LKMT:Linguistics Knowledge-Driven Multi-Task Neural Machine Translation for Urdu and English

Thanks to the strong representation capability of pre-trained language models,supervised machine translation models have achieved outstanding performance.However,the performances of these models drop sharply when the scale of the parallel training corpus is limited.Considering the pre-trained language model has a strong ability for monolingual representation,it is the key challenge for machine translation to construct the in-depth relationship between the source and target language by injecting the lexical and syntactic information into pre-trained language models.To alleviate the dependence on the parallel corpus,we propose a Linguistics Knowledge-Driven MultiTask(LKMT)approach to inject part-of-speech and syntactic knowledge into pre-trained models,thus enhancing the machine translation performance.On the one hand,we integrate part-of-speech and dependency labels into the embedding layer and exploit large-scale monolingual corpus to update all parameters of pre-trained language models,thus ensuring the updated language model contains potential lexical and syntactic information.On the other hand,we leverage an extra self-attention layer to explicitly inject linguistic knowledge into the pre-trained language model-enhanced machine translation model.Experiments on the benchmark dataset show that our proposed LKMT approach improves the Urdu-English translation accuracy by 1.97 points and the English-Urdu translation accuracy by 2.42 points,highlighting the effectiveness of our LKMT framework.Detailed ablation experiments confirm the positive impact of part-of-speech and dependency parsing on machine translation.

Query Expansion Using Wikipedia and a Concept Base in Cross-language Information Retrieval

The present paper describes the use of online free language resources for translating and expanding queries in CLIR （cross-language information retrieval）. In a previous study, we proposed method queries that were translated by two machine translation systems on the Language Gridem. The queries were then expanded using an online dictionary to translate compound words or word phrases. A concept base was used to compare back translation words with the original query in order to delete mistranslated words. In order to evaluate the proposed method, we constructed a CLIR system and used the science documents of the NTCIR1 dataset. The proposed method achieved high precision. However~ proper nouns （names of people and places） appear infrequently in science documents. In information retrieval, proper nouns present unique problems. Since proper nouns are usually unknown words, they are difficult to find in monolingual dictionaries, not to mention bilingual dictionaries. Furthermore, the initial query of the user is not always the best description of the desired information. In order to solve this problem, and to create a better query representation, query expansion is often proposed as a solution. Wikipedia was used to translate compound words or word phrases. It was also used to expand queries together with a concept base. The NTCIRI and NTCIR 6 datasets were used to evaluate the proposed method. In the proposed method, the CLIR system was implemented with a high rate of precision. The proposed syst had a higher ranking than the NTCIRI and NTCIR6 participation systems.

Readability Assessment of Textbooks in Low Resource Languages

Readability is a fundamental problem in textbooks assessment.For low resources languages(LRL),however,little investigation has been done on the readability of textbook.In this paper,we proposed a readability assessment method for Tibetan textbook(a low resource language).We extract features based on the information that are gotten by Tibetan segmentation and named entity recognition.Then,we calculate the correlation of different features using Pearson Correlation Coefficient and select some feature sets to design the readability formula.Fit detection,F test and T test are applied on these selected features to generate a new readability assessment formula.Experiment shows that this new formula is capable of assessing the readability of Tibetan textbooks.

Improving Parallel Corpus Quality for Chinese-Vietnamese Statistical Machine Translation

The performance of a machine translation system heavily depends on the quantity and quality of the bilingual language resource. However,getting a parallel corpus,which has a large scale and is of high quality,is a very difficult task especially for low resource languages such as Chinese-Vietnamese. Fortunately,multilingual user generated contents（ UGC）,such as bilingual movie subtitles,provide us access to automatic construction of the parallel corpus. Although the amount of UGC parallel corpora can be considerable,the original corpus is not suitable for statistical machine translation（ SMT） systems. The corpus may contain translation errors,sentence mismatching,free translations,etc. To improve the quality of the bilingual corpus for SMT systems,three filtering methods are proposed： sentence length difference,the semantic of sentence pairs,and machine learning. Experiments are conducted on the Chinese to Vietnamese translation corpus.Experimental results demonstrate that all the three methods effectively improve the corpus quality,and the machine translation performance（ BLEU score） can be improved by 1. 32.

Benchmarks for Pirá2.0,a Reading Comprehension Dataset about the Ocean,the Brazilian Coast,and Climate Change

Piráis a reading comprehension dataset focused on the ocean,the Brazilian coast,and climate change,built from a collection of scientific abstracts and reports on these topics.This dataset represents a versatile language resource,particularly useful for testing the ability of current machine learning models to acquire expert scientific knowledge.Despite its potential,a detailed set of baselines has not yet been developed for Pirá.By creating these baselines,researchers can more easily utilize Piráas a resource for testing machine learning models across a wide range of question answering tasks.In this paper,we define six benchmarks over the Pirádataset,covering closed generative question answering,machine reading comprehension,information retrieval,open question answering,answer triggering,and multiple choice question answering.As part of this effort,we have also produced a curated version of the original dataset,where we fixed a number of grammar issues,repetitions,and other shortcomings.Furthermore,the dataset has been extended in several new directions,so as to face the aforementioned benchmarks:translation of supporting texts from English into Portuguese,classification labels for answerability,automatic paraphrases of questions and answers,and multiple choice candidates.The results described in this paper provide several points of reference for researchers interested in exploring the challenges provided by the Pirádataset.

National Language Capacity in Global Competition

The term national language capacity refers to the sum total of a nation＇s ability to grasp linguistic resources, provide language services, deal with linguistic issues, develop the language, and related tasks. Key capacities include the possession of linguistic resources, utilization of language and language services, development and use of linguistic resources, the language of the members of the nation, reserves of linguistic talent, language management, development of language enterprises and linguistic influence. The development of informatization has made national language capacity an important part of national strength. It is hard power as well as soft power, playing a very important role in social progress and cultural inheritance, in promoting economic development and technological innovation, and in protecting a country＇s national security and international development.