Dublin City University

About: Dublin City University is a education organization based out in Dublin, Ireland. It is known for research contribution in the topics: Machine translation & Laser. The organization has 5904 authors who have published 17178 publications receiving 389376 citations. The organization is also known as: National Institute for Higher Education, Dublin & DCU.

Blind and Anonymous Identity-Based Encryption and Authorised Private Searches on Public Key Encrypted Data

Abstract: Searchable encryption schemes provide an important mechanism to cryptographically protect data while keeping it available to be searched and accessed. In a common approach for their construction, the encrypting entity chooses one or several keywords that describe the content of each encrypted record of data. To perform a search, a user obtains a trapdoor for a keyword of her interest and uses this trapdoor to find all the data described by this keyword. We present a searchable encryption scheme that allows users to privately search by keywords on encrypted data in a public key setting and decrypt the search results. To this end, we define and implement two primitives: public key encryption with oblivious keyword search (PEOKS) and committed blind anonymous identity-based encryption (IBE). PEOKS is an extension of public key encryption with keyword search (PEKS) in which users can obtain trapdoors from the secret key holder without revealing the keywords. Furthermore, we define committed blind trapdoor extraction, which facilitates the definition of authorisation policies to describe which trapdoor a particular user can request. We construct a PEOKS scheme by using our other primitive, which we believe to be the first blind and anonymous IBE scheme. We apply our PEOKS scheme to build a public key encrypted database that permits authorised private searches, i.e., neither the keywords nor the search results are revealed.

Exploiting Cross-Sentence Context for Neural Machine Translation

Abstract: In translation, considering the document as a whole can help to resolve ambiguities and inconsistencies. In this paper, we propose a cross-sentence context-aware approach and investigate the influence of historical contextual information on the performance of neural machine translation (NMT). First, this history is summarized in a hierarchical way. We then integrate the historical representation into NMT in two strategies: 1) a warm-start of encoder and decoder states, and 2) an auxiliary context source for updating decoder states. Experimental results on a large Chinese-English translation task show that our approach significantly improves upon a strong attention-based NMT system by up to +2.1 BLEU points.

Learning Morphology with Morfette

Abstract: Morfette is a modular, data-driven, probabilistic system which learns to perform joint morphological tagging and lemmatization from morphologically annotated corpora. The system is composed of two learning modules which are trained to predict morphological tags and lemmas using the Maximum Entropy classifier. The third module dynamically combines the predictions of the Maximum-Entropy models and outputs a probability distribution over tag-lemma pair sequences. The lemmatization module exploits the idea of recasting lemmatization as a classification task by using class labels which encode mappings from word forms to lemmas. Experimental evaluation results and error analysis on three morphologically rich languages show that the system achieves high accuracy with no language-specific feature engineering or additional resources.