Cross-lingual Transfer Research Articles

Cross-Lingual Transfer Learning of Non-Native Acoustic Modeling for Pronunciation Error Detection and Diagnosis

In computer-assisted pronunciation training (CAPT), the scarcity of large-scale non-native corpora and human expert annotations are two fundamental challenges to non-native acoustic modeling. Most existing approaches of acoustic modeling in CAPT are based on non-native corpora while there are so many living languages in the world. It is impractical to collect and annotate every non-native speech corpus considering different language pairs. In this work, we address non-native acoustic modeling (both on phonetic and articulatory level) based on transfer learning. In order to effectively train acoustic models of non-native speech without using such data, we propose to exploit two large native speech corpora of learner's native language (L1) and target language (L2) to model cross-lingual phenomena. This kind of transfer learning can provide a better feature representation of non-native speech. Experimental evaluations are carried out for Japanese speakers learning English. We first demonstrate the proposed acoustic-phone model achieves a lower word error rate in non-native speech recognition. It also improves the pronunciation error detection based on goodness of pronunciation (GOP) score. For diagnosis of pronunciation errors, the proposed acoustic-articulatory modeling method is effective for providing detailed feedback at the articulation level.

A Survey of Cross-lingual Word Embedding Models

Cross-lingual representations of words enable us to reason about word meaning in multilingual contexts and are a key facilitator of cross-lingual transfer when developing natural language processing models for low-resource languages. In this survey, we provide a comprehensive typology of cross-lingual word embedding models. We compare their data requirements and objective functions. The recurring theme of the survey is that many of the models presented in the literature optimize for the same objectives, and that seemingly different models are often equivalent, modulo optimization strategies, hyper-parameters, and such. We also discuss the different ways cross-lingual word embeddings are evaluated, as well as future challenges and research horizons.

Cross-Lingual Syntactic Transfer with Limited Resources

We describe a simple but effective method for cross-lingual syntactic transfer of dependency parsers, in the scenario where a large amount of translation data is not available. This method makes use of three steps: 1) a method for deriving cross-lingual word clusters, which can then be used in a multilingual parser; 2) a method for transferring lexical information from a target language to source language treebanks; 3) a method for integrating these steps with the density-driven annotation projection method of Rasooli and Collins (2015). Experiments show improvements over the state-of-the-art in several languages used in previous work, in a setting where the only source of translation data is the Bible, a considerably smaller corpus than the Europarl corpus used in previous work. Results using the Europarl corpus as a source of translation data show additional improvements over the results of Rasooli and Collins (2015). We conclude with results on 38 datasets from the Universal Dependencies corpora.

Cross-lingual sentiment transfer with limited resources

We describe two transfer approaches for building sentiment analysis systems without having gold labeled data in the target language. Unlike previous work that is focused on using only English as the source language and a small number of target languages, we use multiple source languages to learn a more robust sentiment transfer model for 16 languages from different language families. Our approaches explore the potential of using an annotation projection approach and a direct transfer approach using cross-lingual word representations and neural networks. Whereas most previous work relies on machine translation, we show that we can build cross-lingual sentiment analysis systems without machine translation or even high quality parallel data. We have conducted experiments assessing the availability of different resources such as in-domain parallel data, out-of-domain parallel data, and in-domain comparable data. Our experiments show that we can build a robust transfer system whose performance can in some cases approach that of a supervised system.

Multilingual Convolutional, Long Short-Term Memory, Deep Neural Networks for Low Resource Speech Recognition

Stand-alone and the combined model of Convolutional Neural networks (CNNs) and Long Short-Term Memory (LSTM) and Deep neural Networks (DNNs) have shown great improvements in a variety of Speech Recognition tasks. In this paper we also combined these networks but in this paper we used them for multilingual speech recognition, for the prediction and correction (PAC) architecture, in order to calculate the state probability. Our proposed model is known as PAC-MCLDNN. In this paper, we present experiment results for multilingual training on AP16-OLR task. Furthermore, cross-lingual model transfer and multitask learning for under resourced languages such as Uyghur and Vietnam are also performed which further improved the recognition results.

