PulseAugur
EN
LIVE 21:15:18

New model CHARM learns time-series embeddings using JEPA

Researchers have developed CHARM, a Channel-Aware Representation Model, designed for learning general-purpose representations from heterogeneous multivariate time series data. This model utilizes a Transformer encoder that is equivariant to channel order and is trained with a Joint Embedding Predictive Architecture (JEPA) and a novel loss function. The JEPA objective enhances robustness to sensor noise, while description-aware gating offers interpretability by learning inter-channel relationships. AI

IMPACT Introduces a novel approach for time-series representation learning, potentially improving performance in anomaly detection, classification, and forecasting tasks.

RANK_REASON The cluster contains an academic paper detailing a new model and methodology.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

COVERAGE [2]

  1. arXiv cs.LG TIER_1 English(EN) · Utsav Dutta, Gerardo Pastrana, Sina Khoshfetrat Pakazad, Henrik Ohlsson ·

    Giving Sensors a Voice: Multimodal JEPA for Semantic Time-Series Embeddings

    arXiv:2605.31580v1 Announce Type: new Abstract: Transformer-based architectures have advanced sequence modeling in language and vision, yet general-purpose representation learning for heterogeneous multivariate time series remains underexplored. We introduce CHARM (Channel-Aware …

  2. arXiv cs.LG TIER_1 English(EN) · Henrik Ohlsson ·

    Giving Sensors a Voice: Multimodal JEPA for Semantic Time-Series Embeddings

    Transformer-based architectures have advanced sequence modeling in language and vision, yet general-purpose representation learning for heterogeneous multivariate time series remains underexplored. We introduce CHARM (Channel-Aware Representation Model), which incorporates channe…