Symbolic regression as a feature engineering method for machine and deep learning regression tasks
Problem
Manually designing input features for machine and deep learning regression models relies on domain expertise and limits predictive performance.
Why it matters
Better regression accuracy affects prediction quality across science and industry, including materials like superconductors.
What the AI did
Symbolic regression automatically derives features that are fed into ML and deep learning regression models, improving their predictions.
Results
34%–86% RMSE improvement on synthetic datasets; 4%–11.5% improvement on 21 real-world datasets
Impact
Demonstrated accuracy gains across many datasets, including a superconductor prediction use case.
Record
| Venue | Machine Learning Science and Technology |
|---|---|
| Published | 2024 |
| Work type | article |
| Authors | Assaf Shmuel, Oren Glickman, Teddy Lazebnik |
| Citations | 19 (OpenAlex, at import time) |
| References | 122 |
| Model or method | Symbolic regression combined with ML/DL regression models |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1088/2632-2153/ad513a |
| Open access | gold |
| Keywords | symbolic regression, feature engineering, machine learning, deep learning, RMSE, superconductors |
Importance rating
real but narrower advance
General feature-engineering method improving regression accuracy across 21 real datasets; useful but incremental methodological advance, no stated application scale.
More in Mathematics
| # | Paper | Year | Citations | Importance |
|---|---|---|---|---|
| 1 | Discovering faster matrix multiplication algorithms with reinforcement learning | 2022 | 476 |
7
|
| 2 | Solving olympiad geometry without human demonstrations | 2024 | 307 |
7
|
| 3 | Mathematical discoveries from program search with large language models | 2023 | 509 |
6
|
| 4 | Superhuman AI for multiplayer poker | 2019 | 583 |
6
|
| 5 | Efficient generator of mathematical expressions for symbolic regression | 2023 | 20 |
4
|