Open bloemj opened 1 year ago
Sorry Jelke, I totally forgot about this. Here's an update.
From the readme file of the GitHub repository, you can find a list of steps we'd broken down (still a work in progress while we research applications of the network we've created / data we've processed). The tasks were:
We've been having conversations about using clustering for keyword extraction. I also thought to use a pre-trained BERT model for the task (not training my own as suggested in the readme, as I know now that this is unfeasible). But I still need to look into how exactly that could work. You mentioned an application could potentially be a knowledge graph, so that's something we're looking into. The other steps outlined have largely been completed.
Where is it? :)