Novel view synthesis and 3D modeling using implicit neural field representation are shown to be very effective for calibrated multi-view cameras. Such representations are known to benefit from additional geometric and semantic supervision. Most existing methods that exploit additional supervision require dense pixel-wise labels or localized scene priors. These methods cannot benefit from high-level vague scene priors provided in terms of scenes' descriptions. In this work, we aim to leverage the geometric prior of Manhattan scenes to improve the implicit neural radiance field representations. More precisely, we assume that only the knowledge of the scene (under investigation) being Manhattan is known - with no additional information whatsoever - with an unknown Manhattan coordinate frame. Such high-level prior is then used to self-supervise the surface normals derived explicitly in the implicit neural fields. Our modeling allows us to group the derived normals, followed by exploiting their orthogonality constraints for self-supervision. Our exhaustive experiments on datasets of diverse indoor scenes demonstrate the significant benefit of the proposed method over the established baselines.

在大多数情况下，有条件的图像生成可以被认为是对图像理解过程的反转。由于通用图像理解涉及解决多个任务，因此自然要通过多条件来生成图像。但是，由于异质性和（实际上）可用条件标签的稀疏性，多条件图像生成是一个非常具有挑战性的问题。在这项工作中，我们提出了一种新型的神经结构，以解决空间多条件标签的异质性和稀疏性问题。我们选择的空间条件（例如语义和深度）是由它具有更好地控制图像生成过程的承诺所驱动的。所提出的方法使用类似变压器的体系结构操作像素，该架构将可用的标签作为输入令牌接收，以将其合并在学习的标签均匀空间中。然后，合并的标签用于通过有条件的生成对抗训练进行图像生成。在此过程中，通过简单地将与所需位置的缺失标签相对应的输入令牌掉下来处理标签的稀疏性，这要归功于提议的像素操作架构。我们在三个基准数据集上进行的实验证明了我们的方法比最新的基准和比较基线的明显优势。源代码将公开可用。

我们在视觉变压器中引入完全随机层，而不会导致任何严重的性能下降。额外的随机性提高了视觉特征的鲁棒性，并加强了隐私。在该过程中，在训练和推理期间使用具有完全随机参数的线性层，以改变每个多层Perceptron的特征激活。这种随机线性操作保留了由通过共用多层Perceptron的令牌形成的拓扑结构。此操作鼓励学习识别任务依赖令牌的拓扑结构，而不是它们的值，而不是它们的值，这反过来提供了可视化功能的所需的鲁棒性和隐私。在本文中，我们使用我们的特性进行三种不同的应用程序，即对抗鲁棒性，网络校准和特征隐私。我们的功能为这些任务提供令人兴奋的结果。此外，我们展示了联合和转移学习的实验设置，其中具有随机层的视觉变压器再次显示出良好的表现。我们的源代码将公开可用。

Text classifiers have promising applications in high-stake tasks such as resume screening and content moderation. These classifiers must be fair and avoid discriminatory decisions by being invariant to perturbations of sensitive attributes such as gender or ethnicity. However, there is a gap between human intuition about these perturbations and the formal similarity specifications capturing them. While existing research has started to address this gap, current methods are based on hardcoded word replacements, resulting in specifications with limited expressivity or ones that fail to fully align with human intuition (e.g., in cases of asymmetric counterfactuals). This work proposes novel methods for bridging this gap by discovering expressive and intuitive individual fairness specifications. We show how to leverage unsupervised style transfer and GPT-3's zero-shot capabilities to automatically generate expressive candidate pairs of semantically similar sentences that differ along sensitive attributes. We then validate the generated pairs via an extensive crowdsourcing study, which confirms that a lot of these pairs align with human intuition about fairness in the context of toxicity classification. Finally, we show how limited amounts of human feedback can be leveraged to learn a similarity specification that can be used to train downstream fairness-aware models.

We present a method for controlling a swarm using its spectral decomposition -- that is, by describing the set of trajectories of a swarm in terms of a spatial distribution throughout the operational domain -- guaranteeing scale invariance with respect to the number of agents both for computation and for the operator tasked with controlling the swarm. We use ergodic control, decentralized across the network, for implementation. In the DARPA OFFSET program field setting, we test this interface design for the operator using the STOMP interface -- the same interface used by Raytheon BBN throughout the duration of the OFFSET program. In these tests, we demonstrate that our approach is scale-invariant -- the user specification does not depend on the number of agents; it is persistent -- the specification remains active until the user specifies a new command; and it is real-time -- the user can interact with and interrupt the swarm at any time. Moreover, we show that the spectral/ergodic specification of swarm behavior degrades gracefully as the number of agents goes down, enabling the operator to maintain the same approach as agents become disabled or are added to the network. We demonstrate the scale-invariance and dynamic response of our system in a field relevant simulator on a variety of tactical scenarios with up to 50 agents. We also demonstrate the dynamic response of our system in the field with a smaller team of agents. Lastly, we make the code for our system available.

