원문정보
초록
영어
Little attention appears to have been paid to the relevance of learning a good representation function in solving long tail tasks. Therefore, we propose a new loss function to ensure a good representation is learnt while learning to classify. We call this loss function Triplet Class-Wise Difficulty-Based (TriCDB-CE) Loss. It is a combination of the Triplet Loss and Class-wise Difficulty-Based Cross-Entropy (CDB-CE) Loss. We prove its effectiveness empirically by performing experiments on three benchmark datasets. We find improvement in accuracy after comparing with some baseline methods. For instance, in the CIFAR-10-LT, 7 percentage points (pp) increase relative to the CDB-CE Loss was recorded. There is more room for improvement on Places-LT.
목차
1. INTRODUCTION
2. RELATED WORK
3. THEORY
3.1 CLASS-WISE DIFFICULTY-BASED WEIGHTING
3.2 CLASS-WISE DIFFICULTY-BASED CROSS-ENTROPY LOSS
3.3 TRIPLET LOSS
3.4 TRIPLET CLASS-WISE DIFFICULTY-BASED CROSS-ENTROPY LOSS
4. EXPERIMENTS
5. RESULTS AND DISCUSSION
6. CONCLUSION
Acknowledgement
References