Research map / Learning theory and algorithms / Prediction and decision losses
Determinantal point process inference
Scalable MAP inference and learning for determinantal point processes and related online methods.
122 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Prediction and decision losses, keyword cluster “selection · robust · test”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
27 papers in 2023–24 and 29 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 5.0% to 5.0%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Fast Active Set Methods for Online Spike Inference from Calcium Imaging (NeurIPS 2016 · 43 citations)
- 2017 · Coded Distributed Computing for Inverse Problems (NeurIPS 2017 · 44 citations)
- 2018 · Fast Greedy MAP Inference for Determinantal Point Process to Improve Recommendation Diversity (NeurIPS 2018 · 77 citations)
- 2019 · Iterative Projection and Matching: Finding Structure-Preserving Representatives and Its Application to Computer Vision (CVPR 2019 · 18 citations)
- 2020 · StegaStamp: Invisible Hyperlinks in Physical Photographs (CVPR 2020 · 489 citations)
- 2021 · Bringing UMAP Closer to the Speed of Light with GPU Acceleration (AAAI 2021 · 29 citations)
- 2022 · Stability Verification in Stochastic Control Systems via Neural Network Supermartingales (AAAI 2022 · 26 citations)
- 2023 · Quantum Multi-Model Fitting (CVPR 2023 · 12 citations)
- 2024 · Optimizing ADMM and Over-Relaxed ADMM Parameters for Linear Quadratic Problems (AAAI 2024 · 10 citations)
- 2025 · B2Opt: Learning to Optimize Black-box Optimization with Little Budget (AAAI 2025 · 3 citations)
- 2026 · An Exponential Separation Between Quantum and Quantum-Inspired Classical Algorithms for Linear Systems (ICML 2026 · 1 citation)
Earliest papers here
- Fast Active Set Methods for Online Spike Inference from Calcium Imaging (NeurIPS 2016 · 43 citations)
- Parallel and Distributed Block-Coordinate Frank-Wolfe Algorithms (ICML 2016 · 30 citations)
- Split LBI: An Iterative Regularization Path with Structural Sparsity (NeurIPS 2016 · 12 citations)
What has been tried
96 papers here report 777 measured changes (a component added, removed or swapped, compared with a baseline): helped 65%, no clear effect 10%, hurt 19%, mixed 6%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Remove SAC from B2Opt | F4-F9, d=10 · objective value | F4 3.03(5.82) vs 52.7(15.0); F5 1.02(0.20) vs 5.08(0.98); F6 4e3(7e4) vs 1e5(9e4); F7 44.8(8.34) vs 47.2(5.11); F8 0.60( | Hurt | B2Opt: Learning to Optimize Black-box Optimization with Little Budget (AAAI 2025 · 3 citations) |
| Set shrinking frequency to 0.004 rather than disable region shrinking | Ackley, n=500, 30n calls · function value | 5.8 ± 0.1 -> 1.7 ± 0.4 | Helped | Scalable Acceleration for Classification-Based Derivative-Free Optimization (AAAI 2025 · 2 citations) |
| Increase embedding dimension from d=4 to d=10 | P1 regret (N=500, q=5, σ=1) · regret | 83 -> 148 | Hurt | Sample-Constrained Black Box Optimization for Audio Personalization (AAAI 2024 · 2 citations) |
| Use a larger validation set for uncertainty-set calibration | Synthetic shortest path and knapsack · coverage calibration | Helped | Conformal Inverse Optimization (NeurIPS 2024 · 2 citations) | |
| Amortize log-pdf calls in the elliptical slice sampler | Real Estate UCI · speed-up | 2× | Helped | Efficiently Vectorized MCMC on Modern Accelerators (ICML 2025 · 0 citations) |
| vary p across 2, 4, 8, 16, and 32 | Uniform vs Uniform; Uniform vs Gaussian; Taxi Pickup vs Dropoff · approximation quality and update time | Mixed | Fully Dynamic Euclidean Bi-Chromatic Matching in Sublinear Update Time (ICML 2025 · 0 citations) | |
| Increase penalty constant γ | BLO PEFT learning phase (S1) · DPO loss | Hurt | Beyond Value Functions: Single-Loop Bilevel Optimization under Flatness Conditions (NeurIPS 2025 · 0 citations) | |
| replace QUIVER with Accelerated QUIVER, placing two quantization values per dynamic-program step | Normal and LogNormal input vectors · runtime | up to 5.4× | Helped | Optimal and Approximate Adaptive Stochastic Quantization (NeurIPS 2024 · 0 citations) |
| Use PermK instead of independent RandK compressors for downlink | Non-homogeneous quadratics · s2w communication complexity | \frac{d \delta^0 \frac{1}{n} \sum (i=1)^n \norm{\mathbf{A}_i}}{\sqrt{n} \varepsilon} + \frac{d \delta^0 \max (i\in[n]) | Helped | Improving the Worst-Case Bidirectional Communication Complexity for Nonconvex Distributed Optimization under Function Similarity (NeurIPS 2024 · 0 citations) |
| use 2:1 direction splitting instead of lookahead direction selection | Illustrative function from direction-selection example · regret | 5.68 \times 10^{-10} -> 1.8 \times 10^{-4} | Hurt | Adaptive Partitioning Schemes for Optimistic Optimization (ICML 2025) |
| Add the model error term to the objective function | z500 analysis, cyclic forecasting · RMSE | above 200 m2/s2 after a week | Helped | Towards a Self-contained Data-driven Global Weather Forecasting Framework (ICML 2024) |
What reviewers will expect
Datasets and baselines named in the 17 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- MNIST · 2 of 17 papers
Baselines
- CMA-ES · 3 of 17 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 56 papers; from OpenAlex and the papers' affiliations).
- École Polytechnique · 4 papers
- National University of Singapore · 3 papers
- CNRS · 2 papers
- Harvard University · 2 papers
- University of Science and Technology of China · 2 papers
- Princeton University · 2 papers
- University College London · 2 papers
- Inria · 2 papers
- Cornell University · 2 papers
- University of Toronto · 2 papers
Use it
Download BibTeX of 40 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Active learning and subset selection (199 papers)
- Adaptive testing and experiments (201 papers)
- Distributionally robust optimization (184 papers)
- Data attribution and recourse (160 papers)
- Theory of neural network sensitivity (143 papers)
Elsewhere in the topic
- Ranking models from comparisons (127 papers)
- Conformal prediction sets (116 papers)
- Provable post-hoc explanations (92 papers)
- Tree ensemble optimization (77 papers)
- Optimal decision tree learning (75 papers)
- Consistency of surrogate losses (72 papers)
- Losses for decision-making (68 papers)
- Shapley value attribution (63 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).