SOURCE-LINKED INTELLIGENCE

Towards Automatic Evolution Tree Generation from Citation Graphs

arXiv · AI, language, vision and robotics · article · Sep 9, 2026 · UTC

Surveys remain the primary way researchers grasp the lineage of methods within an AI subfield, but they scale poorly against the current rate of publication. Existing taxonomy-induction methods are largely leaf-bound and time-agnostic; they tend to force transitional papers into mature leaves and can create topological inversions between ancestors and descendants. We propose EvoTree, a staged framework that decouples conceptual backbone learning from temporal refinement: a graph-aware encoder with distribution-based hierarchical clustering yields a stable taxonomy backbone; temporal fine-tunin

Read original source ↗ Open in workspace

recordType

paper

region

Global