CARNet framework enhances multivariate time series forecasting with cycle-aware core aggregation

Researchers have introduced CARNet, a novel framework designed for multivariate time series forecasting that addresses the challenge of modeling cross-variate dependencies, especially with strong periodic patterns. CARNet integrates global recurrent cycle information into efficient core-based interaction modeling using Multihead Core Aggregation. Experiments show that CARNet surpasses existing transformer and non-attention baselines in forecasting accuracy across various prediction horizons while maintaining linear complexity. AI

IMPACT Introduces a new method for time series forecasting that improves accuracy and efficiency by incorporating cyclic patterns.

RANK_REASON The item is a research paper detailing a new model (CARNet) for time series forecasting, submitted to arXiv. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CARNet
• CatalyzeX
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• Md. Shahria Sarker Shuvo
• Multihead Core Aggregation
• ScienceCast
• Transformer++

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →