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.

Evolution of User-generated Translation: Fansubs, Translation Hacking and Crowdsourcing

Abstract: Most conspicuous initially with Japanese anime fansubs, fan-based translation has been developing since the 1980s. In the advent of widespread availability of Web 2.0 platforms, fan translation has become a global phenomenon despite its dubious legal status. Extending the emerging interest in fansubs and scanlation in translation studies to the little discussed translation hacking by video game fans, this article brings readers‘ attention to participatory culture manifest in user-generated content in the field of translation and localisation. The article describes the evolution from unsolicited fan translation to solicited community translation now called crowdsourcing and considers them in the framework of user-generated translation (UGT). The article provides interdisciplinary perspectives, drawing on insights from media and game studies to address UGT which could profoundly impact the profession of translation and localisation as new technological environments unleash the technical competence, genre-knowledge and unparalleled devotion of the otherwise untrained Internet crowd as translators.This article is made available under a CC BY-NC 4.0 license.

Lexically Constrained Decoding for Sequence Generation Using Grid Beam Search

Abstract: We present Grid Beam Search (GBS), an algorithm which extends beam search to allow the inclusion of pre-specified lexical constraints. The algorithm can be used with any model which generates sequences token by token. Lexical constraints take the form of phrases or words that must be present in the output sequence. This is a very general way to incorporate auxillary knowledge into a model’s output without requiring any modification of the parameters or training data. We demonstrate the feasibility and flexibility of Lexically Constrained Decoding by conducting experiments on Neural Interactive-Predictive Translation, as well as Domain Adaptation for Neural Machine Translation. Experiments show that GBS can provide large improvements in translation quality in interactive scenarios, and that, even without any user input, GBS can be used to achieve significant gains in performance in domain adaptation scenarios.

Advanced printing and deposition methodologies for the fabrication of biosensors and biodevices.

Abstract: Advanced printing and deposition methodologies are revolutionising the way biological molecules are deposited and leading to changes in the mass production of biosensors and biodevices. This revolution is being delivered principally through adaptations of printing technologies to device fabrication, increasing throughputs, decreasing feature sizes and driving production costs downwards. This review looks at several of the most relevant deposition and patterning methodologies that are emerging, either for their high production yield, their ability to reach micro- and nano-dimensions, or both. We look at inkjet, screen, microcontact, gravure and flexographic printing as well as lithographies such as scanning probe, photo- and e-beam lithographies and laser printing. We also take a look at the emerging technique of plasma modification and assess the usefulness of these for the deposition of biomolecules and other materials associated with biodevice fabrication.

Absorbance Based Light Emitting Diode Optical Sensors and Sensing Devices.

Abstract: The ever increasing demand for in situ monitoring of health, environment and security has created a need for reliable, miniaturised sensing devices. To achieve this, appropriate analytical devices are required that possess operating characteristics of reliability, low power consumption, low cost, autonomous operation capability and compatibility with wireless communications systems. The use of light emitting diodes (LEDs) as light sources is one strategy, which has been successfully applied in chemical sensing. This paper summarises the development and advancement of LED based chemical sensors and sensing devices in terms of their configuration and application, with the focus on transmittance and reflectance absorptiometric measurements.

Hysteresis and the E-to-H transition in radiofrequency inductive discharges

Abstract: Typical inductive discharges, such as are used for plasma processing, exhibit two modes of operation: the true inductive discharge known as the H mode, and a weak capacitive discharge known as the E mode. Experimentally, the transition between these modes as the coil current is increased is clear and is marked by a large increase in discharge power, plasma density and optical emission occurring as the H mode appears. According to simple theory, this transition and the reverse transition occur at a single well defined current. In practice, this is usually not the case. The E-to-H transition occurs at a larger coil current than the H-to-E transition, and a range of currents between these values supports either E or H mode. This effect is called hysteresis. In this paper we show that hysteresis can be understood to arise from nonlinear effects, most notably in the electron power balance equation. We survey various mechanisms that can produce hysteresis and attempt to provide quantitative estimates of their significance.