尽管可解释的AI的大量研究重点是产生有效的解释，但较少的工作致力于人们如何理解和解释解释的问题。在这项工作中，我们通过研究基于显着性数据的解释来关注这个问题。文本模型的特征属性解释旨在传达输入文本的哪些部分比其他部分更具影响力。许多当前的解释方法，例如基于梯度或基于沙普利价值的方法，都提供了重要的衡量标准，这些方法在数学上是众所周知的。但是，一个人接受解释（解释）如何理解它？他们的理解是否与解释试图交流的内容相匹配？我们从经验上研究了输入的各种因素，特征 - 贡献解释和可视化程序对Laypeople对解释的解释的影响。我们询问人群工人对英语和德语的任务进行解释，并根据感兴趣的因素适合他们的回答。我们发现人们经常误解解释：尽管有直接传达重要性的解释，但肤浅和无关的因素（例如单词长度）影响了解释者的重要性分配。然后，我们证明其中一些失真可以减弱：我们提出了一种基于过度感受和低估的模型估计的方法来调整销售的方法，并探索条形图作为热图显着性可视化的替代方法。我们发现两种方法都可以减轻特定因素的扭曲作用，从而使对解释的理解更好地理解。

创意支持工具中的反馈可以帮助人群推动他们的意思。但是，目前的反馈方法需要从促进者或同行中进行人力评估。这不可扩展到大人群。我们提出可解释的定向多样性来自动预测观点的质量和多样性分数，并提供AI解释 - 归因，对比归因和反事实建议 - 反馈意见（低），以及如何获得更高的分数。由于用户迭代地提高其想象，这些解释提供了多面反馈。我们进行了形成性和控制的用户研究，以了解解释的使用和有用性，以提高观念多样性和质量。用户感谢解释反馈帮助重点努力，并提供了改进的方向。这导致解释与没有反馈或反馈仅具有预测的反馈和反馈相比提高了多样性。因此，我们的方法为解释和丰富的反馈开辟了可解释的AI的机会，以获得迭代人群思想和创造力支​​持工具。

情绪分析中最突出的任务是为文本分配情绪，并了解情绪如何在语言中表现出来。自然语言处理的一个重要观察结果是，即使没有明确提及情感名称，也可以通过单独参考事件来隐式传达情绪。在心理学中，被称为评估理论的情感理论类别旨在解释事件与情感之间的联系。评估可以被形式化为变量，通过他们认为相关的事件的人们的认知评估来衡量认知评估。其中包括评估事件是否是新颖的，如果该人认为自己负责，是否与自己的目标以及许多其他人保持一致。这样的评估解释了哪些情绪是基于事件开发的，例如，新颖的情况会引起惊喜或不确定后果的人可能引起恐惧。我们在文本中分析了评估理论对情绪分析的适用性，目的是理解注释者是否可以可靠地重建评估概念，如果可以通过文本分类器预测，以及评估概念是否有助于识别情感类别。为了实现这一目标，我们通过要求人们发短信描述触发特定情绪并披露其评估的事件来编译语料库。然后，我们要求读者重建文本中的情感和评估。这种设置使我们能够衡量是否可以纯粹从文本中恢复情绪和评估，并为判断模型的绩效指标提供人体基准。我们将文本分类方法与人类注释者的比较表明，两者都可以可靠地检测出具有相似性能的情绪和评估。我们进一步表明，评估概念改善了文本中情绪的分类。

Prior work has identified a resilient phenomenon that threatens the performance of human-AI decision-making teams: overreliance, when people agree with an AI, even when it is incorrect. Surprisingly, overreliance does not reduce when the AI produces explanations for its predictions, compared to only providing predictions. Some have argued that overreliance results from cognitive biases or uncalibrated trust, attributing overreliance to an inevitability of human cognition. By contrast, our paper argues that people strategically choose whether or not to engage with an AI explanation, demonstrating empirically that there are scenarios where AI explanations reduce overreliance. To achieve this, we formalize this strategic choice in a cost-benefit framework, where the costs and benefits of engaging with the task are weighed against the costs and benefits of relying on the AI. We manipulate the costs and benefits in a maze task, where participants collaborate with a simulated AI to find the exit of a maze. Through 5 studies (N = 731), we find that costs such as task difficulty (Study 1), explanation difficulty (Study 2, 3), and benefits such as monetary compensation (Study 4) affect overreliance. Finally, Study 5 adapts the Cognitive Effort Discounting paradigm to quantify the utility of different explanations, providing further support for our framework. Our results suggest that some of the null effects found in literature could be due in part to the explanation not sufficiently reducing the costs of verifying the AI's prediction.

