骨科疾病在马匹中常见，通常导致安乐死，这通常可以通过早期的检测来避免。这些条件通常会产生不同程度的微妙长期疼痛。培训视觉疼痛识别方法具有描绘这种疼痛的视频数据是挑战性的，因为所产生的疼痛行为也是微妙的，稀疏出现，变得不同，使得甚至是专家兰德尔的挑战，为数据提供准确的地面真实性。我们表明，一款专业培训的模型，仅涉及急性实验疼痛的马匹（标签不那么暧昧）可以帮助识别更微妙的骨科疼痛显示。此外，我们提出了一个问题的人类专家基线，以及对各种领域转移方法的广泛实证研究以及由疼痛识别方法检测到矫形数据集的清洁实验疼痛中的疼痛识别方法检测到的内容。最后，这伴随着围绕现实世界动物行为数据集所带来的挑战以及如何为类似的细粒度行动识别任务建立最佳实践的讨论。我们的代码可在https://github.com/sofiabroome/painface-recognition获得。

Knowledge Distillation (KD) is a commonly used technique for improving the generalization of compact Pre-trained Language Models (PLMs) on downstream tasks. However, such methods impose the additional burden of training a separate teacher model for every new dataset. Alternatively, one may directly work on the improvement of the optimization procedure of the compact model toward better generalization. Recent works observe that the flatness of the local minimum correlates well with better generalization. In this work, we adapt Stochastic Weight Averaging (SWA), a method encouraging convergence to a flatter minimum, to fine-tuning PLMs. We conduct extensive experiments on various NLP tasks (text classification, question answering, and generation) and different model architectures and demonstrate that our adaptation improves the generalization without extra computation cost. Moreover, we observe that this simple optimization technique is able to outperform the state-of-the-art KD methods for compact models.

Recent work has demonstrated that natural language processing techniques can support consumer protection by automatically detecting unfair clauses in the Terms of Service (ToS) Agreement. This work demonstrates that transformer-based ToS analysis systems are vulnerable to adversarial attacks. We conduct experiments attacking an unfair-clause detector with universal adversarial triggers. Experiments show that a minor perturbation of the text can considerably reduce the detection performance. Moreover, to measure the detectability of the triggers, we conduct a detailed human evaluation study by collecting both answer accuracy and response time from the participants. The results show that the naturalness of the triggers remains key to tricking readers.

Causal structure learning from observational data remains a non-trivial task due to various factors such as finite sampling, unobserved confounding factors, and measurement errors. Constraint-based and score-based methods tend to suffer from high computational complexity due to the combinatorial nature of estimating the directed acyclic graph (DAG). Motivated by the `Cause-Effect Pair' NIPS 2013 Workshop on Causality Challenge, in this paper, we take a different approach and generate a probability distribution over all possible graphs informed by the cause-effect pair features proposed in response to the workshop challenge. The goal of the paper is to propose new methods based on this probabilistic information and compare their performance with traditional and state-of-the-art approaches. Our experiments, on both synthetic and real datasets, show that our proposed methods not only have statistically similar or better performances than some traditional approaches but also are computationally faster.

Traditional electrical power grids have long suffered from operational unreliability, instability, inflexibility, and inefficiency. Smart grids (or smart energy systems) continue to transform the energy sector with emerging technologies, renewable energy sources, and other trends. Artificial intelligence (AI) is being applied to smart energy systems to process massive and complex data in this sector and make smart and timely decisions. However, the lack of explainability and governability of AI is a major concern for stakeholders hindering a fast uptake of AI in the energy sector. This paper provides a review of AI explainability and governance in smart energy systems. We collect 3,568 relevant papers from the Scopus database, automatically discover 15 parameters or themes for AI governance in energy and elaborate the research landscape by reviewing over 150 papers and providing temporal progressions of the research. The methodology for discovering parameters or themes is based on "deep journalism", our data-driven deep learning-based big data analytics approach to automatically discover and analyse cross-sectional multi-perspective information to enable better decision-making and develop better instruments for governance. The findings show that research on AI explainability in energy systems is segmented and narrowly focussed on a few AI traits and energy system problems. This paper deepens our knowledge of AI governance in energy and is expected to help governments, industry, academics, energy prosumers, and other stakeholders to understand the landscape of AI in the energy sector, leading to better design, operations, utilisation, and risk management of energy systems.

