Robust Generalizable Heterogeneous Legal Link Prediction
This work addresses the problem of improving link prediction accuracy and generalizability for legal citation networks, which is incremental as it builds on existing approaches with specific adaptations.
The paper tackles link prediction in heterogeneous legal citation networks by incorporating edge dropout and feature concatenation to learn more robust representations, reducing error rates by up to 45%. It also proposes a method using multilingual node features with an improved asymmetric decoder to generalize predictions to geographically and linguistically disjoint data from New Zealand, enhancing inductive transferability between legal systems.
Recent work has applied link prediction to large heterogeneous legal citation networks \new{with rich meta-features}. We find that this approach can be improved by including edge dropout and feature concatenation for the learning of more robust representations, which reduces error rates by up to 45%. We also propose an approach based on multilingual node features with an improved asymmetric decoder for compatibility, which allows us to generalize and extend the prediction to more, geographically and linguistically disjoint, data from New Zealand. Our adaptations also improve inductive transferability between these disjoint legal systems.