自我跟踪可以提高人们对他们不健康的行为的认识，为行为改变提供见解。事先工作探索了自动跟踪器如何反映其记录数据，但它仍然不清楚他们从跟踪反馈中学到多少，以及哪些信息更有用。实际上，反馈仍然可以压倒，并简明扼要可以通过增加焦点和减少解释负担来改善学习。为了简化反馈，我们提出了一个自动跟踪反馈显着框架，以定义提供反馈的特定信息，为什么这些细节以及如何呈现它们（手动引出或自动反馈）。我们从移动食品跟踪的实地研究中收集了调查和膳食图像数据，并实施了Salientrack，一种机器学习模型，以预测用户从跟踪事件中学习。使用可解释的AI（XAI）技术，SalientRack识别该事件的哪些特征是最突出的，为什么它们导致正面学习结果，并优先考虑如何根据归属分数呈现反馈。我们展示了用例，并进行了形成性研究，以展示Salientrack的可用性和有用性。我们讨论自动跟踪中可读性的影响，以及如何添加模型解释性扩大了提高反馈体验的机会。

关于人类阅读的研究长期以来一直记录在阅读行为表明特定于任务的效果，但是建立一个通用模型来预测人类在给定任务中将显示什么的通用模型。我们介绍了Neat，这是人类阅读中注意力分配的计算模型，基于人类阅读优化了一项任务中关注经济和成功之间的权衡。我们的模型是使用当代神经网络建模技术实施的，并对注意力分配的分配方式在不同任务中如何变化做出明确的测试预测。我们在一项针对阅读理解任务的两个版本的眼影研究中对此进行了测试，发现我们的模型成功说明了整个任务的阅读行为。因此，我们的工作提供了证据表明，任务效果可以建模为对任务需求的最佳适应。

Taking advice from others requires confidence in their competence. This is important for interaction with peers, but also for collaboration with social robots and artificial agents. Nonetheless, we do not always have access to information about others' competence or performance. In these uncertain environments, do our prior beliefs about the nature and the competence of our interacting partners modulate our willingness to rely on their judgments? In a joint perceptual decision making task, participants made perceptual judgments and observed the simulated estimates of either a human participant, a social humanoid robot or a computer. Then they could modify their estimates based on this feedback. Results show participants' belief about the nature of their partner biased their compliance with its judgments: participants were more influenced by the social robot than human and computer partners. This difference emerged strongly at the very beginning of the task and decreased with repeated exposure to empirical feedback on the partner's responses, disclosing the role of prior beliefs in social influence under uncertainty. Furthermore, the results of our functional task suggest an important difference between human-human and human-robot interaction in the absence of overt socially relevant signal from the partner: the former is modulated by social normative mechanisms, whereas the latter is guided by purely informational mechanisms linked to the perceived competence of the partner.

神经语言模型有可能支持人类写作。但是，关于其整合和对写作和产出的影响仍然存在问题。为了解决这个问题，我们设计并比较了两个用于写作的用户界面与移动设备上的AI，这些用户界面操纵主动性和控制级别：1）使用连续生成的文本编写，AI添加了逐字文字和用户转向。 2）编写建议，AI建议短语和用户从列表中选择。在监督的在线研究（n = 18）中，参与者使用了这些原型和无AI的基线。我们收集了触摸互动，关于灵感和作者的评分以及访谈数据。有了AI的建议，人们的写作不那么积极，但觉得他们是作者。连续生成的文本减少了这种感知的作者身份，但编辑行为增加了。在这两种设计中，AI都会增加文本长度，并被认为会影响措辞。我们的发现为UI设计决策对用户体验和共同创造系统的产出的影响增加了新的经验证据。