Reassessing the value of resources for cross-lingual transfer of POS tagging models

When linguistically annotated data is scarce, as is the case for many under-resourced languages, one has to resort to less complete forms of annotations obtained from crawled dictionaries and/or through cross-lingual transfer. Several recent works have shown that learning from such partially supervised data can be effective in many practical situations. In this work, we review two existing proposals for learning with ambiguous labels which extend conventional learners to the weakly supervised setting: a history-based model using a variant of the perceptron, on the one hand; an extension of the Conditional Random Fields model on the other hand. Focusing on the part-of-speech tagging task, but considering a large set of ten languages, we show (a) that good performance can be achieved even in the presence of ambiguity, provided however that both monolingual and bilingual resources are available; (b) that our two learners exploit different characteristics of the training set, and are successful in different situations; (c) that in addition to the choice of an adequate learning algorithm, many other factors are critical for achieving good performance in a cross-lingual transfer setting.

A Distributed Representation-Based Framework for Cross-Lingual Transfer Parsing

This paper investigates the problem of cross-lingual transfer parsing, aiming at inducing dependency parsers for low-resource languages while using only training data from a resource-rich language (e.g., English). Existing model transfer approaches typically don't include lexical features, which are not transferable across languages. In this paper, we bridge the lexical feature gap by using distributed feature representations and their composition. We provide two algorithms for inducing cross-lingual distributed representations of words, which map vocabularies from two different languages into a common vector space. Consequently, both lexical features and non-lexical features can be used in our model for cross-lingual transfer. Furthermore, our framework is flexible enough to incorporate additional useful features such as cross-lingual word clusters. Our combined contributions achieve an average relative error reduction of 10.9% in labeled attachment score as compared with the delexicalized parser, trained on English universal treebank and transferred to three other languages. It also significantly outperforms state-of-the-art delexicalized models augmented with projected cluster features on identical data. Finally, we demonstrate that our models can be further boosted with minimal supervision (e.g., 100 annotated sentences) from target languages, which is of great significance for practical usage.

A Representation Learning Framework for Multi-Source Transfer Parsing

Cross-lingual model transfer has been a promising approach for inducing dependency parsers for low-resource languages where annotated treebanks are not available. The major obstacles for the model transfer approach are two-fold: 1. Lexical features are not directly transferable across languages; 2. Target language-specific syntactic structures are difficult to be recovered. To address these two challenges, we present a novel representation learning framework for multi-source transfer parsing. Our framework allows multi-source transfer parsing using full lexical features straightforwardly. By evaluating on the Google universal dependency treebanks (v2.0), our best models yield an absolute improvement of 6.53% in averaged labeled attachment score, as compared with delexicalized multi-source transfer models. We also significantly outperform the state-of-the-art transfer system proposed most recently.

C-BiLDA extracting cross-lingual topics from non-parallel texts by distinguishing shared from unshared content

We study the problem of extracting cross-lingual topics from non-parallel multilingual text datasets with partially overlapping thematic content (e.g., aligned Wikipedia articles in two different languages). To this end, we develop a new bilingual probabilistic topic model called comparable bilingual latent Dirichlet allocation (C-BiLDA), which is able to deal with such comparable data, and, unlike the standard bilingual LDA model (BiLDA), does not assume the availability of document pairs with identical topic distributions. We present a full overview of C-BiLDA, and show its utility in the task of cross-lingual knowledge transfer for multi-class document classification on two benchmarking datasets for three language pairs. The proposed model outperforms the baseline LDA model, as well as the standard BiLDA model and two standard low-rank approximation methods (CL-LSI and CL-KCCA) used in previous work on this task.

Knowledge Exchanges in Smart-Sourcing Contracts - Learning from the Genius of Al-Biruni

Smart-Sourcing is typically knowledge-intensive and collaborative. Success is not possible without tactics for effective knowledge-exchange throughout the life-cycle of a contract. It involves bringing teams across organizations (vendors, clients, sub-contractors, service providers, suppliers etc) together in order to acquire, share and co-create knowledge. Knowledge exchange across boundaries is not new. The works of Abu Rayhan Al-Biruni (973-1048AD), a Persian polymath scholar who made significant contributions in diverse fields of study, were examined using frameworks of knowledge assets and knowledge flows. The knowledge flows of knowledge assets were studied using Boisot’s I-space model and the knowledge transfer to and from Al-Biruni was studied using Gupta and Govindarajan’s knowledge transfer framework. The factors that have contributed to the tremendous achievements in scholarship by the genius of Al-Biruni were found to be the possession of multiple intelligences, a balance of wisdom with an ability to positively influence his environment and a strong value system within which he operated. His methodology also provides several lessons that can be learned and applied for knowledge diffusion. The research also suggests that possessing multiple intelligences and a balance in wisdom enables the people involved in multi-national, cross-cultural and cross-lingual knowledge transfer to be able to give better form and structure to knowledge enabling easier and better knowledge diffusion.

