Paper Detail

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

Utsav Dutta, Gerardo Pastrana, Sina Khoshfetrat Pakazad, Henrik Ohlsson

arxiv Score 4.8

Published 2026-05-29 · First seen 2026-06-01

General AI

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 Representation Model), which incorporates channel-level textual descriptions into a Transformer encoder equivariant to channel order. CHARM is trained with a Joint Embedding Predictive Architecture (JEPA) and a novel loss promoting informative, temporally stable embeddings; latent-space prediction encourages robustness to sensor noise while description-aware gating provides interpretability through learned inter-channel relationships. Across anomaly detection, classification, and short- and long-term forecasting, the learned embeddings achieve strong performance using only a linear probe. Performance is driven primarily by the JEPA objective and conditioning architecture, with text descriptions serving as channel identifiers for cross-dataset generalization.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
later
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{dutta2026giving,
  title = {Giving Sensors a Voice: Multimodal JEPA for Semantic Time-Series Embeddings},
  author = {Utsav Dutta and Gerardo Pastrana and Sina Khoshfetrat Pakazad and Henrik Ohlsson},
  year = {2026},
  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 Representation Model), which incorporates channel-level textual descriptions into a Transformer encoder equivariant to channel order. CHARM is trained with a Joint Embedding Predictive Architecture (JEPA) and a novel loss promoting informative, temporally stable embed},
  url = {https://arxiv.org/abs/2605.31580},
  keywords = {cs.LG},
  eprint = {2605.31580},
  archiveprefix = {arXiv},
}

Metadata

{}