众所周知，端到端的神经NLP体系结构很难理解，这引起了近年来为解释性建模的许多努力。模型解释的基本原则是忠诚，即，解释应准确地代表模型预测背后的推理过程。这项调查首先讨论了忠诚的定义和评估及其对解释性的意义。然后，我们通过将方法分为五类来介绍忠实解释的最新进展：相似性方法，模型内部结构的分析，基于反向传播的方法，反事实干预和自我解释模型。每个类别将通过其代表性研究，优势和缺点来说明。最后，我们从它们的共同美德和局限性方面讨论了上述所有方法，并反思未来的工作方向忠实的解释性。对于有兴趣研究可解释性的研究人员，这项调查将为该领域提供可访问且全面的概述，为进一步探索提供基础。对于希望更好地了解自己的模型的用户，该调查将是一项介绍性手册，帮助选择最合适的解释方法。

Explainable AI (XAI) is widely viewed as a sine qua non for ever-expanding AI research. A better understanding of the needs of XAI users, as well as human-centered evaluations of explainable models are both a necessity and a challenge. In this paper, we explore how HCI and AI researchers conduct user studies in XAI applications based on a systematic literature review. After identifying and thoroughly analyzing 85 core papers with human-based XAI evaluations over the past five years, we categorize them along the measured characteristics of explanatory methods, namely trust, understanding, fairness, usability, and human-AI team performance. Our research shows that XAI is spreading more rapidly in certain application domains, such as recommender systems than in others, but that user evaluations are still rather sparse and incorporate hardly any insights from cognitive or social sciences. Based on a comprehensive discussion of best practices, i.e., common models, design choices, and measures in user studies, we propose practical guidelines on designing and conducting user studies for XAI researchers and practitioners. Lastly, this survey also highlights several open research directions, particularly linking psychological science and human-centered XAI.

我们提出了一项探索性定性研究，以了解作家如何与下一页建议相互作用。尽管对建议系统对写作的影响进行了一些定量研究，但几乎没有定性的工作来理解作家如何与建议系统互动及其如何影响他们的写作过程 - 特别是针对非本地但英国作家的。我们进行了一项研究，要求业余作家分别写两部电影评论，一本没有建议。我们发现作家以各种复杂的方式与下一页建议互动 - 作家能够抽象建议的多个部分并将其纳入他们的写作中 - 即使他们不同意整个建议。建议系统对写作过程也有各种影响 - 以独特的方式为写作过程的不同方面做出了影响。我们提出了一种用于与GPT-2写作的作家 - 探索互动模型，用于电影评论写作任务，然后是该模型可用于未来研究的方式，并概述了研究和设计的机会。

People constantly use language to learn about the world. Computational linguists have capitalized on this fact to build large language models (LLMs) that acquire co-occurrence-based knowledge from language corpora. LLMs achieve impressive performance on many tasks, but the robustness of their world knowledge has been questioned. Here, we ask: do LLMs acquire generalized knowledge about real-world events? Using curated sets of minimal sentence pairs (n=1215), we tested whether LLMs are more likely to generate plausible event descriptions compared to their implausible counterparts. We found that LLMs systematically distinguish possible and impossible events (The teacher bought the laptop vs. The laptop bought the teacher) but fall short of human performance when distinguishing likely and unlikely events (The nanny tutored the boy vs. The boy tutored the nanny). In follow-up analyses, we show that (i) LLM scores are driven by both plausibility and surface-level sentence features, (ii) LLMs generalize well across syntactic sentence variants (active vs passive) but less well across semantic sentence variants (synonymous sentences), (iii) some, but not all LLM deviations from ground-truth labels align with crowdsourced human judgments, and (iv) explicit event plausibility information emerges in middle LLM layers and remains high thereafter. Overall, our analyses reveal a gap in LLMs' event knowledge, highlighting their limitations as generalized knowledge bases. We conclude by speculating that the differential performance on impossible vs. unlikely events is not a temporary setback but an inherent property of LLMs, reflecting a fundamental difference between linguistic knowledge and world knowledge in intelligent systems.

