Adnan KhanAdnanProduct & AI
San Francisco Bay Area
Get in touch

Source note2 min read

Failures Reveal What Metrics Miss: An Evidence-Driven Agent for Recursive Refinement of ECG Classifiers

Source

arXiv

arxiv.org/abs/2607.24419v1

Read the source

Source

Plain-Language Brief

Researchers present a new framework called RecursiveECG that uses a language model (LLM) to automatically design ECG classifiers by learning from specific failures and objective ECG evidence. The LLM is used as an offline designer, and it refines the classifiers based on concrete failures and deterministic functions derived from curated ECG criteria. This process involves analyzing failed cases, diagnosing classifier limitations, formulating targeted revisions, and only retaining evidence-supported updates. The resulting predictor outperforms strong baselines in various datasets by an average of 10.0%.

What the Document Says

  • The researchers present a framework called RecursiveECG for ECG classifier design using a language model (LLM).
  • RecursiveECG uses the LLM as an offline designer to refine ECG classifiers based on concrete failures and objective ECG evidence.
  • Criteria-to-Measurement Compilation is used to convert curated ECG criteria into deterministic functions that produce reproducible measurements for individual ECGs.
  • Evidence-Grounded Failure Review analyzes failed and comparator cases by considering raw waveforms, measurements, and model outputs.
  • The LLM diagnoses classifier limitations and formulates targeted revisions based on the analysis.
  • Candidate revisions are executed, re-evaluated under a fixed problem contract, and only evidence-supported updates are retained.
  • RecursiveECG outperforms strong baselines across PTB-XL, Georgia, and CPSC2018 datasets by an average of 10.0%.
  • Extensive ablation and transfer studies validate the effectiveness of RecursiveECG's evidence-grounded refinement process.

Key Concepts Mentioned

  • Deep models
  • ECG classification
  • LLM
  • RecursiveECG
  • Evidence-Grounded Failure Review
  • Criteria-to-Measurement Compilation

Claims Made by the Document

  • RecursiveECG outperforms strong baselines by an average of 10.0%.
  • The evidence-grounded refinement process of RecursiveECG is effective.

Evidence Used

  • Results from experiments on PTB-XL, Georgia, and CPSC2018 datasets.

Methods / Approach

  • Criteria-to-Measurement Compilation
  • Evidence-Grounded Failure Review

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Deep models have substantially advanced 12-lead ECG classification, yet their refinement still relies heavily on human experts to inspect failures and iteratively revise classifier designs. Recent LLM-based agents have demonstrated the potential for automated model design, but when guided only by aggregate performance metrics, they lack insight into why individual cases fail and how the classifier should be revised. We present RecursiveECG, an evidence-driven LLM-as-Designer framework in which an LLM serves as an offline model designer that refines ECG classifiers based on concrete failures and objective ECG evidence. To ground failure diagnosis in executable evidence, Criteria-to-Measurement Compilation converts curated ECG criteria into validated deterministic functions that produce reproducible, reference-backed measurements for individual ECGs. Building on these measurements, Evidence-Grounded Failure Review analyzes failed and comparator cases by jointly considering raw waveforms, measurements, and model outputs, enabling the LLM to diagnose classifier limitations and formulate targeted revisions. Candidate revisions are executed and re-evaluated under a fixed problem contract, and only evidence-supported updates are retained. The resulting predictor is frozen after refinement and requires no LLM inference during deployment, while an audit trail links each accepted revision to its supporting evidence. Across PTB-XL, Georgia, and CPSC2018, RecursiveECG consistently outperforms strong baselines, achieving an average relative improvement of 10.0%. Extensive ablation and transfer studies further validate the effectiveness of its evidence-grounded refinement process.