Search

Unsupervised Machine Translation hasbeen advancing our ability to translatewithout parallel data, but state-of-the-artmethods assume an abundance of mono-lingual data. This paper investigates thescenario where monolingual data is lim-ited as well, finding that current unsuper-vised methods suffer in performance un-der this stricter setting. We find that theperformance loss originates from the poorquality of the pretrained monolingual em-beddings, and we propose using linguis-tic information in the embedding train-ing scheme. To support this, we look attwo linguistic features that may help im-prove alignment quality: dependency in-formation and sub-word information. Us-ing dependency-based embeddings resultsin a complementary word representationwhich offers a boost in performance ofaround 1.5 BLEU points compared to stan-dardWORD2VECwhen monolingual datais limited to 1 million sentences per lan-guage. We also find that the inclusion ofsub-word information is crucial to improv-ing the quality of the embeddings

O conceito de desenvolvimento participativo e o discurso da participação dos vários atores da sociedade civil tem vindo a ganhar uma crescente importância ao longo das últimas décadas, sobretudo no contexto de intervenção social. De forma a compreender melhor o conceito de participação nas suas diferentes formas e dimensões, têm sido desenvolvidas várias tipologias que permitem medir e analisar os níveis de participação. A participação de vários atores nos processos de planeamento estratégico e na definição de políticas públicas tem-se afirmado como um desafio para os Estados que promovem este tipo de processos participativos, levando-os a desenvolver estratégias e instrumentos para levar a cabo estes processos. O Programa Rede Social surgiu em Portugal como um exemplo de um programa estatal que privilegia a participação de diversos atores sociais em processos de planeamento estratégico. Desta forma, pretende-se com esta dissertação compreender de que forma o Programa Rede Social, no concelho de Sintra, tem contribuído para o envolvimento dos atores sociais neste tipo de processos, privilegiando uma abordagem metodológica qualitativa. De forma a analisar os níveis de participação dos membros da Rede Social neste programa, a presente dissertação baseou-se na análise modelo da "escada de participação" de Sherry Arnstein. The concept of participatory development and the discourse of participation of several civil society actors as gain an increasing importance over the last decades, particularly in the context of social intervention. In order to better understand the concept of participation in its various forms and dimensions, several typologies have been developed to measure and analyze the levels of participation. The participation of several actors in the processes of strategic planning and in the definition of public policies has established itself as a challenge to States that promote this type of participatory processes. The Program "Rede Social" emerged in Portugal as an example of a state program that requires the participations of different social actors in strategic planning processes. With this thesis, we aim to understand how the Program "Rede Social", in Sintra, has contributed to the involvement of social actors in such processes, privileging a qualitative approach. In order to analyze the levels of participation of the members of the Program "Rede Social", this thesis based itself on the analysis of Sherry Arnstein's model "ladder of participation".

This paper presents a user evaluation of machine translation and post-editing for TV subtitles. Based on a process study where 12 professional subtitlers translated and post-edited subtitles, we compare effort in terms of task time and number of keystrokes. We also discuss examples of specific subtitling features like condensation, and how these features may have affected the post-editing results. In addition to overall MT quality, segmentation and timing of the subtitles are found to be important issues to be addressed in future work.

This paper updates the progress made on the PRINCIPLE project, a 2-year action funded by the European Commission un-der the Connecting Europe Facility (CEF) programme. PRINCIPLE focuses on col-lecting high-quality language resources for Croatian, Icelandic, Irish and Norwe-gian, which have been identified as low-resource languages, especially for build-ing effective machine translation (MT) systems. We report initial achievements of the project and ongoing activities aimed at promoting the uptake of neural MT for the low-resource languages of the project.