Your browser does not support JavaScript. Please to enable it.

Terms & Conditions

The idea you wish to view belongs to a community that requires acceptance of terms and conditions.

RejectAccept

    Help to Improve This Idea.

    Search

     
    Prev | Next

    SANKOFA – Semantic Annotation of Knowledge Featuring Akoma Ntoso

    by Monica Palmirani 02/08/2019 04:30 PM GMT

    • {{:upVoteCount}}
    Username * ()

        Move idea from "Expert Review" stage to:

          Collapse
          I accept the terms and conditions (see side bar). I understand all content I am submitting must be licensed under an open-source software or Creative Commons license as described in the Terms and Conditions:

          on

          Description

          The SANKOFA project will produce a web application (and an API) capable to mark up UN resolutions using the XML vocabulary Akoma Ntoso, to qualify the paragraphs according to the roles they play inside the structure (e.g., preambular or operative) and to classify the semantic aspects of all provisions (e.g. references). We also plan to qualify several peculiar textual parts (e.g. the specification of persons, organizations, dates, quantities, actions, and events). We plan to employ a hybrid solution using NLP techniques to detect structure, references, presentational parts, annexes, inline elements (e.g. symbols, committees, sessions, etc). We use parsers, vocabularies, Name-Entity-Relationship models, regular expressions, patterns, frames, and deep learning techniques to detect the knowledge that can be extracted from the text, and to represent it through semantic annotations.
          In addition to the qualifications of the textual structures and the semantics of relevant fragments, we plan to associate the correct conceptual classes from relevant ontologies (e.g., UNDO, SDGIO, UNBIS) to the corresponding fragments, and to create assertions and relationships (e.g. the decision that was taken, the timing of the decision, which actors were involved, which results are expected). The FRBR model and the ALLOT Top Level Classes are the two fundamental pillars in our approach. Assertion and relationships will be serialized in RDF according to Linked Open Data (LOD) principles. 

          Co-authors to your solution

          Monica Palmirani, Fabio Vitali, Silvio Peroni, Aldo Gangemi, Andrea Nuzzolese - UNIBO

          Link to your concept design and documentation (Required by the final day of the Submission & Collaboration phase)

          https://gitlab.com/CIRSFID/un-challange-2019

          Link to an online working solution or prototype (Required by the final day of the Submission & Collaboration phase):

          https://gitlab.com/CIRSFID/un-challange-2019

          Link to a video or screencast of your solution or prototype (Required by the final day of the Submission & Collaboration phase):

          Link to source code of your solution or prototype above. (If you submitted a link to an online solution or prototype, or to a video of your solution of prototype, you must provide a link to the source code. This item is required by the final day of the submission phase):

          https://gitlab.com/CIRSFID/un-challange-2019

          Ontologies,Patterns,NLP,frames,AkomaNtoso,deep-learning,FRBR,ALLOT,LOD,UNDO

          Move this Idea

          Select a Category

          Close this idea

          When closing an idea, you must determine whether the idea has exited successfully or unsuccessfully.

          Request to become a member

          Type a short message to the owner of this idea.:

          Invite Team Members

            Message
            *Required

            Help to Improve This Idea.

            0%
            100%
            0%
            User Tasks ?
            Required for graduation.
            Task Assigned to Due Date Status
            Judge review 05/24/2019 Completed
            on 05/28/2019
            No ideas found!
            No activities yet.