Research map / Learning theory and algorithms / Prediction and decision losses
Faster Greedy MAP Inference for Determinantal Point Processes
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 |
|---|---|---|---|
| Use first-order log-determinant approximations with partitioning instead of recomputing each marginal gain | Greedy DPP MAP inference · Synthetic DPP, d=10,000 | speedup over Lazy: about $3$ times faster | Helped |
| Use batch-greedy selection | Greedy DPP MAP inference · Synthetic DPP, d=10,000 | speedup over Lazy: $9$ times faster | Helped |
| Use batch-greedy selection | Greedy DPP MAP inference · Synthetic DPP, d=10,000 | log-probability ratio vs Lazy: $2 \%$ loss on accuracy | Hurt |
| Increase the number of partitions (ablation) | Faster Greedy DPP Inference · Synthetic DPP, d=1,000 | log-probability ratio vs Lazy | Helped |
| Increase the batch size (ablation) | Faster Batch-Greedy DPP Inference · Synthetic DPP, d=1,000 | log-probability ratio vs Lazy | Hurt |
| Use the batch algorithm at d=40,000 | Greedy DPP MAP inference · Synthetic DPP, d=40,000 | speedup over Lazy: $19$ times faster | Helped |
| Use the proposed greedy algorithm | Greedy DPP MAP inference · Video summarization, 39 YouTube videos | F-score | No clear effect |
| Share the same random vectors across LDAS runs (ablation) | Faster Batch-Greedy DPP Inference · Comparing log-determinants across partitions | log-determinant comparison variance | Helped |
Closest papers by meaning
- Lazy and Fast Greedy MAP Inference for Determinantal Point Process (NeurIPS 2022 · similarity 0.94)
- Online MAP Inference of Determinantal Point Processes (NeurIPS 2020 · similarity 0.91)
- Scalable Learning and MAP Inference for Nonsymmetric Determinantal Point Processes (ICLR 2021 · similarity 0.91)
- Learning Determinantal Point Processes with Moments and Cycles (ICML 2017 · similarity 0.91)
- Fast Greedy MAP Inference for Determinantal Point Process to Improve Recommendation Diversity (NeurIPS 2018 · similarity 0.90)
- State-of-art minibatches via novel DPP kernels: discretization, wavelets, and rough objectives (NeurIPS 2026 · similarity 0.88)
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.