Research map / Architectures and efficiency / Neural architecture search
AutoML pipeline search
Searches model pipelines and operations beyond standard networks, including multimodal fusion and learning-curve ranking.
57 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Neural architecture search, keyword cluster “differentiable · image · latency”
State History Tried 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.
15 papers in 2023–24 and 10 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 20.0% to 20.0%.
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
- 2017 · Memory-Augmented Attribute Manipulation Networks for Interactive Fashion Search (CVPR 2017 · 149 citations)
- 2018 · Transfer Learning with Neural AutoML (NeurIPS 2018 · 41 citations)
- 2019 · AutoAugment: Learning Augmentation Strategies From Data (CVPR 2019 · 2,815 citations)
- 2020 · AutoML-Zero: Evolving Machine Learning Algorithms From Scratch (ICML 2020 · 124 citations)
- 2021 · Removing Raindrops and Rain Streaks in One Go (CVPR 2021 · 202 citations)
- 2022 · Learning Strides in Convolutional Neural Networks (ICLR 2022 · 42 citations)
- 2023 · Minimizing the Accumulated Trajectory Error To Improve Dataset Distillation (CVPR 2023 · 72 citations)
- 2024 · SasWOT: Real-Time Semantic Segmentation Architecture Search WithOut Training (AAAI 2024 · 23 citations)
- 2025 · LoMix: Learnable Weighted Multi-Scale Logits Mixing for Medical Image Segmentation (NeurIPS 2025 · 1 citation)
- 2026 · DeNAS-ViT: Data Efficient NAS-Optimized Vision Transformer for Ultrasound Image Segmentation (AAAI 2026 · 2 citations)
Earliest papers here
- Memory-Augmented Attribute Manipulation Networks for Interactive Fashion Search (CVPR 2017 · 149 citations)
- Transfer Learning with Neural AutoML (NeurIPS 2018 · 41 citations)
- AutoAugment: Learning Augmentation Strategies From Data (CVPR 2019 · 2,815 citations)
What has been tried
50 papers here report 554 measured changes (a component added, removed or swapped, compared with a baseline): helped 72%, no clear effect 9%, hurt 15%, 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 |
|---|---|---|---|---|
| Combine SasWOT- with the NWOT proxy using AVT normalization | Cityscapes test · mIoU (%) | 66.7 -> 69.8 | Helped | SasWOT: Real-Time Semantic Segmentation Architecture Search WithOut Training (AAAI 2024 · 23 citations) |
| Search on PTB training data rather than WikiText2 | WikiText2, 80% inheriting ratio · perplexity | 6.89 -> 7.08 | Hurt | Search for Efficient Large Language Models (NeurIPS 2024 · 5 citations) |
| Use only μ instead of μ+σ in the acquisition function | MIMIC-IV Tasks @ 5, Tasks @ 10, and Tasks @ 25 · ROC and AVP gain (%) | ROC: +1.69 -> +1.49, +2.55 -> +2.08, +2.80 -> +2.68; AVP: +7.74 -> +7.12, +8.81 -> +7.53, +13.43 -> +12.70 | Mixed | Automated Multi-Task Learning for Joint Disease Prediction on Electronic Health Records (NeurIPS 2024 · 3 citations) |
| add the fourth fusion operator to the best three-operator setting | Synapse 8-organ · DICE (%) | ~0.46% | Helped | LoMix: Learnable Weighted Multi-Scale Logits Mixing for Medical Image Segmentation (NeurIPS 2025 · 1 citation) |
| add parameter-space optimization | Translation · chrF | 32.21 -> 44.68 | Helped | HM3: Hierarchical Multi-Objective Model Merging for Pretrained Models (NeurIPS 2025 · 1 citation) |
| Replace the hybrid block pattern with MBConv-only blocks | ImageNet, 256KB SRAM/1MB Flash · Top-1 accuracy | 66.7% -> 60.8% | Hurt | AtomNet: Designing Tiny Models from Operators Under Extreme MCU Constraints (AAAI 2025 · 0 citations) |
| add pixel-level SNL pruning | CIFAR100, ReLU budget 25K · accuracy | 70.59% -> 71.42% | Helped | Seesaw: Compensating for Nonlinear Reduction with Linear Computations for Private Inference (ICML 2024) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 25 papers; from OpenAlex and the papers' affiliations).
- Carnegie Mellon University · 2 papers
- Hong Kong University of Science and Technology · 2 papers
- The Hong Kong Polytechnic University · 2 papers
- University of Freiburg · 2 papers
- Albert-Ludwigs-Universität Freiburg · 2 papers
Use it
Download BibTeX of 34 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
- Latency-aware architecture search (91 papers)
- Differentiable architecture search (DARTS) (32 papers)
Elsewhere in the topic
- Efficient NAS search strategies (86 papers)
- Bi-level differentiable NAS (84 papers)
- One-shot and few-shot supernets (30 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).