A hybrid Relational-GAT with a projection head that learns relation operators to transform embeddings, impute unseen node vectors, and expand search queries via relational paths.

0 stars 0 forks 0 watchers Python Apache License 2.0
link-prediction llm projection relation-learning
2 Open Issues Need Help Last updated: Sep 10, 2025

Open Issues Need Help

View All on GitHub

AI Summary: The task involves re-training the existing 'medium' model architecture. The primary change for this re-training is to utilize a larger learning rate than what was previously used.

Complexity: 2/5
good first issue

A hybrid Relational-GAT with a projection head that learns relation operators to transform embeddings, impute unseen node vectors, and expand search queries via relational paths.

Python
#link-prediction#llm#projection#relation-learning

AI Summary: This issue focuses on preparing and training a RelGAT model. The model needs to be trained using a learned embedder from `plwordnet_ml.embedder` and should only proceed after the new set of embeddings has been prepared, as detailed in the linked issue #5.

Complexity: 2/5
good first issue

A hybrid Relational-GAT with a projection head that learns relation operators to transform embeddings, impute unseen node vectors, and expand search queries via relational paths.

Python
#link-prediction#llm#projection#relation-learning