Stakeholders of machine learning models desire explainable artificial intelligence (XAI) to produce human-understandable and consistent interpretations. In computational toxicity, augmentation of text-based molecular representations has been used successfully for transfer learning on downstream tasks. Augmentations of molecular representations can also be used at inference to compare differences between multiple representations of the same ground-truth. In this study, we investigate the robustness of eight XAI methods using test-time augmentation for a molecular-representation model in the field of computational toxicity prediction. We report significant differences between explanations for different representations of the same ground-truth, and show that randomized models have similar variance. We hypothesize that text-based molecular representations in this and past research reflect tokenization more than learned parameters. Furthermore, we see a greater variance between in-domain predictions than out-of-domain predictions, indicating XAI measures something other than learned parameters. Finally, we investigate the relative importance given to expert-derived structural alerts and find similar importance given irregardless of applicability domain, randomization and varying training procedures. We therefore caution future research to validate their methods using a similar comparison to human intuition without further investigation. In this research we critically investigate XAI through test-time augmentation, contrasting previous assumptions about using expert validation and showing inconsistencies within models for identical representations. SMILES augmentation has been used to increase model accuracy, but was here adapted from the field of image test-time augmentation to be used as an independent indication of the consistency within SMILES-based molecular representation models.

中文翻译：

---

使用测试时间增强来研究可解释的人工智能：方法、模型和人类直觉之间的不一致

机器学习模型的利益相关者希望可解释的人工智能 (XAI) 能够产生人类可理解且一致的解释。在计算毒性中，基于文本的分子表示的增强已成功用于下游任务的迁移学习。分子表示的增强也可以用于推理，以比较同一基本事实的多个表示之间的差异。在本研究中，我们研究了计算毒性预测领域中分子表示模型的测试时间增强的八种 XAI 方法的稳健性。我们报告了对同一事实的不同表示的解释之间的显着差异，并表明随机模型具有相似的方差。我们假设本研究和过去的研究中基于文本的分子表示更多地反映了标记化而不是学习的参数。此外，我们发现域内预测之间的差异比域外预测之间的差异更大，这表明 XAI 测量的是学习参数以外的其他东西。最后，我们调查了专家衍生的结构警报的相对重要性，并发现无论适用范围、随机性和不同的训练程序如何，都具有相似的重要性。因此，我们提醒未来的研究使用与人类直觉的类似比较来验证他们的方法，而无需进一步调查。在这项研究中，我们通过测试时间增强来批判性地研究 XAI，对比之前关于使用专家验证的假设，并显示相同表示的模型内的不一致之处。 SMILES 增强已用于提高模型准确性，但此处改编自图像测试时间增强领域，用作基于 SMILES 的分子表示模型内一致性的独立指示。