A Knowledge Base for Cultural Heritage Protection against Climate Change
Author(s): Jürgen Moßgraber, Désirée Hilbring, Paraskevi Pouli, Giuseppina Padeletti
Full text: submitted version
Abstract: Environmental factors, worsened by the increasing climate change impact, represent significant threats to European Cultural Heritage (CH) assets. In Europe, the huge number and diversity of CH assets, together with the different climatological sub-regions picture as well as the different adaptation policies to Climate Change adopted (or to be adopted) by the different Nations, generates a very complex scenario. The approach will benefit of a multidisciplinary methodology that will bridge the gap between the two different worlds: the CH stakeholders and the scientific/technological experts since protecting cultural heritage assets and increasing their resilience against effects caused by the climate change is a multidisciplinary task. Experts from many domains need to work together to meets their conservation goals. This paper presents means for facilitating the work for the different experts. A new Ontology has been designed integrating all necessary aspects for improving the resilience of cultural heritages on site. This ontology combines the topics: Cultural Heritage Assets, Stakeholders and Roles, Climate and Weather Effects, Risk Management, Conservation Actions, Materials, Sensors, Models and Observations. Furthermore, a knowledge base has been created, which is based on the developed ontology. It provides different tools for the expert users of the different domains to connect and exchange their information to provide the basis for decision makers to decide, in which ways to protect cultural heritage from climate change effects.
Keywords: Ontology; Knowledge Base; Ontology Visualization; Cultural Heritage
Review 1 (by anonymous reviewer)
This paper presents the design of an ontology in the Cultural Heritage domain, meant to be used in a Knowledge Base for (mostly weather-related) risk assessment. The description of Cultural Heritage classified buildings is integrated with situational information and coupled with sensor data, also organized around a semantic schema. The paper presents an interesting project, using novel Semantic Web technologies to enable access to and sharing of risk assessment data, for specialists in different fields of research. One main claim is that the system is meant to prevent miscommunication between specialists across different fields of science, who have to work together in the context of the preservation of Cultural Heritage buildings. The platform grants access to information that could come from different fields of science, but does not provide a translation service across them or a vocabulary integration service, so - Miscommunication can still occur - It is not clear whether the semantic representation layer over the sensor data integrates reasoning over threshold values to give non-specialist an alert over a possible risk that they would not be able to pick up form raw sensor data (which could come from a domain of science they are not familiar with); these threshold values could depend on the type of sensor delivering the raw data, as many sensors delivering the same type of data still have different formats which are hard to integrate, it is a tricky business. About the model, the authors start the modeling phase from scratch because models such as CIDOC-CRM are judged to be too complex. The best of both worlds would probably be to see which classes from CIDOC-CRM correspond to the generic model that suits their needs (the classes of Man-Made Thing, Place, Event are defined there and are equivalent to some used in their proposed model) and either use them directly or at least link the locally-defined classes to the CIDOC-CRM ones as subclasses, for interoperability with other systems using CIDOC-CRM and for compatibility with standards. As a side note, Cultural Heritage Asset is not a Cultural Heritage, it is a Man-Made Thing, that has been granted a status by a Cultural Heritage representative decision. In the conclusion section, it is stated that users will evaluate ontology-based decision support system and recommendations: it would be very interesting to see concrete examples of data that will be evaluated, from the sensor generated representation to actual risk alerts or recommendations. Overall, the paper lacks concrete examples walking us through the system, with the eyes of different scientific specialists to showcase its additional value. Raising alerts based on weather and environmental conditions seems to be a very hard task and it would be great to show the limits within which the system works: what can it derive alerts or recommendations from, what is not integrated yet, what are the challenges of going from sensor data to human trigger? As for the paper’s structure, it is handy to refer to sections by their section identifier in the introduction (“in section 2” rather than “in section “Title””); also, it might be preferred to refer to ontology leave nodes as “leave nodes” rather than “at the bottom”. There are typos or shaky English sentences throughout the paper (example: “Assets are located in Sites, which in *their* turn can be classified…”, “Regardless the interdisciplinary meeting brought many new insights into the domain”-> into the project, which is across domains?; IoT is introduced without expanding the acronym first,...), it would be worth giving the paper one final proof editing.
