arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2310.19660cs.CL
收藏

通过迭代生成概念瓶颈实现设计上可解释的文本理解

Interpretable-by-Design Text Understanding with Iteratively Generated Concept Bottleneck

University of Pennsylvania(宾夕法尼亚大学)

Josh Magnus Ludan, Qing Lyu, Yue Yang, Liam Dugan, Mark Yatskar, Chris Callison-Burch

AI总结:

提出内在可解释的文本分类框架 TBM,利用 LLM 自动发现并度量稀疏概念,再经线性层预测,在 12 个数据集上以较小性能代价获得可解释性。

摘要:

黑盒深度神经网络在文本分类中表现出色,但由于缺乏可解释性,其在高风险领域的应用受到阻碍。为解决这一问题,我们提出 Text Bottleneck Models(TBM,文本瓶颈模型),这是一个内在可解释的文本分类框架,可同时提供全局解释和局部解释。TBM 并不直接预测输出标签,而是为一小组稀疏的显著概念预测类别值,并在这些概念值之上使用线性层生成最终预测。这些概念可由 Large Language Model(LLM,大语言模型)自动发现和度量,无需人工整理。在 12 个多样化文本理解数据集上的实验表明,TBM 的性能可与 few-shot GPT-4 和经过微调的 DeBERTa 等黑盒基线相媲美,但仍不及经过微调的 GPT-3.5。全面的人工评估验证了 TBM 能够生成与任务相关的高质量概念,并且概念度量与人类判断高度一致,表明 TBM 作出的预测是可解释的。总体而言,我们的研究结果表明,TBM 是一个有前景的新框架,能够以极小的性能权衡增强可解释性。

英文摘要:

Black-box deep neural networks excel in text classification, yet their application in high-stakes domains is hindered by their lack of interpretability. To address this, we propose Text Bottleneck Models (TBM), an intrinsically interpretable text classification framework that offers both global and local explanations. Rather than directly predicting the output label, TBM predicts categorical values for a sparse set of salient concepts and uses a linear layer over those concept values to produce the final prediction. These concepts can be automatically discovered and measured by a Large Language Model (LLM) without the need for human curation. Experiments on 12 diverse text understanding datasets demonstrate that TBM can rival the performance of black-box baselines such as few-shot GPT-4 and finetuned DeBERTa while falling short against finetuned GPT-3.5. Comprehensive human evaluation validates that TBM can generate high-quality concepts relevant to the task, and the concept measurement aligns well with human judgments, suggesting that the predictions made by TBMs are interpretable. Overall, our findings suggest that TBM is a promising new framework that enhances interpretability with minimal performance tradeoffs.

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。

↑