Paper Detail

LSTrans: Efficient Knowledge Transfer for Lightweight and Automated ECG Classification

Yi Zhao, Jiajun Gao, Chenyang Xu, Yuxi Zhou, Hao Wang

arxiv Score 9.2

Published 2026-07-12 · First seen 2026-07-14

General AI

Abstract

Deploying deep learning models for automated electrocardiogram classification on resource-constrained wearable devices remains challenging due to high computational costs. To address this, we propose LSTrans, a lightweight hybrid model designed for efficient and sensitive ECG analysis. LSTrans introduces a specialized 1D convolutional backbone with an interleaved layer architecture to capture both macroscopic rhythmic trends and microscopic morphological variations. This backbone is cascaded with a Transformer encoder to model long-range temporal dependencies, incorporating Low-Rank Adaptation across critical layers to compress the model and reduce the trainable parameter space. We further employ homogeneous and heterogeneous knowledge distillation to transfer diagnostic expertise from high-capacity teacher models to the student. Experimental results on multiple benchmark datasets demonstrate that LSTrans achieves a competitive balance between diagnostic sensitivity and resource efficiency, substantially reducing peak memory footprints and training latency during downstream adaptation. The source code is available for review at https://github.com/zyee00128/LSTrans4BIBM.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
soon
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{zhao2026lstrans,
  title = {LSTrans: Efficient Knowledge Transfer for Lightweight and Automated ECG Classification},
  author = {Yi Zhao and Jiajun Gao and Chenyang Xu and Yuxi Zhou and Hao Wang},
  year = {2026},
  abstract = {Deploying deep learning models for automated electrocardiogram classification on resource-constrained wearable devices remains challenging due to high computational costs. To address this, we propose LSTrans, a lightweight hybrid model designed for efficient and sensitive ECG analysis. LSTrans introduces a specialized 1D convolutional backbone with an interleaved layer architecture to capture both macroscopic rhythmic trends and microscopic morphological variations. This backbone is cascaded wit},
  url = {https://arxiv.org/abs/2607.10784},
  keywords = {cs.LG, cs.AI},
  eprint = {2607.10784},
  archiveprefix = {arXiv},
}

Metadata

{}