A computationally expensive and memory intensive neural network lies behind the recent success of language representation learning. Knowledge distillation, a major technique for deploying such a vast language model in resource-scarce environments, tr
ansfers the knowledge on individual word representations learned without restrictions. In this paper, inspired by the recent observations that language representations are relatively positioned and have more semantic knowledge as a whole, we present a new knowledge distillation objective for language representation learning that transfers the contextual knowledge via two types of relationships across representations: Word Relation and Layer Transforming Relation. Unlike other recent distillation techniques for the language models, our contextual distillation does not have any restrictions on architectural changes between teacher and student. We validate the effectiveness of our method on challenging benchmarks of language understanding tasks, not only in architectures of various sizes but also in combination with DynaBERT, the recently proposed adaptive size pruning method.
Style is an integral part of natural language. However, evaluation methods for style measures are rare, often task-specific and usually do not control for content. We propose the modular, fine-grained and content-controlled similarity-based STyle Eva
Luation framework (STEL) to test the performance of any model that can compare two sentences on style. We illustrate STEL with two general dimensions of style (formal/informal and simple/complex) as well as two specific characteristics of style (contrac'tion and numb3r substitution). We find that BERT-based methods outperform simple versions of commonly used style measures like 3-grams, punctuation frequency and LIWC-based approaches. We invite the addition of further tasks and task instances to STEL and hope to facilitate the improvement of style-sensitive measures.
Natural language relies on a finite lexicon to express an unbounded set of emerging ideas. One result of this tension is the formation of new compositions, such that existing linguistic units can be combined with emerging items into novel expressions
. We develop a framework that exploits the cognitive mechanisms of chaining and multimodal knowledge to predict emergent compositional expressions through time. We present the syntactic frame extension model (SFEM) that draws on the theory of chaining and knowledge from percept'', concept'', and language'' to infer how verbs extend their frames to form new compositions with existing and novel nouns. We evaluate SFEM rigorously on the 1) modalities of knowledge and 2) categorization models of chaining, in a syntactically parsed English corpus over the past 150 years. We show that multimodal SFEM predicts newly emerged verb syntax and arguments substantially better than competing models using purely linguistic or unimodal knowledge. We find support for an exemplar view of chaining as opposed to a prototype view and reveal how the joint approach of multimodal chaining may be fundamental to the creation of literal and figurative language uses including metaphor and metonymy.
Multilingual Neural Machine Translation (MNMT) trains a single NMT model that supports translation between multiple languages, rather than training separate models for different languages. Learning a single model can enhance the low-resource translat
ion by leveraging data from multiple languages. However, the performance of an MNMT model is highly dependent on the type of languages used in training, as transferring knowledge from a diverse set of languages degrades the translation performance due to negative transfer. In this paper, we propose a Hierarchical Knowledge Distillation (HKD) approach for MNMT which capitalises on language groups generated according to typological features and phylogeny of languages to overcome the issue of negative transfer. HKD generates a set of multilingual teacher-assistant models via a selective knowledge distillation mechanism based on the language groups, and then distills the ultimate multilingual model from those assistants in an adaptive way. Experimental results derived from the TED dataset with 53 languages demonstrate the effectiveness of our approach in avoiding the negative transfer effect in MNMT, leading to an improved translation performance (about 1 BLEU score in average) compared to strong baselines.
Unifying acoustic and linguistic representation learning has become increasingly crucial to transfer the knowledge learned on the abundance of high-resource language data for low-resource speech recognition. Existing approaches simply cascade pre-tra
ined acoustic and language models to learn the transfer from speech to text. However, how to solve the representation discrepancy of speech and text is unexplored, which hinders the utilization of acoustic and linguistic information. Moreover, previous works simply replace the embedding layer of the pre-trained language model with the acoustic features, which may cause the catastrophic forgetting problem. In this work, we introduce Wav-BERT, a cooperative acoustic and linguistic representation learning method to fuse and utilize the contextual information of speech and text. Specifically, we unify a pre-trained acoustic model (wav2vec 2.0) and a language model (BERT) into an end-to-end trainable framework. A Representation Aggregation Module is designed to aggregate acoustic and linguistic representation, and an Embedding Attention Module is introduced to incorporate acoustic information into BERT, which can effectively facilitate the cooperation of two pre-trained models and thus boost the representation learning. Extensive experiments show that our Wav-BERT significantly outperforms the existing approaches and achieves state-of-the-art performance on low-resource speech recognition.
