IsiZulu is one of the eleven official languages of South Africa and roughly half the population can speak it. It is the first (home) language for over 10 million people in South Africa. Only a few computational resources exist for isiZulu and its related Nguni languages, yet the imperative for tool development exists. We focus on natural language generation, and the grammar options and preferences in particular, which will inform verbalization of knowledge representation languages and could contribute to machine translation. The verbalization pattern specification shows that the grammar rules are elaborate and there are several options of which one may have preference. We devised verbalization patterns for subsumption, basic disjointness, existential and universal quantification, and conjunction. This was evaluated in a survey among linguists and non-linguists. Some differences between linguists and non-linguists can be observed, with the former much more in agreement, and preferences depend on the overall structure of the sentence, such as singular for subsumption and plural in other cases.
It has been proved that large scale realistic Knowledge Based Machine Translation applications require acquisition of huge knowledge about language and about the world. This knowledge is encoded in computational grammars, lexicons and domain models. Another approach which avoids the need for collecting and analyzing massive knowledge, is the Example Based approach, which is the topic of this paper. We show through the paper that using Example Based in its native form is not suitable for translating into Arabic. Therefore a modification to the basic approach is presented to improve the accuracy of the translation process. The basic idea of the new approach is to improve the technique by which template-based approaches select the appropriate templates.
Inspired by embedded programming languages, an embedded CNL (controlled natural language) is a proper fragment of an entire natural language (its host language), but it has a parser that recognizes the entire host language. This makes it possible to process out-of-CNL input and give useful feedback to users, instead of just reporting syntax errors. This extended abstract explains the main concepts of embedded CNL implementation in GF (Grammatical Framework), with examples from machine translation and some other ongoing work.
In this paper, we describe methods for handling multilingual non-compositional constructions in the framework of GF. We specifically look at methods to detect and extract non-compositional phrases from parallel texts and propose methods to handle such constructions in GF grammars. We expect that the methods to handle non-compositional constructions will enrich CNLs by providing more flexibility in the design of controlled languages. We look at two specific use cases of non-compositional constructions: a general-purpose method to detect and extract multilingual multiword expressions and a procedure to identify nominal compounds in German. We evaluate our procedure for multiword expressions by performing a qualitative analysis of the results. For the experiments on nominal compounds, we incorporate the detected compounds in a full SMT pipeline and evaluate the impact of our method in machine translation process.
In todays digital world automated Machine Translation of one language to another has covered a long way to achieve different kinds of success stories. Whereas Babel Fish supports a good number of foreign languages and only Hindi from Indian languages, the Google Translator takes care of about 10 Indian languages. Though most of the Automated Machine Translation Systems are doing well but handling Indian languages needs a major care while handling the local proverbs/ idioms. Most of the Machine Translation system follows the direct translation approach while translating one Indian language to other. Our research at KMIT R&D Lab found that handling the local proverbs/idioms is not given enough attention by the earlier research work. This paper focuses on two of the majorly spoken Indian languages Marathi and Telugu, and translation between them. Handling proverbs and idioms of both the languages have been given a special care, and the research outcome shows a significant achievement in this direction.
Text classification is a task of automatic classification of text into one of the predefined categories. The problem of text classification has been widely studied in different communities like natural language processing, data mining and information retrieval. Text classification is an important constituent in many information management tasks like topic identification, spam filtering, email routing, language identification, genre classification, readability assessment etc. The performance of text classification improves notably when phrase patterns are used. The use of phrase patterns helps in capturing non-local behaviours and thus helps in the improvement of text classification task. Phrase structure extraction is the first step to continue with the phrase pattern identification. In this survey, detailed study of phrase structure learning methods have been carried out. This will enable future work in several NLP tasks, which uses syntactic information from phrase structure like grammar checkers, question answering, information extraction, machine translation, text classification. The paper also provides different levels of classification and detailed comparison of the phrase structure learning methods.
Named Entity Recognition is always important when dealing with major Natural Language Processing tasks such as information extraction, question-answering, machine translation, document summarization etc so in this paper we put forward a survey of Named Entities in Indian Languages with particular reference to Assamese. There are various rule-based and machine learning approaches available for Named Entity Recognition. At the very first of the paper we give an idea of the available approaches for Named Entity Recognition and then we discuss about the related research in this field. Assamese like other Indian languages is agglutinative and suffers from lack of appropriate resources as Named Entity Recognition requires large data sets, gazetteer list, dictionary etc and some useful feature like capitalization as found in English cannot be found in Assamese. Apart from this we also describe some of the issues faced in Assamese while doing Named Entity Recognition.
