Lyndon Nixon

In this paper we show how the performance of tweet clustering can be improved by leveraging character-based neural networks. Theproposedapproachovercomes the limitations related to the vocabulary explosion in the word-based models and allows for the seamless processing of the multilingual content. Our evaluation results and code are available on-line.

This research note reports on the first, to the best of the author’s knowledge, release of an online image annotation service for destination image measurement. Destination Marketing Organisations (DMOs) today, while actively data mining textual content for insights into visitor sentiment towards their destination or the most popular topics or themes of visitors at that destination, increasingly face usage of digital imagery or videos - yet non-textual content is not as easily ‘understood’ by machines to provide the same insights. The recent emergence of online services for image annotation might be of value to DMOs and researchers but their genericity means that to date e-tourism researchers continue to use manual approaches to media annotation, unable to scale to larger data sets and inconsistent across efforts with respect to chosen visual categories and concepts. We present here initial results which indicate that researchers and organisations could use an online service tuned specifically to the detection of visual concepts related to destination image, allowing them to annotate media at greater scale and analyse and compare results according to a common annotation vocabulary, helping us progress further in this exciting new area of e-tourism research and Tourism Intelligence.

The measurement of destination image from visual media such as online photography is of growing significance to destination managers and marketers who want to make better decisions and attract more visitors to their destination. However, there is no single approach with proven accuracy for doing this. We present a new approach where we fine-tune a deep learning model for a predetermined set of cognitive attributes of destination image. We then train state of the art neural networks using labelled tourist photography and test accuracy by comparing results with a ground truth dataset built for the same set of visual classes. Comparing our fine-tuned model against results which follow past approaches, we demonstrate that the pre-trained models without fine-tuning are not as accurate in capturing all of the destination image’s cognitive attributes. This is, to the best of our knowledge, the first deep learning computer vision model trained specifically to measure the cognitive component of destination image from photography and can act as a benchmark for future systems.

Destination marketers are aware that online communication about their destination is increasingly dependent on visual media rather than text, due to the growing popularity of social networks such as Instagram. An accurate understanding of how the destination is being presented to users in this medium is critical for digital marketing activities, e.g. to know if the desired destination brand is present or if visitors focus on other aspects of the destination than those being promoted in marketing. Unlike text mining, which has well established techniques to extract keywords and associations from text corpora, a consistent approach to understanding the content of images and expressing the resulting destination brand is lacking. This paper presents a visual classifier trained and fine-tuned specifically for destination brand measurement from images using 18 visual classes. It presents an exploratory study of how different destinations are being presented visually on Instagram and discusses how these insights could be used by destination marketers to adapt and improve their digital marketing.

This presentation considers if, and to what extent, social media can change the viewer’s image of a tourism destination as well as which types of visual content are most effective. The results from an online survey, which compared three different test groups and their image of Mexico City, showed that UGC images from Instagram, as well as random Google images, were more effective at improving destination image than UGC images reposted by a DMO. Additionally, the study used image annotation to determine which features in images were most important in terms of their contribution to an improvement in overall destination image.

This work in progress reports on ongoing experimentation with machine learning approaches on time series data, where the time series is a quantification of the success of content about a certain topic published on a certain digital channel over a past time period. The experiment tests how accurate predictive analytical approaches can be to predict the future success of a piece of media content published on the Web or social media platform according to its topics. Our intention is to enable a new innovation in media organizations’ content publication strategies, where the choice of media for a future publication can be informed by such predictive capabilities in order to maximize the potential content's reach to a digital audience.

TV broadcasters and other organizations with online media collections which wish to extend the reach of and engagement with their media assets conduct digital marketing activities. The marketing success depends on the relevance of the topics of the media content to the audience, which is made even more difficult when planning future marketing activities as one needs to know the topics that the future audience will be interested in. This paper presents the innovative application of AI based predictive analytics to identify the topics that will be more popular among future audiences and its use in a digital content marketing strategy of media organisations.

