.png){kind=icon}
.png){kind=icon}
Return
Deep active learning: A reality check
.png){kind=icon}
OA
AI
DOI：10.1016/j.patrec.2025.07.017.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We do a comprehensive evaluation of different active learning acquisition functions. • Surprisingly, we find out that entropy outperforms all other methods. • We give several recommendations to active learning practitioners. • We extend our classification finding to semi-supervised learning, and object detection.
Keywords:
active learning
entropy
acquisition functions
semi-supervised learning
object detection