Machine Translation is the challenging problem for Indian languages. Every day we can see some machine translators being developed, but getting a high quality automatic translation is still a very distant dream . The correct translated sentence for Hindi language is rarely found. In this paper, we are emphasizing on English-Hindi language pair, so in order to preserve the correct MT output we present a ranking system, which employs some machine learning techniques and morphological features. In ranking no human intervention is required. We have also validated our results by comparing it with human ranking.
Machine translation is the process of translating text from one language to another. In this paper, Statistical Machine Translation is done on Assamese and English language by taking their respective parallel corpus. A statistical phrase based translation toolkit Moses is used here. To develop the language model and to align the words we used two another tools IRSTLM, GIZA respectively. BLEU score is used to check our translation system performance, how good it is. A difference in BLEU scores is obtained while translating sentences from Assamese to English and vice-versa. Since Indian languages are morphologically very rich hence translation is relatively harder from English to Assamese resulting in a low BLEU score. A statistical transliteration system is also introduced with our translation system to deal basically with proper nouns, OOV (out of vocabulary) words which are not present in our corpus.
In this paper, we propose a novel neural network model called RNN Encoder-Decoder that consists of two recurrent neural networks (RNN). One RNN encodes a sequence of symbols into a fixed-length vector representation, and the other decodes the representation into another sequence of symbols. The encoder and decoder of the proposed model are jointly trained to maximize the conditional probability of a target sequence given a source sequence. The performance of a statistical machine translation system is empirically found to improve by using the conditional probabilities of phrase pairs computed by the RNN Encoder-Decoder as an additional feature in the existing log-linear model. Qualitatively, we show that the proposed model learns a semantically and syntactically meaningful representation of linguistic phrases.
Machine Translation is one of the major oldest and the most active research area in Natural Language Processing. Currently, Statistical Machine Translation (SMT) dominates the Machine Translation research. Statistical Machine Translation is an approach to Machine Translation which uses models to learn translation patterns directly from data, and generalize them to translate a new unseen text. The SMT approach is largely language independent, i.e. the models can be applied to any language pair. Statistical Machine Translation (SMT) attempts to generate translations using statistical methods based on bilingual text corpora. Where such corpora are available, excellent results can be attained translating similar texts, but such corpora are still not available for many language pairs. Statistical Machine Translation systems, in general, have difficulty in handling the morphology on the source or the target side especially for morphologically rich languages. Errors in morphology or syntax in the target language can have severe consequences on meaning of the sentence. They change the grammatical function of words or the understanding of the sentence through the incorrect tense information in verb. Baseline SMT also known as Phrase Based Statistical Machine Translation (PBSMT) system does not use any linguistic information and it only operates on surface word form. Recent researches shown that adding linguistic information helps to improve the accuracy of the translation with less amount of bilingual corpora. Adding linguistic information can be done using the Factored Statistical Machine Translation system through pre-processing steps. This paper investigates about how English side pre-processing is used to improve the accuracy of English-Tamil SMT system.
We describe a unified and coherent syntactic framework for supporting a semantically-informed syntactic approach to statistical machine translation. Semantically enriched syntactic tags assigned to the target-language training texts improved translation quality. The resulting system significantly outperformed a linguistically naive baseline model (Hiero), and reached the highest scores yet reported on the NIST 2009 Urdu-English translation task. This finding supports the hypothesis (posed by many researchers in the MT community, e.g., in DARPA GALE) that both syntactic and semantic information are critical for improving translation quality---and further demonstrates that large gains can be achieved for low-resource languages with different word order than English.
