scholarly journals Hierarchical Mapping for Crosslingual Word Embedding Alignment

2020 ◽  
Vol 8 ◽  
pp. 361-376
Author(s):  
Ion Madrazo Azpiazu ◽  
Maria Soledad Pera

The alignment of word embedding spaces in different languages into a common crosslingual space has recently been in vogue. Strategies that do so compute pairwise alignments and then map multiple languages to a single pivot language (most often English). These strategies, however, are biased towards the choice of the pivot language, given that language proximity and the linguistic characteristics of the target language can strongly impact the resultant crosslingual space in detriment of topologically distant languages. We present a strategy that eliminates the need for a pivot language by learning the mappings across languages in a hierarchical way. Experiments demonstrate that our strategy significantly improves vocabulary induction scores in all existing benchmarks, as well as in a new non-English–centered benchmark we built, which we make publicly available.

2016 ◽  
Vol 12 (3) ◽  
pp. 15-22
Author(s):  
GENA RHOADES

There are many reasons for students to work in small groups in any class, but when the focus is on teaching them a language, the need to do so, multiplies. During my time as a teacher and teacher trainer, I have heard many reasons why teachers do not want to use group work, and it seems to boil down to a feeling of being unable to control the class. Fortunately, my first few years of teaching were in a program where small-group and whole class interactions were expected. Small classes gave students many opportunities to practice the target language and receive feedback from their peers and instructors.


Author(s):  
Waleed Ammar ◽  
George Mulcaire ◽  
Miguel Ballesteros ◽  
Chris Dyer ◽  
Noah A. Smith

We train one multilingual model for dependency parsing and use it to parse sentences in several languages. The parsing model uses (i) multilingual word clusters and embeddings; (ii) token-level language information; and (iii) language-specific features (fine-grained POS tags). This input representation enables the parser not only to parse effectively in multiple languages, but also to generalize across languages based on linguistic universals and typological similarities, making it more effective to learn from limited annotations. Our parser’s performance compares favorably to strong baselines in a range of data scenarios, including when the target language has a large treebank, a small treebank, or no treebank for training.


2018 ◽  
Author(s):  
Ruochen Xu ◽  
Yiming Yang ◽  
Naoki Otani ◽  
Yuexin Wu

Author(s):  
Isabel García Izquierdo ◽  
Vicent Montalt i Resurrecció

When we translate, we do so for specific communicative situations and purposes; that is, we write translations that will fulfil the needs and conventions of specific textual genres in the target language and culture. The aim of this article, which draws on data and experience from the GENTT project, is to explore the relationship between translation and genre theory in order to understand better how translators are involved in interlinguistic and intercultural communication.Genre theory is attractive to Translation Studies because it links the micro level of writing and text to the macro level of discourse and context, unites process with product and integrates the cognitive, social and profes¬sional approaches to translation. Thus, the notion of genre brings together critical elements in translation such as the reader ’s profile, expectations and preferences; the communicative situation and purpose; and the socio-cultural context. In order to understand better how translators are involved in interlinguistic and intercultural communication, we suggest a remodelling of translation in which the target genre plays a central role.


2021 ◽  
Vol 15 (02) ◽  
pp. 263-290
Author(s):  
Renjith P. Ravindran ◽  
Kavi Narayana Murthy

Word embeddings have recently become a vital part of many Natural Language Processing (NLP) systems. Word embeddings are a suite of techniques that represent words in a language as vectors in an n-dimensional real space that has been shown to encode a significant amount of syntactic and semantic information. When used in NLP systems, these representations have resulted in improved performance across a wide range of NLP tasks. However, it is not clear how syntactic properties interact with the more widely studied semantic properties of words. Or what the main factors in the modeling formulation are that encourages embedding spaces to pick up more of syntactic behavior as opposed to semantic behavior of words. We investigate several aspects of word embedding spaces and modeling assumptions that maximize syntactic coherence — the degree to which words with similar syntactic properties form distinct neighborhoods in the embedding space. We do so in order to understand which of the existing models maximize syntactic coherence making it a more reliable source for extracting syntactic category (POS) information. Our analysis shows that syntactic coherence of S-CODE is superior to the other more popular and more recent embedding techniques such as Word2vec, fastText, GloVe and LexVec, when measured under compatible parameter settings. Our investigation also gives deeper insights into the geometry of the embedding space with respect to syntactic coherence, and how this is influenced by context size, frequency of words, and dimensionality of the embedding space.


