Research map / Learning theory and algorithms / Prediction and decision losses
Efficient Optimization for Rank-Based Loss Functions
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Optimize AP loss instead of 0-1 loss | structured SVM · PASCAL VOC 2011 action classification test | mean AP: 47.934 -> 51.196 | Helped |
| Optimize NDCG loss instead of 0-1 loss | structured SVM · PASCAL VOC 2011 action classification test | mean NDCG: 84.3823 -> 85.521 | Helped |
| Use quicksort-flavored inference instead of the prior AP inference procedure | AP loss-augmented inference · PASCAL VOC 2011 action classification training | total loss-augmented inference time (seconds): 0.7154 -> 0.0625 | Helped |
| Use quicksort-flavored inference instead of the prior NDCG inference procedure | NDCG loss-augmented inference · PASCAL VOC 2011 action classification training | total loss-augmented inference time (seconds): 6.8019 -> 0.0473 | Helped |
| Optimize AP loss instead of 0-1 loss | latent AP-SVM · PASCAL VOC 2007 object detection test | mean AP: 29.4995 -> 36.616 | Helped |
| Use quicksort-flavored inference instead of the prior AP inference procedure | latent AP-SVM training · PASCAL VOC 2007 object detection | loss-augmented inference time (seconds): 7.623 -> 0.5214 | Helped |
| Optimize AP loss instead of 0-1 loss | PL-CNN · CIFAR-10 test | mean AP: 84.22 -> 85.28 | Helped |
| Optimize NDCG loss instead of 0-1 loss | PL-CNN · CIFAR-10 test | mean NDCG: 95.31 -> 96.14 | Helped |
| Use quicksort-flavored inference instead of the prior AP inference procedure | PL-CNN training with AP loss · CIFAR-10 | total training time (hours): 1.974 -> 1.589 | Helped |
| Use quicksort-flavored inference instead of the prior NDCG inference procedure | PL-CNN training with NDCG loss · CIFAR-10 | total training time (hours): 2.217 -> 1.632 | Helped |
Closest papers by meaning
- Rethinking the Optimization of Average Precision: Only Penalizing Negative Instances before Positive Ones Is Enough (AAAI 2022 · similarity 0.88)
- Optimal Classification with Multivariate Losses (ICML 2016 · similarity 0.87)
- On the Complexity of Learning Sparse Functions with Statistical and Gradient Queries (NeurIPS 2024 · similarity 0.84)
- A Gradient Accumulation Method for Dense Retriever under Memory Constraint (NeurIPS 2024 · similarity 0.84)
- POQD: Performance-Oriented Query Decomposer for Multi-vector retrieval (ICML 2025 · similarity 0.84)
- Adaptive Retrieval and Scalable Indexing for k-NN Search with Cross-Encoders (ICLR 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.