Speakers of widespread languages may encounter problems in information retrieval and document understanding when they access documents in the same language from another country. The work described here focuses on the development of resources to support improved document retrieval and understanding by users of Modern Standard Arabic (MSA). The lexicon of an Egyptian Arabic speaker and the lexico...

The automatic recognition of foreign-accented Arabic speech is a challenging task since it involves a large number of nonnative accents. As well, the nonnative speech data available for training are generally insufficient. Moreover, as compared to other languages, the Arabic language has sparked a relatively small number of research efforts. In this paper, we are concerned with the problem of n...

To date, majority of research for Arabic Named Entity Recognition (NER) addresses the task for Modern Standard Arabic (MSA) and mainly focuses on the newswire genre. Despite some common characteristics between MSA and Dialectal Arabic (DA), the significant differences between the two language varieties hinder such MSA specific systems from solving NER for Dialectal Arabic. In this paper, we pre...

Arabs throughout the Arab world speak different dialects of Arabic in their daily conversations. We envision collecting a data set that maps different Arabic variations and dialects to Modern Standard Arabic (MSA). These mappings can be then used to facilitate the communication among Arabs from different regions. In this work, we developed a Game With A Purpose (GWAP) to collect mappings betwee...

Abstract The Chinese writer Lu Xun (1881-1936) is considered one of the most important writers in history literature. His works have been translated over years into many foreign languages, including Arabic. In addition to translating his works, researchers presented studies on thought and literature, Xun’s translations Arabic played an role introducing creativity even modern literature general,...

We describe the CMU submission for the 2014 shared task on language identification in code-switched data. We participated in all four language pairs: Spanish–English, Mandarin–English, Nepali–English, and Modern Standard Arabic–Arabic dialects. After describing our CRF-based baseline system, we discuss three extensions for learning from unlabeled data: semi-supervised learning, word embeddings,...

Arabic Dialects present many challenges for machine translation, not least of which is the lack of data resources. We use crowdsourcing to cheaply and quickly build LevantineEnglish and Egyptian-English parallel corpora, consisting of 1.1M words and 380k words, respectively. The dialectal sentences are selected from a large corpus of Arabic web text, and translated using Amazon’s Mechanical Tur...