Fiammetta Marulli

Cultural Heritage and Social Pulse: A Semantic Approach for CH Sensitivity Discovery in Social Media Data

The demand for real-time business intelligence and the popularity of social media offer room for synthesis. A recent topic of interest in such perspective is Cultural Heritage, as empirically evidenced, in 2015, by Twitter introduction of the hashtag #culturalheritage. The opportunities offered by linking both concepts are acknowledged in the recent scientific literature, but a relative underexposure of quantitative studies devoted to assess the potential and effectiveness of social pulses and activities for supporting organizations promoting Cultural Heritage was recorded. In this work a data driven approach is proposed. Basing on the effective employment of domain specific and general ontologies, it aims to identify a set of Key Performance Indicators (KPIs) for quantitative estimation of Cultural Heritage Sensitivity as expressed by social network users. Quantitative analysis of huge datasets of tweets, combining Natural Language Processing (NLP), semantic technologies, geo-referencing and temporal analysis, and issued in a long period of time from geographical areas of Italy having different densities of CH resources, were performed. Examples of computed measures for characterizing peoples interest and sensitivity into CH subjects, are geographical density of CH resources and temporal proximity to CHrelated events. Obtained results encourage a Business-Intelligence approach.

IoT to Enhance Understanding of Cultural Heritage: Fedro Authoring Platform, Artworks Telling Their Fables

Cultural Heritage has got great importance in recent years, in order to preserve countries history and traditions and to support social and economic improvements. Typical IoT smart technologies represent an effective mean to support understanding of Cultural Heritage, by their capability to involve different users and to catch their explicit and implicit preferences, behaviors and contributions. This paper presents FEDRO, an authoring platform, as part of the intelligent infrastructures developed in DATABENC to support a cultural exhibition of “talking” sculptures held in the Southern Italy, in 2015. FEDRO aims to automatically generate textual and users profiled artworks biographies, employed to feed a smart app for guiding visitors during the exhibition. A preliminary experimentation revealed a tangible improvement in the users’ experience appreciation during the visit. Quality estimations of generated output were also computed exploiting users’ feedbacks, collected through a manual questionnaire, subscribed at the end of their visit.

An Associative Engines Based Approach Supporting Collaborative Analytics in the Internet of Cultural Things

In this paper we illustrate an integrated approach supporting an information system which combines Business Intelligence, Big Data, Internet of Things, GeoSpatial information processing, multimedia resources, structured and unstructured content analysis with Semantic techniques, and Social Network Analysis. This system exploits the specific opportunities provided by Associative in-memory technologies in the context of Cloud Computing and Collaborative processes and experiments the use of semantic technologies to mitigate the efforts involved in merging, and analyzing information incoming from different sources. The primary aim is supporting Cultural Heritage (CH) Asset crowdsourcing, promotion, publication, management and usage. Preliminary results are reported from ongoing experimentation, focused on visitors interests and behaviour analysis performed by monitoring, comparing and combining information from different types of populations and visits: on-site ad-hoc (exhibitions, museums, cultural events, etc.), territorial (eg. Historical downtown, touristic areas including relevant CH resources) and Virtual - Worldwide Internet based visits.

A novel approach for automatic text analysis and generation for the cultural heritage domain

Knowledge is information that has been contextualised in a certain domain, to be used or applied. It represents the basic core of our Cultural Heritage and Natural Language provides us with prime versatile means of construing experience at multiple levels of organization. The natural language generation field consists in the creation of texts providing information contained in other kind of sources (numerical data, graphics, taxonomies and ontologies or even other texts), with the aim of making such texts indistinguishable, as far as possible, from those created by humans. On the other hand, the knowledge extraction, basing on text mining and text analysis tasks, as examples of the many applications born from computational linguistic, provides summarization, categorization, topics extractions from textual resources using linguistic concepts, which deal with the imprecision and ambiguity of human language. This paper presents a research activity focused on exploring and scientifically describing knowledge structure and organization involved in textual resources’ generation. Thus, a novel multidimensional model for the representation of conceptual knowledge, is proposed. Furthermore, a real case study in the Cultural Heritage domain is described to demonstrate the effectiveness and the feasibility of the proposed model and approach.

What's the Matter with Cultural Heritage Tweets? An Ontology -- Based Approach for CH Sensitivity Estimation in Social Network Activities

In 2015, Twitter introduced the hashtag #culturalheritage, thus giving an empirical evidence of the increasing interest of users towards cultural topics and events. In the last years, while an increasing number of studies and initiatives about tweets, in both the academic and business worlds, raised, a lack of quantitative studies devoted to assess their potential and effectiveness for organizations promoting Cultural Heritage was recorded. So, this work describes a quantitative analysis of tweets which combines NLP, semantic technologies, geo-referencing and temporal analysis. The initial set of measures aims to characterize peoples interest and sensitivity into CH subjects, geographical density of CH resources, and temporal proximity to CH-related events. Furthermore, in order to evidence the relevance of the obtained results, they are compared to similar measures computed for different but more general topics -- such as Medicine - while at the same time entail a certain specificity of interest, which cannot be confused with reactions to common mainstream, glamour or massive-impressive events (earthquakes, political elections, wars, amazing news). This kind of analysis was focused on huge datasets of tweets, issued in a long period of time from geographical areas of Italy having different densities of CH resources, The results encourage and sustain a Business-Intelligence approach which is suitable for both no-profit ad business oriented organizations, such as those involved in the DATABENC District.

The Imitation Game to Cultural Heritage: A Human-like Interaction Driven Approach for Supporting Art Recreation

Smart IoT technologies set a milestone in supporting new enjoyment models for Art and Cultural Heritage, providing amazing technological experiences. However, users, while interacting almost purely by mediation of smart devices or augmented VR displays, practically keep themselves out from living the fullness of the surrounding cultural sites experience, establishing no direct dialogues or interactions with artworks. It sounds quite like to be in the living-room, looking at very appealing documentaries, equipped with exciting smart technologies. This paper focuses on the importance of “re-humanize” art recreation models, proposing a human-like interaction driven approach. Holographic projections, reproducing human or fantasy characters, play the human presence imitation game, when users are detected close to any artwork, interacting and dialoguing with them in natural language. An IoT infrastructure, an NLP platform and a Holographic Projection Engine implement a system for supporting holographic projections. Preliminary experiments were promising, thus motivating authors to further investigations.

A Cloud-based Approach for Analyzing Viral Propagation of Linguistic Deviations by Social Networking: Current Challenges and Pitfalls for Text Analysis Tools

Social Networks activities offer rooms a non-trivial testbed for linguistic analysis, introducing significant revolutions into the creation of textual content. Linguistic solecisms, blunders, and generally speaking deviations from standard linguistic norms, are becoming the rule rather than the exception. Social Networks instantly and virally propagate deviations among users, who are increasingly moving away from standard language usage. Performing text analysis on deviated textual documents is a challenging and hard task. In this work, we propose an approach supporting text analysis tasks against a set of deviated textual documents. It exploits a “linguistic blundersonomy”, a taxonomy of linguistic deviations, progressively built by processing textual Big Data provided by social network, in a Cloud-Based environment (SAP-HANA). A preliminary case study for Italian language is presented, showing how the exploitation of a linguistic blundersonomy could improve the precision of a sentiment and opinion mining process, and more generally, of a text analysis process.