该报告解释，实施和扩展了“更紧密的变化界限不一定更好”所介绍的作品（T Rainforth等，2018）。我们提供了理论和经验证据，这些证据增加了重要性的重要性数量$ k $在重要性加权自动编码器（IWAE）中（Burda等，2016）降低了推理中梯度估计量的信噪比（SNR）网络，从而影响完整的学习过程。换句话说，即使增加$ k $减少了梯度的标准偏差，但它也会更快地降低真实梯度的幅度，从而增加梯度更新的相对差异。进行广泛的实验以了解$ k $的重要性。这些实验表明，更紧密的变化界限对生成网络有益，而宽松的边界对推理网络来说是可取的。通过这些见解，可以实施和研究三种方法：部分重要性加权自动编码器（PIWAE），倍增重要性加权自动编码器（MIWAE）和组合重要性加权自动编码器（CIWAE）。这三种方法中的每一种都需要IWAE作为一种特殊情况，但采用不同的重量权重，以确保较高的梯度估计器的SNR。在我们的研究和分析中，这些算法的疗效在多个数据集（如MNIST和Omniglot）上进行了测试。最后，我们证明了三种呈现的IWAE变化能够产生近似后验分布，这些分布与IWAE更接近真正的后验分布，同时匹配IWAE生成网络的性能，或者在PIWAE的情况下可能超过其表现。

近年来，深度学习导致了在城市驾驶场景中移动（即具有运动能力）物体的检测方面取得的巨大进展。监督方法通常需要大型培训集的注释；因此，人们对利用弱，半或自我监督的方法避免这种情况非常兴趣，并取得了很大的成功。虽然弱和半监督的方法需要一些注释，但自我监督的方法已经使用了诸如运动之类的线索来完全减轻注释的需求。但是，完全没有注释通常会降低其性能，而在运动组进行分组期间出现的歧义可以抑制其找到准确的物体边界的能力。在本文中，我们提出了一种称为SCT的新的自制移动对象检测方法。这同时使用运动提示和预期对象大小来提高检测性能，并预测3D方向边界框的密集网格以改善对象发现。我们在Kitti跟踪基准上的最先进的自我监督的移动对象检测方法TCR极大地超过了，并且实现了全面监督的PV-RCNN ++方法的30％以内IOUS <= 0.5。

共同监督的深度学习方法的关节深度和自我运动估计可以产生准确的轨迹，而无需地面真相训练数据。但是，由于通常会使用光度损失，因此当这些损失所产生的假设（例如时间照明一致性，静态场景以及缺少噪声和遮挡）时，它们的性能会显着降解。这限制了它们用于例如夜间序列倾向于包含许多点光源（包括在动态对象上）和较暗图像区域中的低信噪比（SNR）。在本文中，我们展示了如何使用三种技术的组合来允许现有的光度损失在白天和夜间图像中起作用。首先，我们引入了每个像素神经强度转化，以补偿连续帧之间发生的光变化。其次，我们预测了每个像素的残差流图，我们用来纠正由网络估计的自我运动和深度引起的重新注入对应关系。第三，我们将训练图像降低，以提高方法的鲁棒性和准确性。这些更改使我们可以在白天和夜间图像中训练单个模型，而无需单独的编码器或诸如现有方法（例如现有方法）的额外功能网络。我们对具有挑战性的牛津机器人数据集进行了广泛的实验和消融研究，以证明我们方法对白天和夜间序列的疗效。

Understanding the informative structures of scenes is essential for low-level vision tasks. Unfortunately, it is difficult to obtain a concrete visual definition of the informative structures because influences of visual features are task-specific. In this paper, we propose a single general neural network architecture for extracting task-specific structure guidance for scenes. To do this, we first analyze traditional spectral clustering methods, which computes a set of eigenvectors to model a segmented graph forming small compact structures on image domains. We then unfold the traditional graph-partitioning problem into a learnable network, named \textit{Scene Structure Guidance Network (SSGNet)}, to represent the task-specific informative structures. The SSGNet yields a set of coefficients of eigenvectors that produces explicit feature representations of image structures. In addition, our SSGNet is light-weight ($\sim$ 55K parameters), and can be used as a plug-and-play module for off-the-shelf architectures. We optimize the SSGNet without any supervision by proposing two novel training losses that enforce task-specific scene structure generation during training. Our main contribution is to show that such a simple network can achieve state-of-the-art results for several low-level vision applications including joint upsampling and image denoising. We also demonstrate that our SSGNet generalizes well on unseen datasets, compared to existing methods which use structural embedding frameworks. Our source codes are available at https://github.com/jsshin98/SSGNet.

In many domains such as transportation and logistics, search and rescue, or cooperative surveillance, tasks are pending to be allocated with the consideration of possible execution uncertainties. Existing task coordination algorithms either ignore the stochastic process or suffer from the computational intensity. Taking advantage of the weakly coupled feature of the problem and the opportunity for coordination in advance, we propose a decentralized auction-based coordination strategy using a newly formulated score function which is generated by forming the problem into task-constrained Markov decision processes (MDPs). The proposed method guarantees convergence and at least 50% optimality in the premise of a submodular reward function. Furthermore, for the implementation on large-scale applications, an approximate variant of the proposed method, namely Deep Auction, is also suggested with the use of neural networks, which is evasive of the troublesome for constructing MDPs. Inspired by the well-known actor-critic architecture, two Transformers are used to map observations to action probabilities and cumulative rewards respectively. Finally, we demonstrate the performance of the two proposed approaches in the context of drone deliveries, where the stochastic planning for the drone league is cast into a stochastic price-collecting Vehicle Routing Problem (VRP) with time windows. Simulation results are compared with state-of-the-art methods in terms of solution quality, planning efficiency and scalability.

