Research map / Detection and segmentation / Open-set and few-shot detection
Slot attention methods
Extend slot attention with grounding, invariance, and query optimization for object discovery.
20 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2020–2026) · an approach in Open-set and few-shot detection, keyword cluster “object centric · slot · representations”
State History Tried Baselines 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.
14 papers in 2023–24 and 4 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 2.4% to 0.7%.
Open questions the data points to
- Papers report SA on Multi-dSprites (FG-ARI) anywhere from 81 to 92.2: comparisons against it depend on whose number you take.
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
- 2020 · Object-Centric Learning with Slot Attention (NeurIPS 2020 · 211 citations)
- 2022 · Slot-VPS: Object-Centric Representation Learning for Video Panoptic Segmentation (CVPR 2022 · 21 citations)
- 2023 · Improving Object-centric Learning with Query Optimization (ICLR 2023 · 4 citations)
- 2024 · SPOT: Self-Training with Patch-Order Permutation for Object-Centric Learning with Autoregressive Transformers (CVPR 2024 · 16 citations)
- 2025 · MetaSlot: Break Through the Fixed Number of Slots in Object-Centric Learning (NeurIPS 2025 · 2 citations)
- 2026 · From Vicious to Virtuous Cycles: Synergistic Representation Learning for Unsupervised Video Object-Centric Learning (ICLR 2026 · 0 citations)
Earliest papers here
- Object-Centric Learning with Slot Attention (NeurIPS 2020 · 211 citations)
- Slot-VPS: Object-Centric Representation Learning for Video Panoptic Segmentation (CVPR 2022 · 21 citations)
- Improving Object-centric Learning with Query Optimization (ICLR 2023 · 4 citations)
What has been tried
18 papers here report 216 measured changes (a component added, removed or swapped, compared with a baseline): helped 65%, no clear effect 4%, hurt 26%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add sequence permutations to the autoregressive decoder | COCO · MBOi | 30.1±0.5 -> 33.7±0.1 | Helped | SPOT: Self-Training with Patch-Order Permutation for Object-Centric Learning with Autoregressive Transformers (CVPR 2024 · 16 citations) |
| Use zero-slot masking instead of zero-mask decoding | Reconstruction and object grouping · reconstruction quality | Hurt | Adaptive Slot Attention: Object Discovery with Dynamic Slot Number (CVPR 2024 · 10 citations) | |
| set codebook size to 1024 instead of 512 | COCO · FG-ARI | 40.3 -> 40.4 | No clear effect | MetaSlot: Break Through the Fixed Number of Slots in Object-Centric Learning (NeurIPS 2025 · 2 citations) |
| Randomly shuffle binding vectors across patches at ratio 0.5 | ADE20K semantic segmentation · mIoU | 44.14 -> 41.03 | Hurt | Does Object Binding Naturally Emerge in Large Pretrained Vision Transformers? (NeurIPS 2025 · 1 citation) |
| Use 3 update iterations instead of 7 at test time | CLEVR6 test · FG-ARI | 97.1 -> 99.8 | Helped | Object-Centric Learning with Slot Mixture Module (ICLR 2024 · 0 citations) |
| Apply cycle objectives to all slot-attention iterations instead of only the last iteration | ClevrTex · FG-ARI | 0.5453 -> 0.7245 | Helped | Cycle Consistency Driven Object Discovery (ICLR 2024 · 0 citations) |
| Evaluate using slot attention masks rather than decoder alpha masks | Pascal VOC2012 · mBO_c | 0.422 -> 0.436 | Helped | Identifiable Object-Centric Representation Learning via Probabilistic Slot Attention (NeurIPS 2024 · 0 citations) |
Baseline numbers that disagree
The same baseline on the same benchmark and metric, reported more than 5% apart by different papers here. Training budgets, data splits and reimplementations differ, so state which number you compare with.
| Baseline | Benchmark | Metric | Reported | Papers |
|---|---|---|---|---|
| SA | Multi-dSprites | FG-ARI | 81 – 92.2 | 2 papers, e.g. ICLR 2024 and ICML 2023 |
What reviewers will expect
Datasets and baselines named in the 8 papers of this approach from 2024–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- MOVi-C · 4 of 8 papers
- MOVi-E · 3 of 8 papers
- MS COCO 2017 · 3 of 8 papers
- PASCAL VOC 2012 · 3 of 8 papers
- CLEVR · 3 of 8 papers
- ClevrTex · 3 of 8 papers
- ObjectsRoom · 3 of 8 papers
- CLEVR10 · 2 of 8 papers
Baselines
- DINOSAUR · 5 of 8 papers
- SLATE · 4 of 8 papers
- Slot Attention · 4 of 8 papers
- SA · 2 of 8 papers
- DINOSAUR-MLP · 2 of 8 papers
- Rotating Features · 2 of 8 papers
- SlotDiffusion · 2 of 8 papers
- MONet · 2 of 8 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 18 papers; from OpenAlex and the papers' affiliations).
- Google DeepMind · 3 papers
- Amazon · 2 papers
- University of Montreal · 2 papers
- Institute of Science and Technology · 2 papers
- Imperial College London · 2 papers
- Mila · 2 papers
- University of Amsterdam · 2 papers
Use it
Download BibTeX of 20 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
- Object-centric learning theory and video (86 papers)
Elsewhere in the topic
- Object-level representation learning (219 papers)
- Interactive and referring segmentation (161 papers)
- Open-set and unknown object detection (160 papers)
- Amodal and open-world instance segmentation (137 papers)
- Industrial anomaly detection (117 papers)
- Unsupervised object and part discovery (114 papers)
- Domain-adaptive object detection (103 papers)
- Open-vocabulary object detection (96 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).