Research map / Learning theory and algorithms / Prediction and decision losses
Chunk Dynamic Updating for Group Lasso with ODEs
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use ODE path following to update the group Lasso solution when adding a data chunk instead of retraining from scratch | Group Lasso · Yolanda, BNG(libras move), satellite image, BNG(wisconsin), and MR#1 | aRMSE: 10.027 vs 10.027; 4.031 vs 4.032; 1.550 vs 1.550; 29.918 vs 29.918; 18.105 vs 18.105 | No clear effect |
| Use ODE path following to update the group Lasso solution when removing a data chunk instead of retraining from scratch | Group Lasso · Yolanda, BNG(libras move), satellite image, BNG(wisconsin), and MR#1 | aRMSE: 10.230 vs 10.230; 4.117 vs 4.117; 1.667 vs 1.667; 29.976 vs 29.976; 18.866 vs 18.866 | No clear effect |
| Use ODE-based CIGL instead of batch retraining | Group Lasso chunk-incremental learning · Yolanda, BNG(libras move), and cpu act runtime experiments | running time | Helped |
| Use ODE-based CDGL instead of batch retraining | Group Lasso chunk-decremental learning · Yolanda, BNG(libras move), and cpu act runtime experiments | running time | Helped |
| Increase feature dimension while holding the data size fixed | CIGL and CDGL · MR#1–3, m > n | running time | Hurt |
| Use DA-GL rather than CIGL for incremental updating | Group Lasso · Yolanda, BNG(libras move), satellite image, BNG(wisconsin), and MR#1 | aRMSE: 10.066±0.229 vs 10.027; 4.045±0.097 vs 4.031; 1.553±0.092 vs 1.550; 29.942±0.975 vs 29.918; 20.347±2.497 vs 18.105 | Mixed |
Closest papers by meaning
- Fast Sparse Group Lasso (NeurIPS 2019 · similarity 0.88)
- Bilevel learning of the Group Lasso structure (NeurIPS 2018 · similarity 0.87)
- GAP Safe Screening Rules for Sparse-Group Lasso (NeurIPS 2016 · similarity 0.87)
- Dual Feature Reduction for the Sparse-group Lasso and its Adaptive Variant (ICML 2025 · similarity 0.86)
- Selective inference for group-sparse linear models (NeurIPS 2016 · similarity 0.85)
- SequentialAttention++ for Block Sparsification: Differentiable Pruning Meets Combinatorial Optimization (NeurIPS 2024 · similarity 0.82)
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.