greater gains on a posttest given three weeks following instruction fell short of statistical significance. Exp. 3 (N=54) compared the effectiveness of a FD lesson using dynamic images with a lesson without such images and found no significant differences between the groups. The three experiments demonstrate that CL-based instruction on force dynamics provides a viable alternative to conventional instruction in which target semantics are conveyed through translation.

Enhancing Image Quality of Photoacoustic Tomography Using Sub-Pitch Array Translation Approach: Simulation and Experimental Validation.

The purpose of this paper is to present a more convenient and practical alternate way of increasing the lateral discrete array sampling while using a typical λ pitch linear array transducer at receive for photoacoustic tomography (PAT) application. We have employed a linear translation-based approach, in which the array transducer is translated by sub-pitch amount to create an augmented RF frame data having denser lateral spatial sampling. The denser λ/2 and λ/4 pitch data were reconstructed and compared against conventional λ pitch reconstructed PAT image using simulation and tissue mimicking phantom experiments in terms of improvements in resolution and contrast. The results from experiments demonstrate a 34.48% improvement in lateral resolution (LR), measured in terms of full-width at half-maximum of the lateral profile of point spread function, and a maximum of 7-dB improvement in contrast is achieved while using a λ/2-pitch configuration when compared to the conventional λ-pitch configuration. It was demonstrated that λ/2- and λ/4-pitch configurations result in better LR and contrast than λ-pitch configuration. Based on the results obtained, the proposed method has the potential to serve as an easy-to-integrate and simple way of achieving better image quality without requiring to increase the system complexity with existing transducer array probe technology in regular clinical scanners.

A Multilayer Convolutional Encoder-Decoder Neural Network for Grammatical Error Correction

We improve automatic correction of grammatical, orthographic, and collocation errors in text using a multilayer convolutional encoder-decoder neural network. The network is initialized with embeddings that make use of character N-gram information to better suit this task. When evaluated on common benchmark test data sets (CoNLL-2014 and JFLEG), our model substantially outperforms all prior neural approaches on this task as well as strong statistical machine translation-based systems with neural and task-specific features trained on the same data. Our analysis shows the superiority of convolutional neural networks over recurrent neural networks such as long short-term memory (LSTM) networks in capturing the local context via attention, and thereby improving the coverage in correcting grammatical errors. By ensembling multiple models, and incorporating an N-gram language model and edit features via rescoring, our novel method becomes the first neural approach to outperform the current state-of-the-art statistical machine translation-based approach, both in terms of grammaticality and fluency.

Language and domain aware lightweight ontology matching

Concepts and relations in ontologies and in other knowledge organisation systems are usually annotated with natural language labels. Most ontology matchers rely on such labels in element-level matching techniques. State-of-the-art approaches, however, tend to make implicit assumptions about the language used in labels (usually English) and are either domain-agnostic or are built for a specific domain. When faced with labels in different languages, most approaches resort to general-purpose machine translation services to reduce the problem to monolingual English-only matching. We investigate a thoroughly different and highly extensible solution based on semantic matching where labels are parsed by multilingual natural language processing and then matched using language-independent and domain aware background knowledge acting as an interlingua. The method is implemented in NuSM, the language and domain aware evolution of the SMATCH semantic matcher, and is evaluated against a translation-based approach. We also design and evaluate a fusion matcher that combines the outputs of the two techniques in order to boost precision or recall beyond the results produced by either technique alone.

Language and Domain Aware Lightweight Ontology Matching

Concepts and relations in ontologies and in other knowledge organisation systems are usually annotated with natural language labels. Most ontology matchers rely on such labels in element-level matching techniques. State-of-the-art approaches, however, tend to make implicit assumptions about the language used in labels (usually English) and are either domain-agnostic or are built for a specic domain. When faced with labels in dierent languages, most approaches resort to general-purpose machine translation services to reduce the problem to monolingual English-only matching. We investigate a thoroughly dierent and highly extensible solution based on semantic matching where labels are parsed by multilingual natural language processing and then matched using language-independent and domain-aware background knowledge acting as an interlingua. The method is implemented in NuSM, the language and domain aware evolution of the SMATCH semantic matcher, and is evaluated against a translation-based approach. We also design and evaluate a fusion matcher that combines the outputs of the two techniques in order to boost precision or recall beyond the results produced by either technique alone.

의미 유사도를 활용한 Distant Supervision 기반의 트리플 생성 성능 향상

기존의 패턴기반 트리플 생성 시스템은 distant supervision의 가정으로 인해 오류 패턴을 생성하여 트리플 생성 시스템의 성능을 저하시키는 문제점이 있다. 이 문제점을 해결하기 위해 본 논문에서는 패턴과 프로퍼티 사이의 의미 유사도 기반의 패턴 신뢰도를 측정하여 오류 패턴을 제거하는 방법을 제안한다. 의미 유사도 측정은 비지도 학습 방법인 워드임베딩과 워드넷 기반의 어휘 의미 유사도 측정 방법을 결합하여 사용한다. 또한 한국어 패턴과 영어 프로퍼티 사이의 언어 및 어휘 불일치 문제를 해결하기 위해 정준 상관 분석과 사전 기반의 번역을 사용한다. 실험 결과에 따르면 제안한 의미 유사도 기반의 패턴 신뢰도 측정 방법이 기존의 방법보다 10% 높은 정확률의 트리플 집합을 생성하여, 트리플 생성 성능 향상을 증명하였다.