In robotics and computer vision communities, extensive studies have been widely conducted regarding surveillance tasks, including human detection, tracking, and motion recognition with a camera. Additionally, deep learning algorithms are widely utilized in the aforementioned tasks as in other computer vision tasks. Existing public datasets are insufficient to develop learning-based methods that handle various surveillance for outdoor and extreme situations such as harsh weather and low illuminance conditions. Therefore, we introduce a new large-scale outdoor surveillance dataset named eXtremely large-scale Multi-modAl Sensor dataset (X-MAS) containing more than 500,000 image pairs and the first-person view data annotated by well-trained annotators. Moreover, a single pair contains multi-modal data (e.g. an IR image, an RGB image, a thermal image, a depth image, and a LiDAR scan). This is the first large-scale first-person view outdoor multi-modal dataset focusing on surveillance tasks to the best of our knowledge. We present an overview of the proposed dataset with statistics and present methods of exploiting our dataset with deep learning-based algorithms. The latest information on the dataset and our study are available at https://github.com/lge-robot-navi, and the dataset will be available for download through a server.

Task-oriented dialogue systems often assist users with personal or confidential matters. For this reason, the developers of such a system are generally prohibited from observing actual usage. So how can they know where the system is failing and needs more training data or new functionality? In this work, we study ways in which realistic user utterances can be generated synthetically, to help increase the linguistic and functional coverage of the system, without compromising the privacy of actual users. To this end, we propose a two-stage Differentially Private (DP) generation method which first generates latent semantic parses, and then generates utterances based on the parses. Our proposed approach improves MAUVE by 3.8$\times$ and parse tree node-type overlap by 1.4$\times$ relative to current approaches for private synthetic data generation, improving both on fluency and semantic coverage. We further validate our approach on a realistic domain adaptation task of adding new functionality from private user data to a semantic parser, and show gains of 1.3$\times$ on its accuracy with the new feature.

Task-oriented dialogue (TOD) systems are mainly based on the slot-filling-based TOD (SF-TOD) framework, in which dialogues are broken down into smaller, controllable units (i.e., slots) to fulfill a specific task. A series of approaches based on this framework achieved remarkable success on various TOD benchmarks. However, we argue that the current TOD benchmarks are limited to surrogate real-world scenarios and that the current TOD models are still a long way from unraveling the scenarios. In this position paper, we first identify current status and limitations of SF-TOD systems. After that, we explore the WebTOD framework, the alternative direction for building a scalable TOD system when a web/mobile interface is available. In WebTOD, the dialogue system learns how to understand the web/mobile interface that the human agent interacts with, powered by a large-scale language model.

Any classifier can be "smoothed out" under Gaussian noise to build a new classifier that is provably robust to $\ell_2$-adversarial perturbations, viz., by averaging its predictions over the noise via randomized smoothing. Under the smoothed classifiers, the fundamental trade-off between accuracy and (adversarial) robustness has been well evidenced in the literature: i.e., increasing the robustness of a classifier for an input can be at the expense of decreased accuracy for some other inputs. In this paper, we propose a simple training method leveraging this trade-off to obtain robust smoothed classifiers, in particular, through a sample-wise control of robustness over the training samples. We make this control feasible by using "accuracy under Gaussian noise" as an easy-to-compute proxy of adversarial robustness for an input. Specifically, we differentiate the training objective depending on this proxy to filter out samples that are unlikely to benefit from the worst-case (adversarial) objective. Our experiments show that the proposed method, despite its simplicity, consistently exhibits improved certified robustness upon state-of-the-art training methods. Somewhat surprisingly, we find these improvements persist even for other notions of robustness, e.g., to various types of common corruptions.

Test-time adaptation (TTA) has attracted significant attention due to its practical properties which enable the adaptation of a pre-trained model to a new domain with only target dataset during the inference stage. Prior works on TTA assume that the target dataset comes from the same distribution and thus constitutes a single homogeneous domain. In practice, however, the target domain can contain multiple homogeneous domains which are sufficiently distinctive from each other and those multiple domains might occur cyclically. Our preliminary investigation shows that domain-specific TTA outperforms vanilla TTA treating compound domain (CD) as a single one. However, domain labels are not available for CD, which makes domain-specific TTA not practicable. To this end, we propose an online clustering algorithm for finding pseudo-domain labels to obtain similar benefits as domain-specific configuration and accumulating knowledge of cyclic domains effectively. Moreover, we observe that there is a significant discrepancy in terms of prediction quality among samples, especially in the CD context. This further motivates us to boost its performance with gradient denoising by considering the image-wise similarity with the source distribution. Overall, the key contribution of our work lies in proposing a highly significant new task compound domain test-time adaptation (CD-TTA) on semantic segmentation as well as providing a strong baseline to facilitate future works to benchmark.