A crowdsourcing translation approach is an effective tool for globalization of site content, but it is also an important source of parallel linguistic data. For the given site, processed with a crowdsourcing system, a sentence-aligned corpus can be fetched, which covers a very narrow domain of terminology and language patterns - a site-specific domain. These data can be used for training and estimation of site-specific statistical machine translation engine
Neural machine translation is a relatively new approach to statistical machine translation based purely on neural networks. The neural machine translation models often consist of an encoder and a decoder. The encoder extracts a fixed-length representation from a variable-length input sentence, and the decoder generates a correct translation from this representation. In this paper, we focus on analyzing the properties of the neural machine translation using two models; RNN Encoder--Decoder and a newly proposed gated recursive convolutional neural network. We show that the neural machine translation performs relatively well on short sentences without unknown words, but its performance degrades rapidly as the length of the sentence and the number of unknown words increase. Furthermore, we find that the proposed gated recursive convolutional network learns a grammatical structure of a sentence automatically.
The authors of (Cho et al., 2014a) have shown that the recently introduced neural network translation systems suffer from a significant drop in translation quality when translating long sentences, unlike existing phrase-based translation systems. In this paper, we propose a way to address this issue by automatically segmenting an input sentence into phrases that can be easily translated by the neural network translation model. Once each segment has been independently translated by the neural machine translation model, the translated clauses are concatenated to form a final translation. Empirical results show a significant improvement in translation quality for long sentences.
We present power low rank ensembles (PLRE), a flexible framework for n-gram language modeling where ensembles of low rank matrices and tensors are used to obtain smoothed probability estimates of words in context. Our method can be understood as a generalization of n-gram modeling to non-integer n, and includes standard techniques such as absolute discounting and Kneser-Ney smoothing as special cases. PLRE training is efficient and our approach outperforms state-of-the-art modified Kneser Ney baselines in terms of perplexity on large corpora as well as on BLEU score in a downstream machine translation task.
The increasing diversity of languages used on the web introduces a new level of complexity to Information Retrieval (IR) systems. We can no longer assume that textual content is written in one language or even the same language family. In this paper, we demonstrate how to build massive multilingual annotators with minimal human expertise and intervention. We describe a system that builds Named Entity Recognition (NER) annotators for 40 major languages using Wikipedia and Freebase. Our approach does not require NER human annotated datasets or language specific resources like treebanks, parallel corpora, and orthographic rules. The novelty of approach lies therein - using only language agnostic techniques, while achieving competitive performance. Our method learns distributed word representations (word embeddings) which encode semantic and syntactic features of words in each language. Then, we automatically generate datasets from Wikipedia link structure and Freebase attributes. Finally, we apply two preprocessing stages (oversampling and exact surface form matching) which do not require any linguistic expertise. Our evaluation is two fold: First, we demonstrate the system performance on human annotated datasets. Second, for languages where no gold-standard benchmarks are available, we propose a new method, distant evaluation, based on statistical machine translation.
We explore how to improve machine translation systems by adding more translation data in situations where we already have substantial resources. The main challenge is how to buck the trend of diminishing returns that is commonly encountered. We present an active learning-style data solicitation algorithm to meet this challenge. We test it, gathering annotations via Amazon Mechanical Turk, and find that we get an order of magnitude increase in performance rates of improvement.
Building machine translation (MT) test sets is a relatively expensive task. As MT becomes increasingly desired for more and more language pairs and more and more domains, it becomes necessary to build test sets for each case. In this paper, we investigate using Amazon's Mechanical Turk (MTurk) to make MT test sets cheaply. We find that MTurk can be used to make test sets much cheaper than professionally-produced test sets. More importantly, in experiments with multiple MT systems, we find that the MTurk-produced test sets yield essentially the same conclusions regarding system performance as the professionally-produced test sets yield.
This paper describes our resource-building results for an eight-week JHU Human Language Technology Center of Excellence Summer Camp for Applied Language Exploration (SCALE-2009) on Semantically-Informed Machine Translation. Specifically, we describe the construction of a modality annotation scheme, a modality lexicon, and two automated modality taggers that were built using the lexicon and annotation scheme. Our annotation scheme is based on identifying three components of modality: a trigger, a target and a holder. We describe how our modality lexicon was produced semi-automatically, expanding from an initial hand-selected list of modality trigger words and phrases. The resulting expanded modality lexicon is being made publicly available. We demonstrate that one tagger---a structure-based tagger---results in precision around 86% (depending on genre) for tagging of a standard LDC data set. In a machine translation application, using the structure-based tagger to annotate English modalities on an English-Urdu training corpus improved the translation quality score for Urdu by 0.3 Bleu points in the face of sparse training data.