Destination promotion videos (DPVs) are increasingly being used for online marketing and seen by travellers during the information search process. Yet, scholarly attention to DPVs is scarce and the research question of “how do DPVs influence viewers’ destination image change?” is unresolved. To fill these voids, this study (1) examines the projected image of Macau based on the video content analysis of their latest DPV; and (2) investigates the impact of viewing a DPV on viewers’ perceived destination image and on their behavioural intention to visit Macau. The efficacy of repeating—a framing method introduced by Entman (1991)—in influencing change in the DPV viewers’ image of the destination is highlighted. Findings from the experiment indicate that the content of DPVs and repeating certain shots are effective in positively enhancing travellers’ perceived destination image as well as triggering potential travellers’ interest in further researching and visiting the destination.

To optimise digital content marketing for broadcasters, the Horizon 2020 funded ReTV project developed an end-to-end process termed “Trans-Vector Publishing” and made it accessible through a Web-based tool termed “Content Wizard”. This paper presents this tool with a focus on each of the innovations in data and AI-driven media analysis to address each key step in the digital content marketing workflow: topic selection, content search and video summarisation. First, we use predictive analytics over online data to identify topics the target audience will give the most attention to at a future time. Second, we use neural networks and embeddings to find the video asset closest in content to the identified topic. Third, we use a GAN to create an optimally summarised form of that video for publication, e.g. on social networks. The result is a new and innovative digital content marketing workflow which meets the needs of media organisations in this age of interactive online media where content is transient, malleable and ubiquitous.

This survey considers the vision of TV broadcasting where content is personalised and personalisation is data-driven, looks at the AI and data technologies making this possible and surveys the current uptake and usage of those technologies. We examine the current state-of-the-art in standards and best practices for data-driven technologies and identify remaining limitations and gaps for research and innovation. Our hope is that this survey provides an overview of the current state of AI and data-driven technologies for use within broadcasters and media organisations. It also provides a pathway to the needed research and innovation activities to fulfil the vision of data-driven personalisation of TV content.

The correct classification of the 17 Sustainable Development Goals (SDG) proposed by the United Nations (UN) is still a challenging and compelling prospect due to the Shared Task’s imbalanced dataset. This paper presents a good method to create a baseline using RoBERTa and data augmentation that offers a good overall performance on this imbalanced dataset. What is interesting to notice is that even though the alignment between synthetic gold and real gold was only marginally better than what would be expected by chance alone, the final scores were still okay.

Annotation styles express guidelines that direct human annotators in what rules to follow when creating gold standard annotations of text corpora. These guidelines not only shape the gold standards they help create, but also influence the training and evaluation of Named Entity Linking (NEL) tools, since different annotation styles correspond to divergent views on the entities present in the same texts. Such divergence is particularly present in texts from the media domain that contain references to creative works. In this work we present a corpus of 1000 annotated documents selected from the media domain. Each document is presented with multiple gold standard annotations representing various annotation styles. This corpus is used to evaluate a series of Named Entity Linking tools in order to understand the impact of the differences in annotation styles on the reported accuracy when processing highly ambiguous entities such as names of creative works. Relaxed annotation guidelines that include overlap styles lead to better results across all tools.

This paper explores the current landscape of visualizing large language models (LLMs). The main objective was threefold. Firstly, we investigate how we can visualize LLM-specific techniques such as prompt engineering, instruction tuning, or guidance. Secondly, LLM causality, interpretability, and explainability are examined through visualization. And finally, we showcase the role of visualization in illuminating the integration of multiple modalities. We are interested in discovering the papers that present visualization systems instead of those that use visualization to showcase a part of their work. Our survey aims to synthesize the state-of-the-art in LLM visualization, offering a compact resource for exploring future research avenues.

As global travel emerges from the pandemic, pent up interest in travel will lead to consumers making their choice between global destinations. Instagram is a key source of destination inspiration. DMO marketing success on this channel relies on projecting a destination image that resonates with this target group. However, usual text-based marketing intelligence on this channel does not work as content is consumed first and foremost as a visual projection. The author has built a deep learning based visual classifier for destination image measurement from photos. In this paper, we compare projected and perceived destination images in Instagram photography for four of the most Instagrammed destinations worldwide. We find that whereas the projected destination image aligns well to the perceived image, there are specific aspects of the destinations that are of more interest to Instagrammers than reflected in the current destination marketing.

