提出无互锁的生成预测框架，提升文本理性化效果。
Interlocking-free Selective Rationalization Through Genetic-based Learning
- 用遗传全局搜索分离训练生成器与预测器
- 在合成与真实数据集上超越多个先进模型
- 无需额外学习开销，适合追求高效可靠推理场景
选择性理性化的一种流行端到端架构是先生成后预测的流水线，由生成器提取关键片段并输入预测器。这种协作系统因模块间主导关系导致次优平衡点，即互锁现象。现有方法仅缓解该问题，常依赖基于特征的启发式、采样或临时正则化。本文提出首个无互锁的理性化架构GenSPP，无需任何学习开销。其通过遗传全局搜索实现生成器与预测器的解耦训练，避免互锁。在合成和真实世界基准上的实验表明，该模型优于多个当前最优方法。
原文摘要 · Abstract (English)
A popular end-to-end architecture for selective rationalization is the select-then-predict pipeline, comprising a generator to extract highlights fed to a predictor. Such a cooperative system suffers from suboptimal equilibrium minima due to the dominance of one of the two modules, a phenomenon known as interlocking. While several contributions aimed at addressing interlocking, they only mitigate its effect, often by introducing feature-based heuristics, sampling, and ad-hoc regularizations. We present GenSPP, the first interlocking-free architecture for selective rationalization that does not require any learning overhead, as the above-mentioned. GenSPP avoids interlocking by performing disjoint training of the generator and predictor via genetic global search. Experiments on a synthetic and a real-world benchmark show that our model outperforms several state-of-the-art competitors.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。