Compiling Conformant Probabilistic Planning Problems into Classical Planning

In CPP, we are given a set of actions (assumed deterministic in this paper), a distribution over initial states, a goal condition, and a real value 0 < θ ≤1. We seek a plan π such that following its execution, the goal probability is at least θ. Motivated by the success of the translation-based approach for conformant planning, introduced by Palacios and Geffner, we suggest a new compilation scheme from CPP to classical planning. Our compilation scheme maps CPP into cost-bounded classical planning, where the cost-bound represents the maximum allowed probability of failure. Empirically, this technique shows mixed, but promising results, performing very well on some domains, and less so on others when compared to the state of the art PFF planner. It is also very flexible due to its generic nature, allowing us to experiment with diverse search strategies developed for classical planning. Our results show that compilation-based technique offer a new viable approach to CPP and, possibly, more general probabilistic planning problems.

Modelling of Ontology-based Service Compositions using Petri Net

OWL-S, one of the most significant Semantic web service ontologies , provides Web Service providers with a core ontological framework and guidelines for describing the properties and capabilities of their web Services in unambiguous, computer interpretable form. In this work we present a translation-based approach for modelling the semantic Web services described in the OWL-S language. This approach employs Petri net as the fundamental means of modelling and defines a set of translation rules to map OWL-S elements into equivalent Petri net representations. They capture the main aspect of service invocations and the control flow of the service model. A case study based on a real-world OWL-S example is also conducted to examine the effectiveness of the translation-based model. DOI: http://dx.doi.org/10.5755/j01.eee.19.2.3474

Replanning in Domains with Partial Information and Sensing Actions

Replanning via determinization is a recent, popular approach for online planning in MDPs. In this paper we adapt this idea to classical, non-stochastic domains with partial information and sensing actions, presenting a new planner: SDR (Sample, Determinize, Replan). At each step we generate a solution plan to a classical planning problem induced by the original problem. We execute this plan as long as it is safe to do so. When this is no longer the case, we replan. The classical planning problem we generate is based on the translation-based approach for conformant planning introduced by Palacios and Geffner. The state of the classical planning problem generated in this approach captures the belief state of the agent in the original problem. Unfortunately, when this method is applied to planning problems with sensing, it yields a non-deterministic planning problem that is typically very large. Our main contribution is the introduction of state sampling techniques for overcoming these two problems. In addition, we introduce a novel, lazy, regression-based method for querying the agent's belief state during run-time. We provide a comprehensive experimental evaluation of the planner, showing that it scales better than the state-of-the-art CLG planner on existing benchmark problems, but also highlighting its weaknesses with new domains. We also discuss its theoretical guarantees.

A configurable translation-based cross-lingual ontology mapping system to adjust mapping outcomes

Ontologies are widely considered as the building blocks of the semantic web, and with them, comes the data interoperability issue. As ontologies are not necessarily always labelled in the same natural language, one way to achieve semantic interoperability is by means of cross-lingual ontology mapping. Translation techniques are often used as an intermediate step to translate the conceptual labels within an ontology. This approach essentially removes the natural language barrier in the mapping environment and enables the application of monolingual ontology mapping tools. This paper shows that the key to this translation-based approach to cross-lingual ontology mapping lies with selecting appropriate ontology label translations in a given mapping context. Appropriateness of the translations in the context of cross-lingual ontology mapping differs from the ontology localisation point of view, as the former aims to generate correct mappings whereas the latter aims to adapt specifications of conceptualisations to target communities. This paper further demonstrates that the mapping outcome using the translation-based cross-lingual ontology mapping approach is conditioned on the translations selected for the intermediate label translation step. In particular, this paper presents the design, implementation and evaluation of a novel cross-lingual ontology mapping system: SOCOM++. SOCOM++ provides configurable properties that can be manipulated by a user in the process of selecting label translations in an effort to adjust the subsequent mapping outcome. It is shown through the evaluation that for the same pair of ontologies, the mappings between them can be adjusted by tuning the translations for the ontology labels. This finding is not yet shown in the previous research.

Modeling and Performance Evaluation of BPEL Processes: A Stochastic-Petri-Net-Based Approach

Business Process Execution Language (BPEL) is considered as the de facto standard for Web service composition. To analyze the performance of composite service processes specified in BPEL gives the way to tell whether the process meets the performance requirements. In this paper, we propose a translation-based approach for performance analysis of BPEL processes, which employs a general stochastic Petri net (GSPN) as the intermediate representation. A set of translation rules is defined for constructs and activities of BPEL so that the processes specified in BPEL can be translated into the GSPN representations. Based on the GSPN representation of BPEL processes, we introduce a state-space method to calculate the expected-process-normal-completion-time as the performance estimate. In the case study, we obtain experimental data and conduct a confidence interval analysis to validate the feasibility and accuracy of the translation-based approach.

Teaching General English in Academic Context: Schema-Based or Translation-Based Approach?

This study is one of the first attempts to investigate the effect of schema-based general English teaching and testing on English language learning in an Iranian academic context. A total of 90 undergraduate theology students studying general English at Ferdowsi University of Mashhad (FUM) were assigned to an experimental and a control group. For the same course book two instructional approaches were adopted during an academic semester; schema-based instruction (SBI) for experimental and translation-based instruction (TBI) for control groups. At the semester’s commencement a schema-based cloze multiple choice item test (MCIT) was developed on the textbook and administered as a pre-test to ensure that the experimental group did not differ significantly from the control group. The same test was administered again at the semester’s termination along with an unseen reading comprehension test (URCT) to find out which approach was more effective. Final results showed that the performance of the experimental group was superior to the control group on both MCIT and URCT. The implications of the findings are discussed in the light of teaching English as a foreign language within an academic context.