Transformers · Survey
Attention Is All You Need: Revisited for Multilingual Corpora
How self-attention scales when training and evaluating across typologically diverse languages.
Read & discuss →
Context & Meaning
Papers · Methods · Debate
Research papers
Selected readings and discussion starters. Open a card, then continue the conversation in the community.
Featured papers
Four starting points — titles you can discuss, annotate and extend.
NER · Methods
Named Entity Recognition under Low-Resource Constraints
Annotation strategies and transfer learning when labelled data is scarce or domain-shifted.
Read & discuss →
Multimodal · Accessibility
Visual Terminology for Accessible Knowledge Graphs
Pairing controlled terms with image and audio so concepts stay clear beyond plain text.
Read & discuss →
Evaluation · Practice
Evaluation Gaps in Applied Language Models
Where benchmarks diverge from production metrics — and what practitioners measure instead.
Read & discuss →
How to use this page
From paper to conversation
Treat each card as a seed — not a finished review.
01
Open a title
Pick a paper card that matches your current work or curiosity.
02
Note claims
Capture methods, limits and what would change your mind.
03
Share evidence
Post a short note or link in the matching community space.
04
Invite critique
Ask for counter-examples, datasets or related papers.
Add a paper to the list
Suggest a title, share a preprint, or open a methods thread in the community.
Go to community