2017 ◽  
Vol 38 (4) ◽  
pp. 787-833 ◽  
Author(s):  
KYRIAKOS ANTONIOU ◽  
NAPOLEON KATSOS

ABSTRACTThe present study compares the performance of multilingual children speaking Cypriot Greek, Standard Modern Greek, and English (and sometimes an additional language), bilectal children speakers of Cypriot Greek and Standard Modern Greek, and Standard Modern Greek-speaking monolingual children on a task that measures the comprehension of different types of implicature. Despite lower scores in language ability in the target language, multilingual and bilectal children performed at rates comparable to the monolinguals with implicature. Regression analyses indicated a positive correlation between implicature, language proficiency, and age (but not executive control), albeit language ability did not affect implicature within multilinguals. We suggest an interpretation according to which multilingual, bilectal, and monolingual children maintain a comparable level of implicature understanding, but they do so by relying on different resources. Finally, a principal component analysis on different implicature types revealed a single factor of implicature performance. This outcome has implications for pragmatic theory.


English Today ◽  
2016 ◽  
Vol 33 (1) ◽  
pp. 36-42
Author(s):  
Syeda Farzana Bukhari ◽  
Xiaoguang Cheng

English is used by more than one and a half billion people as a first, second or foreign language for communication purposes (Strevens, 1992). In this context, the purpose of teaching English has shifted from mastery of the grammatical rules to the ability to use the target language for successful communication. Consequently, the communication aspect of teaching and learning English has become the key issue in the domain of second language acquisition (Yashima, 2002: 54). Therefore, the issue of whether the learners will communicate in English when they have the chance to do so and to what extent they are willing to communicate gain importance. These questions have led to the emergence of an important construct in the field of L2 instruction, i.e. willingness to communicate (hereafter, WTC), which is defined as a learner's ‘readiness to enter into discourse at a particular time with a specific person or persons, using a L2’ (MacIntyre et al., 1998: 547). MacIntyre and his associates even proposed that WTC in L2 should be conceptualized as ‘the primary goal’ of language instruction (MacIntyre et al., 1998: 545). This paper explores the important concept of WTC by looking into Pakistani students' WTC in Canada.


2021 ◽  
Vol 6 ◽  
Author(s):  
Naomi Vingron ◽  
Pauline Palma ◽  
Jason W. Gullifer ◽  
Veronica Whitford ◽  
Deanna Friesen ◽  
...  

Bilinguals juggle knowledge of multiple languages, including syntactic constructions that can mismatch (e.g., the red car, la voiture rouge; Mary sees it, Mary le voit). We used eye-tracking to examine whether French-English (n = 23) and English-French (n = 21) bilingual adults activate non-target language syntax during English L2 (Experiment 1) and L1 (Experiment 2) reading, and whether this differed from functionally monolingual English reading (Experiment 3, n = 26). People read English sentences containing syntactic constructions that were either partially shared across languages (adjective-noun constructions) or completely unshared (object-pronoun constructions). These constructions were presented in an intact form, or in a violated form that was French-consistent or French-inconsistent. For both L2 and L1 reading, bilinguals read French-consistent adjective-noun violations relatively quickly, suggesting cross-language activation. This did not occur when the same people read object-pronoun constructions manipulated in the same manner. Surprisingly, English readers exposed to French in their lifetime but functionally monolingual, also read French-consistent violations for adjective-noun constructions faster, particularly for some items. However, when we controlled for item differences in the L2 and L1 reading data, cross-language effects observed were similar to the original data pattern. Moreover, individual differences in L2 experience modulated both L2 and L1 reading for adjective-noun constructions, consistent with a cross-language activation interpretation of the data. These findings are consistent with the idea of syntactic cross-language activation during reading for some constructions. However, for several reasons, cross-language syntactic activation during comprehension may be overall more variable and challenging to investigate methodologically compared to past work on other forms of cross-language activation (i.e., single words).


Author(s):  
Bettina Matthias

While the field of drama and theater continues to inspire many foreign language teachers, improvisational theater has not received more than passing attention as a resource providing interesting warm-ups and games to be used periodically in our classroom. This article makes a case for using the format of an improvisational theater workshop in beginning foreign language teaching. The example of a three-week experimental workshop in January 2006 suggests that improvisational theater and systematic work with its basic directive ‘Show, don’t tell!’ encourage students to communicate in a foreign language environment before they may feel prepared to do so in the target language itself. Physical engagement with a situation opens up communicative possibilities, and it eventually enables students to overcome cognitive and psychological barriers to successfully move towards greater linguistic proficiency and communicative freedom. While the field of drama and theater continues to inspire many foreign language teachers, improvisational theater has not received more than passing attention as a resource providing interesting warm-ups and games to be used periodically in our classroom. This article makes a case for using the format of an improvisational theater workshop in beginning foreign language teaching. The example of a three-week experimental workshop in January 2006 suggests that improvisational theater and systematic work with its basic directive ‘Show, don’t tell!’ encourage students to communicate in a foreign language environment before they may feel prepared to do so in the target language itself. Physical engagement with a situation opens up communicative possibilities, and it eventually enables students to overcome cognitive and psychological barriers to successfully move towards greater linguistic proficiency and communicative freedom.


Sign in / Sign up

Export Citation Format

Share Document