| Paper title | Paper url | Summary |
|---|
| Neural Architecture Search without Training (ICML 2021) | [arXiv](https://arxiv.org/abs/2006.04647) | Proposes zero-cost NAS by predicting network performance from untrained weights, enabling architecture search in seconds without training. |
| AutoFormer: Searching Transformers for Visual Recognition (ICCV 2021) | [arXiv](https://arxiv.org/abs/2107.00651) | One-shot NAS framework for Vision Transformers, producing architectures that outperform DeiT/ViT baselines on ImageNet. |
| FBNetV3: Joint Architecture-Recipe Search using Predictor Pretraining (CVPR 2021) | [arXiv](https://arxiv.org/abs/2006.02049) | Jointly searches architectures and training recipes, matching EfficientNet with fewer FLOPs and improving detection tasks. |
| NAS-Bench-360: Benchmarking NAS on Diverse Tasks (NeurIPS D&B 2022) | [arXiv](https://arxiv.org/abs/2110.05668) | Benchmarking suite covering 10 diverse domains, revealing limits of NAS generalization beyond vision tasks. |
| DEHB: Evolutionary Hyperband for Scalable, Robust and Efficient HPO (IJCAI 2021) | [arXiv](https://arxiv.org/abs/2011.09854) | Combines Differential Evolution with Hyperband for efficient HPO, achieving up to 1000× speedups over random search. |
| HyperBO: Pre-trained Gaussian Processes for Bayesian Optimization (JMLR 2024) | [JMLR](https://jmlr.org/papers/v25/23-0212.html) | Uses meta-learning to pre-train GP priors for Bayesian optimization, finding hyperparameters ~3× more efficiently. |
| Auto-sklearn 2.0: Hands-free AutoML via Meta-Learning (JMLR 2022) | [JMLR](https://jmlr.org/papers/v23/20-1121.html) | Next-gen Auto-sklearn with meta-learning warm starts and resource-aware scheduling, achieving faster and more accurate AutoML. |
| FLAML: A Fast and Lightweight AutoML Library (MLSys 2021) | [arXiv](https://arxiv.org/abs/2009.09288) | Lightweight AutoML library focusing on cost-efficient tabular learning, outperforming frameworks under tight time budgets. |
| CAAFE: Context-Aware Automated Feature Engineering with LLMs (NeurIPS 2023) | [arXiv](https://arxiv.org/abs/2306.16487) | Leverages large language models to automatically generate interpretable new features, improving tabular model performance. |
| LoRA: Low-Rank Adaptation of Large Language Models (ICLR 2022) | [arXiv](https://arxiv.org/abs/2106.09685) | Efficiently fine-tunes large models by injecting low-rank trainable matrices, reducing tuned parameters up to 10,000×. |
| Quick-Tune: Quickly Learning Which Pretrained Model to Finetune and How (ICLR 2024) | [OpenReview](https://openreview.net/forum?id=hf5SjDKQVY) | Automates selection of pretrained models and fine-tuning strategies using meta-learned performance predictors. |
| TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Second (ICLR 2023) | [OpenReview](https://openreview.net/forum?id=Wp8sQyx6ydl) | Foundation model for tabular data, solving small tabular tasks in one forward pass without training. |
| TabPFN v2: Accurate Predictions on Small Data with a Tabular Foundation Model (Nature 2025) | [Nature](https://www.nature.com/articles/s41586-025-08505-z) | Extends TabPFN to larger datasets, regression, missing values, and time-series, showing broader foundation model utility. |
| FedNAS: Federated Deep Learning via Neural Architecture Search (ICLR 2022) | [OpenReview](https://openreview.net/forum?id=F9D-1LGeXx) | Automates NAS in federated settings, letting clients collaboratively find architectures under data heterogeneity. |
| HyperFD: Privacy-Preserving Online AutoML for Face Detection (CVPR 2022) | [CVPR](https://openaccess.thecvf.com/content/CVPR2022/html/Yan_HyperFD_Privacy-Preserving_Online_AutoML_for_Domain-Specific_Face_Detection_CVPR_2022_paper.html) | Privacy-preserving AutoML for face detection, tuning models using meta-features without sharing raw images. |
| AutoGluon-TimeSeries: AutoML for Probabilistic Time Series Forecasting (AutoML Conf 2023) | [PMLR](https://proceedings.mlr.press/v220/shchur23a.html) | Open-source AutoML library for time-series, ensembling diverse models to generate accurate point and probabilistic forecasts. |
| AutoClust: A Framework for Automated Clustering based on Cluster Validity Indices (ICDM 2020) | [PDF](https://ieeexplore.ieee.org/document/9338424) | Automates clustering algorithm and hyperparameter selection using validity indices, making clustering AutoML practical. |
| AMLB: An AutoML Benchmark (JMLR 2024) | [JMLR](https://jmlr.org/papers/v25/22-0493.html) | Standard benchmark framework comparing 9 AutoML systems on 100+ tasks, enabling fair and reproducible evaluation. |
| SmartCal: A Novel Automated Approach to Classifier Probability Calibration (AutoML 2025) | [OpenReview](https://openreview.net/forum?id=SmartCal2025) | An AutoML framework for the calibration of supervised classification machine learning models. |
