Question answering (QA) models for reading comprehension tend to learn shortcut solutions rather than the solutions intended by QA datasets. QA models that have learned shortcut solutions can achieve human-level performance in shortcut examples where shortcuts are valid, but these same behaviors degrade generalization potential on anti-shortcut examples where shortcuts are invalid. Various methods have been proposed to mitigate this problem, but they do not fully take the characteristics of shortcuts themselves into account. We assume that the learnability of shortcuts, i.e., how easy it is to learn a shortcut, is useful to mitigate the problem. Thus, we first examine the learnability of the representative shortcuts on extractive and multiple-choice QA datasets. Behavioral tests using biased training sets reveal that shortcuts that exploit answer positions and word-label correlations are preferentially learned for extractive and multiple-choice QA, respectively. We find that the more learnable a shortcut is, the flatter and deeper the loss landscape is around the shortcut solution in the parameter space. We also find that the availability of the preferred shortcuts tends to make the task easier to perform from an information-theoretic viewpoint. Lastly, we experimentally show that the learnability of shortcuts can be utilized to construct an effective QA training set; the more learnable a shortcut is, the smaller the proportion of anti-shortcut examples required to achieve comparable performance on shortcut and anti-shortcut examples. We claim that the learnability of shortcuts should be considered when designing mitigation methods.

Question answering (QA) models are shown to be insensitive to large perturbations to inputs; that is, they make correct and confident predictions even when given largely perturbed inputs from which humans can not correctly derive answers. In addition, QA models fail to generalize to other domains and adversarial test sets, while humans maintain high accuracy. Based on these observations, we assume that QA models do not use intended features necessary for human reading but rely on spurious features, causing the lack of generalization ability. Therefore, we attempt to answer the question: If the overconfident predictions of QA models for various types of perturbations are penalized, will the out-of-distribution (OOD) generalization be improved? To prevent models from making confident predictions on perturbed inputs, we first follow existing studies and maximize the entropy of the output probability for perturbed inputs. However, we find that QA models trained to be sensitive to a certain perturbation type are often insensitive to unseen types of perturbations. Thus, we simultaneously maximize the entropy for the four perturbation types (i.e., word- and sentence-level shuffling and deletion) to further close the gap between models and humans. Contrary to our expectations, although models become sensitive to the four types of perturbations, we find that the OOD generalization is not improved. Moreover, the OOD generalization is sometimes degraded after entropy maximization. Making unconfident predictions on largely perturbed inputs per se may be beneficial to gaining human trust. However, our negative results suggest that researchers should pay attention to the side effect of entropy maximization.

In various fields of data science, researchers are often interested in estimating the ratio of conditional expectation functions (CEFR). Specifically in causal inference problems, it is sometimes natural to consider ratio-based treatment effects, such as odds ratios and hazard ratios, and even difference-based treatment effects are identified as CEFR in some empirically relevant settings. This chapter develops the general framework for estimation and inference on CEFR, which allows the use of flexible machine learning for infinite-dimensional nuisance parameters. In the first stage of the framework, the orthogonal signals are constructed using debiased machine learning techniques to mitigate the negative impacts of the regularization bias in the nuisance estimates on the target estimates. The signals are then combined with a novel series estimator tailored for CEFR. We derive the pointwise and uniform asymptotic results for estimation and inference on CEFR, including the validity of the Gaussian bootstrap, and provide low-level sufficient conditions to apply the proposed framework to some specific examples. We demonstrate the finite-sample performance of the series estimator constructed under the proposed framework by numerical simulations. Finally, we apply the proposed method to estimate the causal effect of the 401(k) program on household assets.

我们提出了一种使用变异隐式神经表示（INR）的动作条件人类运动产生方法。变分形式主义可以使INR的动作条件分布，从中可以轻松地采样表示形式以产生新的人类运动序列。我们的方法通过构造提供可变的长度序列生成，因为INR的一部分已针对随时间嵌入的整个任意长度进行了优化。相反，以前的作品报告了建模可变长度序列的困难。我们证实，使用变压器解码器的方法优于人类Act12，NTU-RGBD和UESTC数据集的所有相关方法，从现实主义和生成动作的多样性方面。令人惊讶的是，即使我们使用MLP解码器的方法也始终优于最先进的基于变压器的自动编码器。特别是，我们表明，在现实主义和多样性方面，我们方法生成的可变长度运动比最先进方法产生的固定长度运动更好。 https://github.com/pacerv/implicitmotion上的代码。