Accurate speed estimation of road vehicles is important for several reasons. One is speed limit enforcement, which represents a crucial tool in decreasing traffic accidents and fatalities. Compared with other research areas and domains, the number of available datasets for vehicle speed estimation is still very limited. We present a dataset of on-road audio-video recordings of single vehicles passing by a camera at known speeds, maintained stable by the on-board cruise control. The dataset contains thirteen vehicles, selected to be as diverse as possible in terms of manufacturer, production year, engine type, power and transmission, resulting in a total of $ 400 $ annotated audio-video recordings. The dataset is fully available and intended as a public benchmark to facilitate research in audio-video vehicle speed estimation. In addition to the dataset, we propose a cross-validation strategy which can be used in a machine learning model for vehicle speed estimation. Two approaches to training-validation split of the dataset are proposed.

Tabular biomedical data is often high-dimensional but with a very small number of samples. Although recent work showed that well-regularised simple neural networks could outperform more sophisticated architectures on tabular data, they are still prone to overfitting on tiny datasets with many potentially irrelevant features. To combat these issues, we propose Weight Predictor Network with Feature Selection (WPFS) for learning neural networks from high-dimensional and small sample data by reducing the number of learnable parameters and simultaneously performing feature selection. In addition to the classification network, WPFS uses two small auxiliary networks that together output the weights of the first layer of the classification model. We evaluate on nine real-world biomedical datasets and demonstrate that WPFS outperforms other standard as well as more recent methods typically applied to tabular data. Furthermore, we investigate the proposed feature selection mechanism and show that it improves performance while providing useful insights into the learning task.

Genome-wide studies leveraging recent high-throughput sequencing technologies collect high-dimensional data. However, they usually include small cohorts of patients, and the resulting tabular datasets suffer from the "curse of dimensionality". Training neural networks on such datasets is typically unstable, and the models overfit. One problem is that modern weight initialisation strategies make simplistic assumptions unsuitable for small-size datasets. We propose Graph-Conditioned MLP, a novel method to introduce priors on the parameters of an MLP. Instead of randomly initialising the first layer, we condition it directly on the training data. More specifically, we create a graph for each feature in the dataset (e.g., a gene), where each node represents a sample from the same dataset (e.g., a patient). We then use Graph Neural Networks (GNNs) to learn embeddings from these graphs and use the embeddings to initialise the MLP's parameters. Our approach opens the prospect of introducing additional biological knowledge when constructing the graphs. We present early results on 7 classification tasks from gene expression data and show that GC-MLP outperforms an MLP.

Classification of cancer cellularity within tissue samples is currently a manual process performed by pathologists. This process of correctly determining cancer cellularity can be time intensive. Deep Learning (DL) techniques in particular have become increasingly more popular for this purpose, due to the accuracy and performance they exhibit, which can be comparable to the pathologists. This work investigates the capabilities of two DL approaches to assess cancer cellularity in whole slide images (WSI) in the SPIE-AAPM-NCI BreastPathQ challenge dataset. The effects of training on augmented data via rotations, and combinations of multiple architectures into a single network were analyzed using a modified Kendall Tau-b prediction probability metric known as the average prediction probability PK. A deep, transfer learned, Convolutional Neural Network (CNN) InceptionV3 was used as a baseline, achieving an average PK value of 0.884, showing improvement from the average PK value of 0.83 achieved by pathologists. The network was then trained on additional training datasets which were rotated between 1 and 360 degrees, which saw a peak increase of PK up to 4.2%. An additional architecture consisting of the InceptionV3 network and VGG16, a shallow, transfer learned CNN, was combined in a parallel architecture. This parallel architecture achieved a baseline average PK value of 0.907, a statistically significantly improvement over either of the architectures' performances separately (p<0.0001 by unpaired t-test).

在我们不断变化的气候中，使用模型来评估天气和气候对社会和企业的后续后果的风险及其后续后果至关重要。这种模型的操作在历史上是定制的，并限制在特定的计算基础架构，驱动数据集和预定义的配置上。这些约束通过缩放模型运行并将模型掌握在感兴趣的用户手中。在这里，我们提出了一个基于云的模块化框架，用于部署和操作地理空间模型，最初应用于气候影响。气候冲击建模框架（CIMF）可以以动态和灵活的方式部署模块化工作流程。用户可以以简化的方式指定工作流程组件，然后可以轻松地将这些组件组织成不同的配置，以以不同的方式和不同的尺度评估风险。这还可以使不同的模型（物理模拟或机器学习模型）和工作流程连接以产生合并的风险评估。洪水建模被用作端到端的示例，以证明CIMF的操作。