我们介绍了Sparrow，这是一个寻求信息的对话代理，与提示的语言模型基线相比，训练有素，更有帮助，正确和无害。我们使用从人类反馈中的强化学习来培训我们的模型，以帮助人类评估者判断代理人的行为。首先，为了使我们的代理人更有帮助和无害，我们将良好对话的要求分解为代理人应遵循的自然语言规则，并分别向评估者询问每个规则。我们证明，这种崩溃使我们能够收集对代理行为的更多针对性的人类判断，并允许更有效的规则条件奖励模型。其次，我们的代理商在收集对模型声明的偏好判决时提供了支持事实主张的来源的证据。对于事实问题，麻雀提供的证据支持了78％的时间。比基线比基线更享受麻雀，同时对人类的对抗性探测更具弹性，在探测时只有8％的时间违反了我们的规则。最后，我们进行了广泛的分析，表明尽管我们的模型学会遵守我们的规则，但它可以表现出分布偏见。

人类语言中发现的最强大的模式之一是ZIPF的缩写定律，即更短的单词的趋势。自ZIPF开创性研究以来，该定律被视为压缩的体现，即形式的长度最小化 - 自然交流的普遍原则。尽管对语言进行优化的说法已经变得时尚，但衡量语言优化程度的尝试却相当稀缺。在这里，我们证明压缩在无例外的大量语言中表现出来，并且独立于测量单位。这两个单词长度都可以在书面语言的字符以及口语的持续时间中检测到。此外，为了衡量优化程度，我们得出了一个随机基线的简单公式，并提出了两个分数归一化的分数，即，它们相对于最小值和随机基线都进行了归一化。我们分析了这些和其他分数的理论和统计优势和缺点。利用最佳分数，我们首次量化了语言中单词长度的最佳程度。这表明当单词长度以字符测量时，语言平均被优化至62％或67％（取决于源），当单词长度及时测量时，平均而言，平均而言，平均而言，平均而言，平均而言，平均而言，平均至65％。通常，口语持续时间比字符中的书面单词长度更优化。除了这里报告的分析外，我们的工作还铺平了衡量其他物种发声或手势的最佳程度的方法，并将其与书面，口语或签名的人类语言进行比较。

There has been a recent resurgence in the area of explainable artificial intelligence as researchers and practitioners seek to make their algorithms more understandable. Much of this research is focused on explicitly explaining decisions or actions to a human observer, and it should not be controversial to say that looking at how humans explain to each other can serve as a useful starting point for explanation in artificial intelligence. However, it is fair to say that most work in explainable artificial intelligence uses only the researchers' intuition of what constitutes a 'good' explanation. There exists vast and valuable bodies of research in philosophy, psychology, and cognitive science of how people define, generate, select, evaluate, and present explanations, which argues that people employ certain cognitive biases and social expectations towards the explanation process. This paper argues that the field of explainable artificial intelligence should build on this existing research, and reviews relevant papers from philosophy, cognitive psychology/science, and social psychology, which study these topics. It draws out some important findings, and discusses ways that these can be infused with work on explainable artificial intelligence.

Incivility remains a major challenge for online discussion platforms, to such an extent that even conversations between well-intentioned users can often derail into uncivil behavior. Traditionally, platforms have relied on moderators to -- with or without algorithmic assistance -- take corrective actions such as removing comments or banning users. In this work we propose a complementary paradigm that directly empowers users by proactively enhancing their awareness about existing tension in the conversation they are engaging in and actively guides them as they are drafting their replies to avoid further escalation. As a proof of concept for this paradigm, we design an algorithmic tool that provides such proactive information directly to users, and conduct a user study in a popular discussion platform. Through a mixed methods approach combining surveys with a randomized controlled experiment, we uncover qualitative and quantitative insights regarding how the participants utilize and react to this information. Most participants report finding this proactive paradigm valuable, noting that it helps them to identify tension that they may have otherwise missed and prompts them to further reflect on their own replies and to revise them. These effects are corroborated by a comparison of how the participants draft their reply when our tool warns them that their conversation is at risk of derailing into uncivil behavior versus in a control condition where the tool is disabled. These preliminary findings highlight the potential of this user-centered paradigm and point to concrete directions for future implementations.

