arXiv cs.AI· Xiaoliang Wang, Zhuo Chen, Chuanyang Zheng, Jiangwei Lao, Quanbin Wang, Honglin Lin, Jiajia Liu, Zhihong Lu, Hao Jin, Jian Wang, Jinjie Gu, Lijun Wu·· 4 小时前AI 评分35
ECG-RSI：通过心电波形定向演化实现递归自我改进
ECG-RSI: Recursive Self-Improvement through Targeted Electrocardiogram Waveform Evolution
AI 导读
研究者提出 ECG-RSI，一个结合经验证波形合成与反馈引导干预搜索的心电图模型递归自我改进框架，用信号重测验证波形编辑。在七个基准上，三轮迭代将 Qwen3.5-35B-A3B 的平均问答准确率从 34.57% 提升至 56.04%，平均分类 F1 从 24.27 提升至 34.03。结果显示，继续反馈修正带来的外部收益递减，作者主张将错误诊断与干预选择分开。
正文 · 原文
Authors:Xiaoliang Wang, Zhuo Chen, Chuanyang Zheng, Jiangwei Lao, Quanbin Wang, Honglin Lin, Jiajia Liu, Zhihong Lu, Hao Jin, Jian Wang, Jinjie Gu, Lijun Wu
Abstract:Electrocardiogram (ECG) models need supervision that targets evolving weaknesses and remains faithful to waveform evidence. We introduce ECG-RSI, a recursive self-improvement framework combining verified waveform synthesis with feedback-guided intervention search. Signal remeasurement verifies waveform edits; construction errors guide generator repair, while learner errors motivate interventions compared on a separate selection set. Across seven benchmarks, three rounds raise Qwen3.5-35B-A3B's mean question-answering accuracy from 34.57% to 56.04% and mean classification F1 from 24.27 to 34.03. Across the retained trajectory, the preferred supervision shifts from direct reinforcement to confusion-focused contrasts. Continued feedback correction yields diminishing external gains. These results motivate separating error diagnosis from intervention selection and checking whether gains on the feedback set transfer to independent benchmarks.
| Comments: | Project: this https URL GitHub: this https URL |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Signal Processing (eess.SP); Quantitative Methods (q-bio.QM) |
| Cite as: | arXiv:2610.17718 [cs.AI] |
| (or arXiv:2610.17718v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.17718 arXiv-issued DOI via DataCite |
Submission history
From: Zhuo Chen [view email]
[v1]
Tue, 13 Oct 2026 04:50:27 UTC (972 KB)
[v2]
Mon, 26 Oct 2026 13:37:47 UTC (1,000 KB)
来源：arXiv cs.AI · arxiv.org