In Arabic Language, diacritics are used to specify meanings as well as pronunciations. However, diacritics are often omitted from written texts, which increases the number of possible meanings and pronunciations. This leads to an ambiguous text and m
akes the computational process on undiacritized text more difficult. In this paper, we propose a Linguistic Attentional Model for Arabic text Diacritization (LAMAD). In LAMAD, a new linguistic feature representation is presented, which utilizes both word and character contextual features. Then, a linguistic attention mechanism is proposed to capture the important linguistic features. In addition, we explore the impact of the linguistic features extracted from the text on Arabic text diacritization (ATD) by introducing them to the linguistic attention mechanism. The extensive experimental results on three datasets with different sizes illustrate that LAMAD outperforms the existing state-of-the-art models.
Linguistic typology is an area of linguistics concerned with analysis of and comparison between natural languages of the world based on their certain linguistic features. For that purpose, historically, the area has relied on manual extraction of lin
guistic feature values from textural descriptions of languages. This makes it a laborious and time expensive task and is also bound by human brain capacity. In this study, we present a deep learning system for the task of automatic extraction of linguistic features from textual descriptions of natural languages. First, textual descriptions are manually annotated with special structures called semantic frames. Those annotations are learned by a recurrent neural network, which is then used to annotate un-annotated text. Finally, the annotations are converted to linguistic feature values using a separate rule based module. Word embeddings, learned from general purpose text, are used as a major source of knowledge by the recurrent neural network. We compare the proposed deep learning system to a previously reported machine learning based system for the same task, and the deep learning system wins in terms of F1 scores with a fair margin. Such a system is expected to be a useful contribution for the automatic curation of typological databases, which otherwise are manually developed.
Translation memory systems (TMS) are the main component of computer-assisted translation (CAT) tools. They store translations allowing to save time by presenting translations on the database through matching of several types such as fuzzy matches, wh
ich are calculated by algorithms like the edit distance. However, studies have demonstrated the linguistic deficiencies of these systems and the difficulties in data retrieval or obtaining a high percentage of matching, especially after the application of syntactic and semantic transformations as the active/passive voice change, change of word order, substitution by a synonym or a personal pronoun, for instance. This paper presents the results of a pilot study where we analyze the qualitative and quantitative data of questionnaires conducted with professional translators of Spanish, French and Arabic in order to improve the effectiveness of TMS and explore all possibilities to integrate further linguistic processing from ten transformation types. The results are encouraging, and they allowed us to find out about the translation process itself; from which we propose a pre-editing processing tool to improve the matching and retrieving processes.
We present an extended version of a tool developed for calculating linguistic distances and asymmetries in auditory perception of closely related languages. Along with evaluating the metrics available in the initial version of the tool, we introduce
word adaptation entropy as an additional metric of linguistic asymmetry. Potential predictors of speech intelligibility are validated with human performance in spoken cognate recognition experiments for Bulgarian and Russian. Special attention is paid to the possibly different contributions of vowels and consonants in oral intercomprehension. Using incom.py 2.0 it is possible to calculate, visualize, and validate three measurement methods of linguistic distances and asymmetries as well as carrying out regression analyses in speech intelligibility between related languages.
The present work aims at assigning a complexity score between 0 and 1 to a target word or phrase in a given sentence. For each Single Word Target, a Random Forest Regressor is trained on a feature set consisting of lexical, semantic, and syntactic in
formation about the target. For each Multiword Target, a set of individual word features is taken along with single word complexities in the feature space. The system yielded the Pearson correlation of 0.7402 and 0.8244 on the test set for the Single and Multiword Targets, respectively.
