The Surface Realization Shared Tasks of 2018 and 2019 were Natural Language Generation shared tasks with the goal of exploring approaches to surface realization from Universal-Dependency-like trees to surface strings for several languages.In the 2018 shared task there was very little difference in the absolute performance of systems trained with and without additional, synthetically created data, and a new rule prohibiting the use of synthetic data was introduced for the 2019 shared task.Contrary to the findings of the 2018 shared task, we show, in experiments on the English 2018 dataset, that the use of synthetic data can have a substantial positive effect -an improvement of almost 8 BLEU points for a previously state-of-the-art system.We analyse the effects of synthetic data, and we argue that its use should be encouraged rather than prohibited so that future research efforts continue to explore systems that can take advantage of such data.

Neural Natural Language Generation (NLG) systems are well known for their unreliability.To overcome this issue, we propose a data augmentation approach which allows us to restrict the output of a network and guarantee reliability.While this restriction means generation will be less diverse than if randomly sampled, we include experiments that demonstrate the tendency of existing neural generation approaches to produce dull and repetitive text, and we argue that reliability is more important than diversity for this task.The system trained using this approach scored 100% in semantic accuracy on the E2E NLG Challenge dataset, the same as a template system.

In natural language generation (NLG), the task is to generate utterances from a more abstract input, such as structured data.An added challenge is to generate utterances that contain an accurate representation of the input, while reflecting the fluency and variety of human-generated text.In this paper, we report experiments with NLG models that can be used in task oriented dialogue systems.We explore the use of additional input to the model to encourage diversity and control of outputs.While our submission does not rank highly using automated metrics, qualitative investigation of generated utterances suggests the use of additional information in neural network NLG systems to be a promising research direction.

‘Big Data’ can mean something quite different in the context of Humanities. The way Humanities scholars frame their inquiries often leverages collections that are an order of magnitude smaller than the full, industrial scale, there is significant value to be found in the Humanistic sense of ‘Big’. In particular, the variety of the data, and the richness of the explorations, means that high-quality knowledge systems are required. More meaning is needed than the surface analytics often demonstrated in other ‘Big’ scenarios. This paper examines how a specific collection related to the 1916 Rising in Ireland was analyzed. The result was a process to extract entities that underpinned a highly-effective personalized knowledge-driven exploration of that collection by users. It demonstrates the mutual benefit of natural language at scale with rich humanistic inquiry to communicate improved experiences for a much broader range of users than would otherwise be possible.

Commons-based Peer Production is the process by which internet communities create media and software artefacts. Learning is integral to the success of these communities as it encourages contribution on an individual level, helps to build and sustain commitment on a group level and provides a means for adaption at an organisational level. While some communities have established ways to support organisational learning -- through a forum or thread reserved for community discussion -- few have investigated how more in-depth visual and analytic interfaces could help formalise this process. In this paper, we explore how social network visualisation can be used to encourage reflection and thus support organisational learning in online communities. We make the following contributions: First, we describe Commons-Based Peer Production, in terms of a socio-technical learning system that includes individual, group and organisational learning. Second, we present a novel visualisation environment that embeds social network visualisation in an asynchronous collaborative architecture. Third, we present results from an evaluation and discuss the potential for visualisation to support the process of organisational reflection in online communities.