可解释的人工智能（XAI）的新兴领域旨在为当今强大但不透明的深度学习模型带来透明度。尽管本地XAI方法以归因图的形式解释了个体预测，从而确定了重要特征的发生位置（但没有提供有关其代表的信息），但全局解释技术可视化模型通常学会的编码的概念。因此，两种方法仅提供部分见解，并留下将模型推理解释的负担。只有少数当代技术旨在将本地和全球XAI背后的原则结合起来，以获取更多信息的解释。但是，这些方法通常仅限于特定的模型体系结构，或对培训制度或数据和标签可用性施加其他要求，这实际上使事后应用程序成为任意预训练的模型。在这项工作中，我们介绍了概念相关性传播方法（CRP）方法，该方法结合了XAI的本地和全球观点，因此允许回答“何处”和“ where”和“什么”问题，而没有其他约束。我们进一步介绍了相关性最大化的原则，以根据模型对模型的有用性找到代表性的示例。因此，我们提高了对激活最大化及其局限性的共同实践的依赖。我们证明了我们方法在各种环境中的能力，展示了概念相关性传播和相关性最大化导致了更加可解释的解释，并通过概念图表，概念组成分析和概念集合和概念子区和概念子区和概念子集和定量研究对模型的表示和推理提供了深刻的见解。它们在细粒度决策中的作用。

Deep Learning and Machine Learning based models have become extremely popular in text processing and information retrieval. However, the non-linear structures present inside the networks make these models largely inscrutable. A significant body of research has focused on increasing the transparency of these models. This article provides a broad overview of research on the explainability and interpretability of natural language processing and information retrieval methods. More specifically, we survey approaches that have been applied to explain word embeddings, sequence modeling, attention modules, transformers, BERT, and document ranking. The concluding section suggests some possible directions for future research on this topic.

尽管试图提高政治性别平等，但全球努力仍在努力确保女性的同等代表。这很可能与对权威妇女的性别偏见有关。在这项工作中，我们介绍了在线政治讨论中出现的性别偏见的全面研究。为此，我们在有关男性和女性政客的对话中收集了1000万条有关Reddit的评论，这使得对自动性别偏见检测进行了详尽的研究。我们不仅讨论了厌恶女性的语言，还解决了其他偏见的表现，例如以看似积极的情绪和主导地位归因于女性政客或描述符归因的差异的形式的仁慈性别歧视。最后，我们对调查语言和语言外暗示的政客进行了多方面的性别偏见研究。我们评估了5种不同类型的性别偏见，评估社交媒体语言和话语中存在的覆盖范围，组合，名义，感性和词汇偏见。总体而言，我们发现，与以前的研究相反，覆盖范围和情感偏见表明对女性政客的公共兴趣平等。名义和词汇分析的结果并没有明显的敌对或仁慈的性别歧视，这表明这种兴趣不像男性政客那样专业或尊重。女性政客通常以其名字命名，并与他们的身体，衣服或家庭有关。这是一种与男性相似的治疗方法。在现在被禁止的极右翼子列表中，这种差异最大，尽管性别偏见的差异仍然出现在右和左倾的子列表中。我们将策划的数据集释放给公众以进行未来研究。

With the increasing demand for predictable and accountable Artificial Intelligence, the ability to explain or justify recommender systems results by specifying how items are suggested, or why they are relevant, has become a primary goal. However, current models do not explicitly represent the services and actors that the user might encounter during the overall interaction with an item, from its selection to its usage. Thus, they cannot assess their impact on the user's experience. To address this issue, we propose a novel justification approach that uses service models to (i) extract experience data from reviews concerning all the stages of interaction with items, at different granularity levels, and (ii) organize the justification of recommendations around those stages. In a user study, we compared our approach with baselines reflecting the state of the art in the justification of recommender systems results. The participants evaluated the Perceived User Awareness Support provided by our service-based justification models higher than the one offered by the baselines. Moreover, our models received higher Interface Adequacy and Satisfaction evaluations by users having different levels of Curiosity or low Need for Cognition (NfC). Differently, high NfC participants preferred a direct inspection of item reviews. These findings encourage the adoption of service models to justify recommender systems results but suggest the investigation of personalization strategies to suit diverse interaction needs.