TY - RPRT TI - Feature Alignment: Rethinking Efficient Active Learning via Proxy in the Context of Pre-trained Models AU - Ziting Wen AU - Oscar Pizarro AU - Stefan Williams PY - 2024 UR - https://arxiv.org/abs/2403.01101 ID - 2403.01101 ER -