Naive automated machine learning
Problem
Automatically finding the machine learning pipeline with the best generalization performance on a given dataset, a core AutoML task.
Why it matters
Automating pipeline design removes the need for expert manual tuning and determines the quality of downstream ML applications.
What the AI did
Naive AutoML optimizes each pipeline component in isolation and combines the best algorithm per slot, drastically reducing the search space compared to black-box optimization.
Results
Comparable and sometimes better performance than current state-of-the-art approaches (per abstract, truncated mid-sentence).
Impact
Shows simple isolated optimization can rival sophisticated AutoML optimizers, per the abstract.
Record
| Venue | Machine Learning |
|---|---|
| Published | 2022 |
| Work type | article |
| Authors | Felix Mohr, Marcel Wever |
| Citations | 24 (OpenAlex, at import time) |
| References | 52 |
| Model or method | Naive AutoML |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1007/s10994-022-06200-0 |
| Open access | hybrid |
| Keywords | AutoML, pipeline optimization, search space, machine learning, Bayesian optimization |
Importance rating
significant — thousands to millions affected
AutoML pipeline selection underpins quality of downstream ML applications across many domains; abstract frames it as an essential AutoML task but gives no scale numbers.
More in Software & Coding
| # | Paper | Year | Citations | Importance |
|---|---|---|---|---|
| 1 | ImageNet classification with deep convolutional neural networks | 2017 | 106,345 |
8
|
| 2 | Deep Residual Learning for Image Recognition | 2016 | 221,134 |
8
|
| 3 | Faster sorting algorithms discovered using deep reinforcement learning | 2023 | 175 |
8
|
| 4 | A deep learning‐based approach for software vulnerability detection using code metrics | 2022 | 42 |
7
|
| 5 | Deep Learning-Based Attack Detection and Classification in Android Devices | 2023 | 41 |
7
|