ASDA：自动化技能蒸馏与适应用于金融推理
ASDA: Automated Skill Distillation and Adaptation for Financial Reasoning
AI总结：
ASDA通过迭代误差纠正学习生成结构化技能 artifact，无需修改模型权重，在金融推理任务中实现显著提升，优于无训练基线。
摘要：
适应大型语言模型（LLMs）到专门的金融推理通常需要昂贵的微调，产生模型锁定的专长。训练自由的替代方案已出现，但我们的实验表明，领先方法（GEPA和ACE）在FAMMA金融推理基准上仅获得微小提升，暴露了无结构文本优化在复杂多步骤领域推理中的局限性。我们引入了自动化技能蒸馏与适应（ASDA），一个框架，通过迭代误差纠正学习自动生成结构化技能 artifact，而无需修改模型权重。教师模型分析学生模型在金融推理任务中的失败，通过子领域和错误类型聚类错误，并合成包含推理过程、代码模板和示例的技能文件，这些文件在推理期间动态注入。在FAMMA上评估，ASDA在算术推理上实现高达+17.33%的提升，在非算术推理上实现+5.95%的提升，明显优于所有训练自由基线。生成的技能 artifact 是可读的、版本受控的，并且与Agent Skills开放标准兼容，为任何拥有标记领域数据集的组织提供了一条实用且可审计的领域适应路径，无需权重访问或重新训练。
英文摘要：
Adapting large language models (LLMs) to specialized financial reasoning typically requires expensive fine-tuning that produces model-locked expertise. Training-free alternatives have emerged, yet our experiments show that leading methods (GEPA and ACE) achieve only marginal gains on the FAMMA financial reasoning benchmark, exposing the limits of unstructured text optimization for complex, multi-step domain reasoning. We introduce Automated Skill Distillation and Adaptation (ASDA), a framework that automatically generates structured skill artifacts through iterative error-corrective learning without modifying model weights. A teacher model analyzes a student model's failures on financial reasoning tasks, clusters errors by subfield and error type, and synthesizes skill files containing reasoning procedures, code templates, and worked examples, which are dynamically injected during inference. Evaluated on FAMMA, ASDA achieves up to +17.33% improvement on arithmetic reasoning and +5.95% on non-arithmetic reasoning, substantially outperforming all training-free baselines. The resulting skill artifacts are human-readable, version-controlled, and compatible with the Agent Skills open standard, offering any organization with a labeled domain dataset a practical and auditable path to domain adaptation without weight access or retraining.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
