Research map / Time series / Long-term forecasting
C3RL: Rethinking the Combination of Channel-independence and Channel-mixing from Representation Learning
Shusen Ma, Yunbo Zhao, Yu Kang
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 C3RL (siamese CM/CI contrastive branch) to CI models | S-Mamba, DLinear, PatchTST, iTransformer, RLinear · 9 forecasting datasets, MSE and MAE | best-score rate: 43.6% -> 81.4% | Improved |
| add C3RL to CM-strategy models | Informer, Autoformer · ETT, Weather | best-score rate: 23.8% -> 76.3% | Improved |
| add C3RL to S-Mamba | S-Mamba · all datasets | best-score rate: 34.7% -> 81.9% | Improved |
| add C3RL to DLinear | DLinear · ETTh2 horizon 720 | MSE: 0.733 -> 0.667 | Improved |
| add C3RL to Informer | Informer · ETTh2 horizon 24 | MSE: 0.761 -> 0.408 | Improved |
| add C3RL to Informer | Informer · Weather horizon 720 | MSE: 0.526 -> 0.589 | Hurt |
| add C3RL to RLinear | RLinear · ETTh1 horizon 96 | MSE: 0.366 -> 0.367 | No clear effect |
| add C3RL to PatchTST | PatchTST · Illness horizon 36 | MSE: 1.452 -> 1.317 | Improved |
| add C3RL to PatchTST | PatchTST · Illness horizon 24 | MSE: 1.401 -> 1.522 | Hurt |
| use adaptive loss weighting instead of equal weights (ablation) | C3RL on S-Mamba · ETTh1 horizon 96 | MSE: 0.402 -> 0.386 | Improved |
| use adaptive loss weighting instead of equal weights (ablation) | C3RL on DLinear · ETTh1 horizon 720 | MSE: 0.544 -> 0.472 | Improved |
| use adaptive loss weighting instead of equal weights (ablation) | C3RL on DLinear · ETTh1 horizon 96 | MSE: 0.374 -> 0.374 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Informer — Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting (AAAI 2021) · better on 2 of 3 shared results
- Autoformer — Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecasting (NeurIPS 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Self-Supervised Contrastive Learning for Long-term Forecasting (ICLR 2024 · similarity 0.87)
- TimeCHEAT: A Channel Harmony Strategy for Irregularly Sampled Multivariate Time Series Analysis (AAAI 2025 · similarity 0.86)
- Inference via Interpolation: Contrastive Representations Provably Enable Planning and Inference (NeurIPS 2024 · similarity 0.86)
- CSformer: Combining Channel Independence and Mixing for Robust Multivariate Time Series Forecasting (AAAI 2025 · similarity 0.86)
- From Similarity to Superiority: Channel Clustering for Time Series Forecasting (NeurIPS 2024 · similarity 0.85)
- TimeFilter: Patch-Specific Spatial-Temporal Graph Filtration for Time Series Forecasting (ICML 2025 · similarity 0.85)
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.
