Pre-trained language models (LMs) are currently integral to many natural language processing systems. Although multilingual LMs were also introduced to serve many languages, these have limitations such as being costly at inference time and the size and diversity of non-English data involved in their pre-training. We remedy these issues for a collection of diverse Arabic varieties by introducing two powerful deep bidirectional transformer-based models, ARBERT and MARBERT. To evaluate our models, we also introduce ARLUE, a new benchmark for multi-dialectal Arabic language understanding evaluation. ARLUE is built using 42 datasets targeting six different task clusters, allowing us to offer a series of standardized experiments under rich conditions. When fine-tuned on ARLUE, our models collectively achieve new state-of-theart results across the majority of tasks (37 out of 48 classification tasks, on the 42 datasets). Our best model acquires the highest ARLUE score (77.40) across all six task clusters, outperforming all other models including XLM-R Large (∼ 3.4× larger size). Our models are publicly available at https://github.com/UBC-NLP/marbert and ARLUE will be released through the same repository.

IoT promises a new era of connectivity that goes beyond laptops and smart connected devices to connected vehicles, smart homes, smart cities and connected healthcare. The huge volume of data that is collected from millions of IoT devices raises information security and privacy concerns for users. This paper presents a new scalable encryption technique, called Flexible encryption Technique (FlexenTech), to protect IoT data during storage and in transit. FlexenTech is suitable for resource constrained devices and networks. It offers a low encryption time, defends against common attacks such as replay attacks and defines a configurable mode, where any number of rounds or key sizes may be used. Experimental analysis of FlexenTech shows its robustness in terms of its multiple configurable confidentiality levels by allowing various configurations.This configurability provides several advantages for resource constrained devices, including reducing the encryption computation time by up to 9.7% when compared to its best rivals in the literature.

Semantic textual similarity is the basis of countless applications and plays an important role in diverse areas, such as information retrieval, plagiarism detection, information extraction and machine translation. This article proposes an innovative word embedding-based system devoted to calculate the semantic similarity in Arabic sentences. The main idea is to exploit vectors as word representations in a multidimensional space in order to capture the semantic and syntactic properties of words. IDF weighting and Part-of-Speech tagging are applied on the examined sentences to support the identification of words that are highly descriptive in each sentence. The performance of our proposed system is confirmed through the Pearson correlation between our assigned semantic similarity scores and human judgments.

No abstract

This article describes our proposed system named LIM-LIG. This system is designed for SemEval 2017 Task1: Semantic Textual Similarity (Track1). LIM-LIG proposes an innovative enhancement to word embedding-based model devoted to measure the semantic similarity in Arabic sentences. The main idea is to exploit the word representations as vectors in a multidimensional space to capture the semantic and syntactic properties of words. IDF weighting and Part-of-Speech tagging are applied on the examined sentences to support the identification of words that are highly descriptive in each sentence. LIM-LIG system achieves a Pearsons correlation of 0.74633, ranking 2nd among all participants in the Arabic monolingual pairs STS task organized within the SemEval 2017 evaluation campaign.

We describe models focused at the understudied problem of translating between monolingual and code-mixed language pairs. More specifically, we offer a wide range of models that convert monolingual English text into Hinglish (code-mixed Hindi and English). Given the recent success of pretrained language models, we also test the utility of two recent Transformer-based encoder-decoder models (i.e., mT5 and mBART) on the task finding both to work well. Given the paucity of training data for code-mixing, we also propose a dependency-free method for generating code-mixed texts from bilingual distributed representations that we exploit for improving language model performance. In particular, armed with this additional data, we adopt a curriculum learning approach where we first finetune the language models on synthetic data then on gold code-mixed data. We find that, although simple, our synthetic code-mixing method is competitive with (and in some cases is even superior to) several standard methods (backtranslation, method based on equivalence constraint theory) under a diverse set of conditions. Our work shows that the mT5 model, finetuned following the curriculum learning procedure, achieves best translation performance (12.67 BLEU). Our models place first in the overall ranking of the English-Hinglish official shared task.

Masked language models (MLM) have become an integral part of many natural language processing systems. Although multilingual MLMs have been introduced to serve many languages, these have limitations as to their capacity and the size and diversity of non-English data they are pre-trained on. In this work, we remedy these issues for Arabic by introducing two powerful deep bidirectional transformer-based models, AR-BERT and MARBERT, that have superior performance to all existing models. To evaluate our models, we propose ArBench, a new benchmark for multi-dialectal Arabic language understanding. ArBench is built using 41 datasets targeting 5 different tasks/task clusters, allowing us to offer a series of standardized experiments under rich conditions. When fine-tuned on ArBench, ARBERT and MAR-BERT collectively achieve new SOTA with sizeable margins compared to all existing models such as mBERT, XLM-R (Base and Large), and AraBERT on 37 out of 45 classification tasks on the 41 datasets (%82.22). Our models are publicly available for research.

Measuring the amount of shared information between two documents is a key to address a number of Natural Language Processing (NLP) challenges such as Information Retrieval (IR), Semantic Textual Similarity (STS), Sentiment Analysis (SA) and Plagiarism Detection (PD). In this paper, we report a plagiarism detection system based on two layers of assessment: 1) Fingerprinting which simply compares the documents fingerprints to detect the verbatim reproduction; 2) Word embedding which uses the semantic and syntactic properties of words to detect much more complicated reproductions. Moreover, Word Alignment (WA), Inverse Document Frequency (IDF) and Part-of-Speech (POS) weighting are applied on the examined documents to support the identification of words that are most descriptive in each textual unit. In the present work, we focused on Arabic documents and we evaluated the performance of the system on a data-set of holding three types of plagiarism: 1) Simple reproduction (copy and paste); 2) Word and phrase shuffling; 3) Intelligent plagiarism including synonym substitution, diacritics insertion and paraphrasing. The results show a recall of 88% and a precision of 86%. Compared to the results obtained by the systems participating in the Arabic Plagiarism Detection Shared Task 2015, our system outperforms all of them with a plagiarism detection score (Plagdet) of 83%.