Review 2 (by anonymous reviewer)
This paper introduces an ontology for modeling CH assets and climate change-related information and describes a knowledge base that has been created by employing such ontology. I think that the paper is well-written, and it tackles an interesting, concrete, and important problem. Personally, I find the examples useful to understand the possible applications of the ontology. However, I think that the paper would benefit from some more information about the ontology itself. I understand that describing 74 classes is beyond the scope of the paper, but some overview description would be useful. Likewise, the description of the knowledge base is extremely short, and focuses mainly on the interface, while I would appreciate reading more details and statistics about its content.
Review 3 (by anonymous reviewer)
Given the other reviewers' scores and the authors' rebuttal I stick with my original evaluation. --- The paper discusses the design of the HERACLES ontology, which aggregates multiple domain knowledge to support multidisciplinary collaboration in the area of CH preservation. The authors claim that the ontology supports experts with various expertise and from various disciplines to connect and exchange knowledge in the area of CH preservation and derive qualified decisions from the corresponding knowledge base. The paper is well written and easy to read. The line of argumentatuoin is clear and comprehensible. Nevertheless the paper has some flaws: The related work section is rather unspecific. It basically lists a number of projects and initiatives that deal with ICT support for CH preservation purposes but it does not sufficiently relate to the technologies applied and how they relate to the claims described in the paper. The methodology chosen to design the ontology as well as the ontology itself is described on a rather generic level. It would have been very interesting to get to know more about the design process and intentions for choosing specific ontologies as described in sect. 4. One of the major flaws of the paper is the missing evaluation. At the current stage of development it is hard to tell whether the ontology serves the claims described in the paper. This paper would be better suited for a poster presentation.
Review 4 (by Tomi Kauppinen)
This paper presents an approach to model concepts for modelling issues climate change might cause for cultural heritage. The paper looks at times like a project proposal rather than a report on work completed. Beyond that I have two main concerns about the paper: 1) The idea of the ontologies for the semantic web is essentially that ontologies are published on the (semantic) web. There are many existing mechanisms and tools for doing so. I strongly encourage authors to publish their work for instance with the help of LODE http://www.essepuntato.it/lode or some other alternative, automated way of sharing and documenting vocabularies and ontologies online. This would greatly facilitate giving feedback and essentially assessing the work by authors. 2) Secondly, and related to the first point, it was not clear from paper what kind of design decisions authors follow. For instance, why are the authors not using any support from top-level ontologies, and considered creating their ontology by reusing existing / developing new ontology design patterns? Authors talk about OGC but now it is not clear how sensor observations have been modelled beyond what is mentioned in Figure 4. One example of modelling sensor observations is a work a co-authored several years ago: https://www.researchgate.net/publication/235914188_Sensors_Tell_More_than_They_Sense_Modeling_and_Reasoning_about_Sensor_Observations_for_Understanding_Weather_Events I am not proposing here to follow the same approach but top-level structures can indeed help to overcome basic ontology engineering challenges. In figure 4, for instance, how are causesEffect and leadsToEffect different? I understand that in that specific instance heavy precipitation led to landslide but more details can support understanding of why this was the cause, and thus help supporting resilience. All in all, I think this is a very interesting work and research direction, and has certainly already some practical value. However, for mainly the above reasons it is not really publishable for time-being. I encourage the authors to continue improving the ontology and to test it against different use cases and application scenarios - then I think the paper will be in a more mature format for publishing. Minor details: - Wording in the paper needs to be improved (an example sentence: "Using a tool, which supports online collaboration with graphical ontology visualization, creation of input forms, etc. speeds up this process.")
Review 5 (by Anna Tordai)
This is a metareview for the paper that summarizes the opinions of the individual reviewers. The authors present an ontology that aggregates domain knowledge to support multidisciplinary collaboration in Cultural Heritage preservation. The reviewers agree that the paper is well written, tackles an interesting problem and the application has practical value. However, the reviewers also note that the paper lacks an evaluation and that it would benefit from a more detailed description of the ontology and the ontology design methodology. Laura Hollink & Anna Tordai