HEPA architecture predicts critical time-series events using self-supervision

Researchers have developed HEPA, a novel self-supervised architecture for predicting critical events in multivariate time series data. This architecture uses a causal Transformer encoder pretrained with a Joint-Embedding Predictive Architecture (JEPA) to forecast future representations, enabling it to learn from unlabeled data. HEPA has demonstrated superior performance across 14 benchmarks, including water contamination and cyberattack detection, outperforming existing models like PatchTST and Chronos-2 with significantly less labeled data and fewer tuned parameters. AI

IMPACT Enables more accurate prediction of rare critical events in time series data with less labeled data.

RANK_REASON The cluster contains a research paper detailing a new architecture for time series analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• iTransformer
• Jonas Petersen
• MAE
• PatchTST

• paper
• model release

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