恢复质量差的图像与一组混合伪影对于可靠的诊断起着至关重要的作用。现有的研究集中在特定的恢复问题上，例如图像过度，去核和暴露校正，通常对伪影类型和严重性有很强的假设。作为盲X射线恢复的先驱研究，我们提出了一个通用图像恢复和分类的联合模型：恢复分类为分类的生成对抗网络（R2C-GAN）。这种共同优化的模型使恢复后保持任何疾病完整。因此，由于X射线图像质量的提高，这自然会导致更高的诊断性能。为了实现这一关键目标，我们将恢复任务定义为图像到图像的翻译问题，从差异，模糊或暴露不足/暴露不足的图像到高质量的图像域。提出的R2C-GAN模型能够使用未配对的训练样本在两个域之间学习前进和逆变换。同时，联合分类在恢复过程中保留了疾病标签。此外，R2C-GAN配备了操作层/神经元，可降低网络深度，并进一步增强恢复和分类性能。拟议的联合模型对2019年冠状病毒病（COVID-19）分类的卡塔-COV19数据集进行了广泛的评估。拟议的恢复方法达到了90％以上的F1得分，这显着高于任何深层模型的性能。此外，在定性分析中，R2C-GAN的恢复性能得到了一群医生的批准。我们在https://github.com/meteahishali/r2c-gan上共享软件实施。

深度学习已在许多神经影像应用中有效。但是，在许多情况下，捕获与小血管疾病有关的信息的成像序列的数量不足以支持数据驱动的技术。此外，基于队列的研究可能并不总是具有用于准确病变检测的最佳或必需成像序列。因此，有必要确定哪些成像序列对于准确检测至关重要。在这项研究中，我们旨在找到磁共振成像（MRI）序列的最佳组合，以深入基于学习的肿瘤周围空间（EPV）。为此，我们实施了一个有效的轻巧U-NET，适用于EPVS检测，并全面研究了来自易感加权成像（SWI），流体侵入的反转恢复（FLAIR），T1加权（T1W）和T2的不同信息组合 - 加权（T2W）MRI序列。我们得出的结论是，T2W MRI对于准确的EPV检测最为重要，并且在深神经网络中掺入SWI，FLAIR和T1W MRI可能会使精度的提高无关。

自从各种任务的自动化开始以来，自动驾驶车辆一直引起人们的兴趣。人类容易疲惫，在道路上的响应时间缓慢，最重要的是，每年约有135万道路交通事故死亡，这已经是一项危险的任务。预计自动驾驶可以减少世界上驾驶事故的数量，这就是为什么这个问题对研究人员感兴趣的原因。目前，自动驾驶汽车在使车辆自动驾驶时使用不同的算法来实现各种子问题。我们将重点关注增强学习算法，更具体地说是Q学习算法和增强拓扑的神经进化（NEAT），即进化算法和人工神经网络的组合，以训练模型代理，以学习如何在给定路径上驱动。本文将重点介绍上述两种算法之间的比较。

面部情感识别是识别心理学用来诊断患者的重要工具之一。面部和面部情感识别是机器学习卓越的领域。由于不同的环境，例如照明条件，姿势变化，偏航运动和遮挡，面部情绪识别是对数字图像处理的开放挑战。深度学习方法已显示出图像识别的显着改善。但是，准确性和时间仍然需要改进。这项研究旨在在训练期间提高面部情绪识别的准确性，并使用Extreme Learning Machine（CNNeelm）增强的修改后的卷积神经网络减少处理时间。该系统需要（CNNeelm）提高培训期间图像注册的准确性。此外，该系统通过拟议的CNNeelm模型认识到六种面部情绪快乐，悲伤，厌恶，恐惧，惊喜和中立。研究表明，与经过改进的随机梯度下降（SGD）技术相比，总体面部情绪识别精度的提高了2％。借助Extreme Learning Machine（ELM）分类器，处理时间从113ms中降至65ms，可以从20fps的视频剪辑中平滑地对每个帧进行分类。使用预先训练的InceptionV3模型，建议使用JAFFE，CK+和FER2013表达数据集训练所提出的CNNeelm模型。仿真结果显示出准确性和处理时间的显着改善，使该模型适合视频分析过程。此外，该研究解决了处理面部图像所需的大量处理时间的问题。

可编程逻辑控制器（PLC）推动对社会至关重要的工业过程，例如水处理和分配，电力和燃料网络。搜索引擎（例如Shodan）强调说，可编程逻辑控制器（PLC）经常暴露于Internet，这是安全设置的错误配置的主要原因之一。这导致了一个问题 - 为什么这些错误配置会发生，具体而言，安全控制的可用性是否起作用？迄今为止，尚未研究配置PLC安全机制的可用性。我们通过基于任务的研究和随后的半结构化访谈（n = 19）介绍了第一次调查。我们探索PLC连接配置和两个关键安全机制的可用性（即访问级别和用户管理）。我们发现使用不熟悉的标签，布局和误导性术语加剧了已经复杂的配置安全机制的过程。我们的结果揭示了对安全控制的各种看法，以及设计约束（例如，安全性和缺乏定期更新）如何（由于这种系统的长期性质），为实现现代HCI和可用性原则的实现提供了重大挑战。基于这些发现，我们提供了设计建议，以在工业环境中与IT同行提供可用的安全性。