TY - JOUR AU - Zhang, Quanshi AU - Wang, Wenguan AU - Zhu, Song-Chun PY - 2018/04/29 Y2 - 2024/03/29 TI - Examining CNN Representations With Respect to Dataset Bias JF - Proceedings of the AAAI Conference on Artificial Intelligence JA - AAAI VL - 32 IS - 1 SE - AAAI Technical Track: Machine Learning DO - 10.1609/aaai.v32i1.11833 UR - https://ojs.aaai.org/index.php/AAAI/article/view/11833 SP - AB - <p> Given a pre-trained CNN without any testing samples, this paper proposes a simple yet effective method to diagnose feature representations of the CNN. We aim to discover representation flaws caused by potential dataset bias. More specifically, when the CNN is trained to estimate image attributes, we mine latent relationships between representations of different attributes inside the CNN. Then, we compare the mined attribute relationships with ground-truth attribute relationships to discover the CNN's blind spots and failure modes due to dataset bias. In fact, representation flaws caused by dataset bias cannot be examined by conventional evaluation strategies based on testing images, because testing images may also have a similar bias. Experiments have demonstrated the effectiveness of our method. </p> ER -