PhD student, The University of Hong Kong
1 paper at NeurIPS 2025
We introduce SEAL, a semantic-aware hierarchical learning framework that leverages real-world coarse-to-fine taxonomies to provide stronger supervision for GCD.