Possibility of Crosslingual Transfer: A Comparative Study of Chinese (L1) and English (L2) Lessons

This article reports a comparative study of first language (L1) Chinese and second language (L2) English lessons in Hong Kong secondary schools. Aiming to explore how to take advantage of L1 academic proficiency for the benefit of L2 development, the study asked how Chinese and English lessons compare in terms of the content knowledge covered and the levels and depth of such coverage. Through a corpus‐aided approach, 3 differences were identified, namely, the “aboutness” of the lessons through a keyword analysis, the “abstractness” reflected in the use of metalanguage, and the “richness in coverage” of meanings in classrooms. The findings indicated that the Chinese lessons were either text‐based, involving discussion of literary concepts in metalanguage, or task‐based, engaging students in debating about a current issue, while English lessons were based on linguistic knowledge and composed mainly of discrete grammatical drills through simple classroom routines. This article argues that the 3 differences identified were deficiencies in the L2 classrooms. However, they could also be areas where crosslingual transfer might be possible. Constant exposure to literary text‐related concepts and cognitively complex tasks in the L1 may lead to students’ development of academic language proficiency; such proficiency could be capitalized for the development of academic proficiency in the L2.

The intersection of cognitive and sociocultural factors in the development of reading comprehension among immigrant students

The present paper synthesizes the international research literature on the educational achievement of immigrant and minority language students by articulating three propositions for which there is strong empirical evidence: (a) print access and literacy engagement play a key role in promoting reading comprehension; (b) the development of bilingual students’ L1 proficiency plays a positive role in L2 academic development; and (c) societal power relations play a direct causal role in promoting school failure among students from subordinated communities. This interpretation of the empirical evidence is contrasted with the conclusions of recent North American and European reviews. For example, the comprehensive review of literacy development among minority students conducted in the United States by the National Literacy Panel on Language-Minority Children and Youth acknowledged the legitimacy of bilingual education as a policy option but said very little about the role of literacy engagement in promoting reading comprehension. By contrast, various reports of the Programme for International Student Achievement (PISA) conducted by the Organisation for Economic Cooperation and Development (OECD) highlighted the importance of reading engagement for reading achievement but discounted bilingual education as a feasible or realistic policy option. The instructional implications of the present review include the need for educators to promote print access and literacy engagement, teach for cross-lingual transfer, and affirm students’ identities in classroom interactions.

A Probe into Lexical Depth: What is the Direction of Transfer for L1 Literacy and L2 Development?

This paper examines the intersection between heritage language (HL) learning and the development of English and Farsi deep lexical knowledge. The study compares two groups of Farsi-English bilingual children with different HL educational backgrounds and a group of English-only children by testing their paradigmatic-syntagmatic knowledge of words. A statistical analysis of the children's deep lexical knowledge was conducted in light of their HL literacy experience, second language (English) schooling, and length of residence. The findings revealed that longer length of residence and L2 schooling correlates with better performance on the L2 measures of lexical depth, whereas longer residence in the home country and first language (L1) formal schooling do not correlate with better performance on the Farsi measures. The study concluded that, in the long term, the learning of a heritage language, in combination with L2 academic instruction is more effective to the cross-lingual transfer of academic skills.

Cross-lingual transfer effects between English and Italian cognates and noncognates

Two experiments were implemented to examine the proposition that transfer effects between English and Italian translations are restricted to words from the same lexical paradigm. The experiments involved standard repetition priming, a traditional laboratory procedure, and a comparative analysis of the impact of high frequency and low frequency words in one language on performance in an alternative language. The two procedures supported the same conclusion, and showed that transfer effects are restricted to morphologically related words, and that, for the subjects used in our experiments, transfer was asymmetric, involving facilitation for second language words only. The results are discussed with reference to functional accounts of the mental lexicon.