The global aim is to respond to the challenge posed by the H2020 Call ”To develop new perspectives and improved methodologies for capturing the wider societal value of culture, including but also beyond its economic impact”
Search documents and data for cultural artifacts
Analysis of cultural artifacts using Artificial Intelligence.
Transition variables from documents using AI
Social impact indicators from transition variables
Aditional tools showing AI impact on culture.
Mesoc website - all data and results from MESOC project
Search the documents on cultural artifacts
Search documents and data for cultural artifacts
Search documents based on social impact
Search documents based on cultural domain
Search by cultural domain and social impact
Semantic search on articles. Semantic search describes a search engine’s attempt to generate the most accurate search engine results possible by understanding based on searcher intent, query context, and the relationship between words.Semantic search denotes search with meaning, as distinguished from lexical search where the search engine looks for literal matches of the query words or variants of them, without understanding the overall meaning of the query. Some authors regard semantic search as a set of techniques for retrieving knowledge from richly structured data sources
Analysis of documents with Artificial Intelligence
The system reviews documents to assess the social impact of culture. The review includes analysis of all and individual documents, clustering, summarizing, keyword analysis, and other analytical tools and methods.
Text clustering is the task of grouping a set of unlabelled texts in such a way that texts in the same cluster are more similar to each other than to those in other clusters. Text clustering algorithms process text and determine if natural clusters (groups) exist in the data.
As computers work with numbers, text has to be transformed into multidimensional numbers as vectors. We use here 4096 dimensional space.
The idea is that documents can be represented numerically as vectors of features. The similarity in text can be compared by measuring the distance between these feature vectors. Objects that are near each other should belong to the same cluster. Objects that are far from each other should belong to different clusters
Cluster analysis of full text documents
Documents defining each cluster
Most significant documents for cluster
Summary of full text documents for each cluster
topic model is a type of statistical model for discovering the abstract "topics" that occur in a collection of documents. Topic modeling is a frequently used text-mining tool for discovery of hidden semantic structures in a text body.
Given that a document is about a particular topic, one would expect particular words to appear in the document more or less frequently
The topics produced by topic modeling techniques are clusters of similar words. A topic model captures this intuition in a mathematical framework, which allows examining a set of documents and discovering, based on the statistics of the words in each, what the topics might be and what each document's of topics is.
Topic analysis using latent Dirichlet allocation
Keywords defining each topic cluster
Most significant documents for topic cluster
Summary for each topic cluster
Keywords and phrases for each cultural category
Questions and answers on each cultural domain
Text summary based on cultural domain
Keywords and phrases for each cultural category and social impact
Questions and answers on each cultural category and social impact
Text summary based on each cultural category and social impact
contextual elements, which can be measured ensuring that the cultural policy or practice under inspection is generating public value and/or affecting, at least to some extent, the target individuals or groups.
Transition variables show us the paths of transformation and the channels of materialization of the impact, in a richer and more complex analysis than the cause-effect linearity.
Transition processes are complex and non linear, but induce changes across time.
Transition variables enable a better contextualisation of the concrete processes in concrete places and periods.They can be observed in shorter periods of time than the expected impacts.
We can obtain them from the experiences recorded in scientific literature and in grey material reports (evaluation reports, memos, programs...).
We are using a pre-trained BERT (Bidirectional Encoder Representations from Transformers) model for a text classification task. The main goal of the model in a text classification task is to categorize a text into one of the predefined labels or tags, and in our case to read sentences and to classify them as normal sentences or potential transition variables.
BERT is pretrained on unlabeled data extracted from BooksCorpus, which has 800M words, and from Wikipedia, which has 2,500M words. We are fine-tuning it, by training on our sample of 3500 sentences each of which is manually marked, as trans. variable or anything else.
Transition variables for all documents
Search and view of transition variables based on social impact
Search and view of transition variables based on cultural domain
Search and view of transition variables based on cultural domain and social impact
Summary of transition variables based on social impact
Summary of transition variables based on cultural domain
Summary of transition variables based on cultural domain and social impact
Cluster of transitional variable topics
Keywords and topics by cluster of transitional variables
Transitional variables sentences divided on cluster topic
Transitional variables summary on cluster topic
Clustered transitional variables
Transitional variables defining each cluster
Transformative process - Using AI to extract sentences defining social impacts
‘Social impacts’ is the term which describes the changes in the quality of life of the local residents. Changes that affect individuals’ surroundings (architecture, arts, customs, rituals etc.) constitute cultural impacts.
The enormous range of impacts include arts and crafts through to the fundamental behaviour and beliefs of individuals and collective groups (Sharpley, 2008; Sharpley & Telfer, 2014).
We are using semantic search engines to identify social impacts by finding a numerical representation of text queries using state-of-the-art language models, indexing them in a high-dimensional vector space and measuring how similar a query vector is to the indexed documents.
List of redefined impacts
Search and view of Social impact sentences
Search and view of sentences based on predefined impacts
Search and view of sentences transition variables based on impacts
Keywords from impact sentences based on social predefined impacts
Graph showing number of sentences for predefined impacts
Aditional Artificial Intelligence tools for usage in culture.