If you are having trouble with the sign in process, please <ContactLink>contact the SSH Open Marketplace team</ContactLink>.

If you are having trouble with the sign up process, please <ContactLink>contact the SSH Open Marketplace team</ContactLink>.

Subject

Social Sciences & Humanities Open Marketplace

Successfully submitted!

The first step is to download the source code that allows you to have all the features. The code is available in https://github.com/atrium-research/T3.4.1_KeywordsTranslation. You can download it in ZIP format or clone it using the git command. In this way, you have all the features available on your computer!

The second step is to choose the article you want to work with from the GoTriple platform (https://www.gotriple.eu/). We suppose here your article is already on GoTriple. If you have basic programming skills you can use the GoTriple API (https://api.gotriple.eu/), otherwise you can go with the simple interface! Search the article in the platform. By default, each article in GoTriple comes with an ID and a list of keywords. Once you have the article ID, go to the next step. 

Now, open a Python notebook (you can use Jupyter or Colab). You can look at the file test_notebook to see a tutorial of how to use the tool. The first step is to execute the function get_item_by_id with the id of your article as parameter. This formats the abstract and the keywords so that they can be ingested by the other steps of the chain. 

Now, it's time to choose which tool you want to use to link the keywords to linked data resources! At the present state, the code gives you two options: the first is using DBPedia Spotlight, a simple and light tool which is specialized for named entity linking. This tool is easy and free, but its performance is poor if keywords are not in English.
The second option is to free the power of LLM! You have two choices here: one is to use Mistral, a free model. This method does not require having an account on the OpenAI API, but it is quite slow (nearly 30 seconds to get the answer) and its performance is of course poorer than that of a GPT model. The code also allows you to use your OpenAI key to ask directly to your favourite GPT. In order to execute any of these methods, look at the guide in test_notebook.ipynb. You can choose if you want to give the LLM the content of the abstract, so that it makes more informed choices based on the context. 

If you have some programming skills, you can contribute to the tool and adapt the function in main_functions.py to use different LLMs!

If you execute any of the functions, you will get as result a set of keywords each with its own link to DBPedia (for DBPediaSpotlight) and Wikidata (for all the methods). If you go to the link you can find information about the concept (an explanation, translation in different languages, link to other resources) which you can use for your purposes!

Paolo Pedinotti

Suppose you have a research article which is written in a language other than English and you have a set of keywords for that article. How could you grasp the main topics of the article from the keywords? First, you would like to translate them, but this is not a trivial task since words can be ambiguous, and finding the exact sense requires an understanding of the whole article! Moreover, suppose you want to link your article to other resources, for example other articles that share the same topic. Now suppose you want to make your resource accessible so that people can find keywords in different languages. This workflow provides a solution to all this scenarios, since it gives you a way to map keywords to linked data (Wikidata and DBPedia entities). 

LLM-Powered Mapping of Keywords of a Research Article to Linked Data

Workflow steps(5)

1 Download the source code

2 Choose the article you want to analyze

3 Format Keywords and Abstract

4 Choose your Tool for Linking

5 Enjoy your Links!