Destination positioning refers to destinations identifying their most distinct attributes and focusing on these in their marketing activities in order to distinguish themselves from competitors, develop a brand identity and highlight uniqueness. In this paper, we consider 9 European capitals and analyse their visual marketing on Instagram to identify how truly distinct their destinations are being presented online. By comparing between them as well as comparing to the perceived destination image measured from visitor photos on the same platform, we present a methodology for identifying each destinations distinct attributes and measuring how well DMOs are positioning themselves with respect to competing destinations, with recommendations for improving their positioning.

The Social Web is increasingly taking up the daily time of consumers and is becoming a primary source of impressions about tourism destinations. The recent shift towards more visual content, as evidenced in the fast growing social network Instagram being largely a photo sharing site, means that DMOs need to consider how photos of their destinations can be influencing consumers’ destination image. We present what may be the first study on how the selection of images from a social media site (Instagram) to promote a destination can be used to influence destination image. As a basis for our study, we have selected the DMO channels of Jordan and Costa Rica in Instagram. Through focus groups and a Likert-scale survey, we draw first conclusions on which types of photos are most effective to positively promote a destination and how the consumers’ previous image of a destination could affect this

Background. Sentic computing relies on welldefined affective models of different complexity - polarity to distinguish positive and negative sentiment,
for example, or more nuanced models to capture expressions of human emotions. When used to measure
communication success, even the most granular affective model combined with sophisticated machine learning approaches may not fully capture an organisation’s
strategic positioning goals. Such goals often deviate from
the assumptions of standardised affective models. While
certain emotions such as Joy and Trust typically represent desirable brand associations, specific communication goals formulated by marketing professionals often
go beyond such standard dimensions. For instance, the
brand manager of a television show may consider fear
or sadness to be desired emotions for its audience.
Method. This article introduces expansion techniques
for affective models, combining common and commonsense knowledge available in knowledge graphs with
language models and affective reasoning, improving coverage and consistency as well as supporting domainspecific interpretations of emotions.
Results and Conclusions. An extensive evaluation
compares the performance of different expansion techniques: (i) a quantitative evaluation based on the revisited Hourglass of Emotions model to assess perfor
mance on complex models that cover multiple affective categories, using manually compiled gold standard
data, and (ii) a qualitative evaluation of a domainspecific affective model for television programme brands.
The results of these evaluations demonstrate that the
introduced techniques support a variety of embeddings
and pre-trained models. The paper concludes with a
discussion on applying this approach to other scenarios
where affective model resources are scarce.

In order to optimise the distribution of video assets online, media organizations need tailor their offerings for specific digital channels and better understand the interests of their audiences at particular points in time, which are often influenced by contemporary new stories and trends on social media. For this purpose, the research project ReTV has developed a Web-based tool termed ’Content Wizard’ which demonstrates an end-to-end, semi-automated workflow for video content creation, adaptation and distribution across digital channels. Digital assets can be selected based on predicted future trending topics, re-purposed according to the different digital channels they will be published upon and scheduled for the optimal future publication date. The result is an innovative video publication workflow that meets the marketing needs of media organisations in this age of transient online media spread across multiple channels.

This paper describes a solution for synchronized delivery of Web content related to objects and topics present in a parallel online television programme, which we call "Linked Television". A content platform prepares the TV programme by analyzing, annotating and enriching it with links to Web content through a combination of innovative Web services. An editor tool allows manual correction and completion of the enrichment. A Web-based player allows multiple devices to synchronise the video and its related content across different screens. The result is a richer TV experience for the "second screen" generation who, initiated by TV viewing, like to explore further content online. We back this up by pilots using news and cultural heritage programming which have been validated in trials by viewers as enhancing their TV experience.

The news media landscape tends to focus on long-running narratives. Correctly processing new information, therefore, requires considering multiple lenses when analyzing media content. Traditionally it would have been considered sufficient to extract the topics or entities contained in a text in order to classify it, but today it is important to also look at more sophisticated annotations related to fine-grained geolocation, events, stories and the relations between them. In order to leverage such lenses we propose a new corpus that offers a diverse set of annotations over texts collected from multiple media sources. We also showcase the framework used for creating the corpus, as well as how the information from the various lenses can be used in order to support different use cases in the EU project InVID for verifying the veracity of online video.