Research map / Label-efficient and robust learning / Recognition and long-tail
Contrastive Symbolic Regression: Aligned Representations, Adaptive Prediction, and Diverse Ensembles
Hengzhe Zhang, Qi Chen, Bing Xue, Wolfgang Banzhaf, Mengjie Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add contrastive learning to KNN on constructed features (ablation) | SR_KNN · 58 PMLB real-world regression datasets | normalized median R2: 0.4613 -> 0.6064 | Improved |
| add contrastive learning to LOOCV-KNN (ablation) | SR_LOO-KNN · 58 PMLB real-world regression datasets | normalized median R2: 0.5905 -> 0.6410 | Improved |
| add contrastive learning to ridge+KNN (ablation) | SR_LR+KNN · 58 PMLB real-world regression datasets | normalized median R2: 0.8241 -> 0.8310 | Improved |
| use contrastive learning instead of PLS projection (ablation) | SR_LR+KNN · 58 PMLB real-world regression datasets | normalized median R2: 0.8244 -> 0.8310 | Improved |
| use contrastive learning instead of LPP projection (ablation) | SR_LR+KNN · 58 PMLB real-world regression datasets | normalized median R2: 0.7647 -> 0.8310 | Improved |
| add ridge regression boosting before KNN (ablation) | SR_KNN · 58 PMLB real-world regression datasets | normalized median R2: 0.4613 -> 0.8241 | Improved |
| add linear-rank weighted KNN evaluation and adaptive K selection (ablation) | SR_LR+CL+KNN · 58 PMLB real-world regression datasets | normalized median R2: 0.8310 -> 0.8884 | Improved |
| use DPP-based ensemble selection instead of top-T selection or RBF DPP (ablation) | CSR · 58 PMLB real-world regression datasets | average rank of test R2 | Improved |
| use an ensemble instead of a single model (ablation) | CSR · 58 PMLB real-world regression datasets | test R2 | Improved |
| use naive KNN without contrastive learning or ridge (ablation) | SR_KNN · 58 PMLB real-world regression datasets | normalized median R2: 0.4613 vs SR_LR 0.6602 | Hurt |
Closest papers by meaning
- Rethinking Symbolic Regression: Morphology and Adaptability in the Context of Evolutionary Algorithms (ICLR 2023 · similarity 0.88)
- RAG-SR: Retrieval-Augmented Generation for Neural Symbolic Regression (ICLR 2025 · similarity 0.87)
- Ab Initio Nonparametric Variable Selection for Scalable Symbolic Regression with Large $p$ (ICML 2025 · similarity 0.85)
- FunctionEvolve: Structure-Guided Symbolic Regression with LLMs (NeurIPS 2026 · similarity 0.85)
- SMILE: Bridging Continuous Optimization and Discrete Symbolic Recovery (NeurIPS 2026 · similarity 0.84)
- Symbolic regression via MDLformer-guided search: from minimizing prediction error to minimizing description length (ICLR 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
