In this paper, we propose an end-to-end Retrieval-Augmented Visual Language Model (REVEAL) that learns to encode world knowledge into a large-scale memory, and to retrieve from it to answer knowledge-intensive queries. REVEAL consists of four key components: the memory, the encoder, the retriever and the generator. The large-scale memory encodes various sources of multimodal world knowledge (e.g. image-text pairs, question answering pairs, knowledge graph triplets, etc) via a unified encoder. The retriever finds the most relevant knowledge entries in the memory, and the generator fuses the retrieved knowledge with the input query to produce the output. A key novelty in our approach is that the memory, encoder, retriever and generator are all pre-trained end-to-end on a massive amount of data. Furthermore, our approach can use a diverse set of multimodal knowledge sources, which is shown to result in significant gains. We show that REVEAL achieves state-of-the-art results on visual question answering and image captioning.

深度学习的最新进展依赖于大型标签的数据集来培训大容量模型。但是，以时间和成本效益的方式收集大型数据集通常会导致标签噪声。我们提出了一种从嘈杂的标签中学习的方法，该方法利用特征空间中的训练示例之间的相似性，鼓励每个示例的预测与其最近的邻居相似。与使用多个模型或不同阶段的训练算法相比，我们的方法采用了简单，附加的正规化项的形式。它可以被解释为经典的，偏置标签传播算法的归纳版本。我们在数据集上彻底评估我们的方法评估合成（CIFAR-10，CIFAR-100）和现实（迷你网络，网络vision，Clotsing1m，Mini-Imagenet-Red）噪声，并实现竞争性或最先进的精度，在所有人之间。

真实世界的图像通常是通过对每级图像数量的显着不平衡的特征，导致长尾的分布。长尾视觉识别的有效和简单的方法是分别学习特征表示和分类器，分别使用实例和类平衡采样。在这项工作中，我们介绍一个新的框架，通过键观察，即使用实例采样学习的特征表示远远不受长尾设置的最佳选择。我们的主要贡献是一种新的培训方法，称为类别平衡蒸馏（CBD），其利用知识蒸馏来增强特征表示。 CBD允许特征表示在第二阶段的老师指导的第二次培训阶段演变。第二阶段使用类平衡的采样，以专注于非代表性的类。此框架可以自然地适应多个教师的使用，从模型的集合中解锁信息以增强识别能力。我们的实验表明，所提出的技术始终如一地优于本领域的长尾识别基准，例如想象群 - LT，Inaturatibry17和Inaturation18。

Semi-supervised learning is becoming increasingly important because it can combine data carefully labeled by humans with abundant unlabeled data to train deep neural networks. Classic methods on semi-supervised learning that have focused on transductive learning have not been fully exploited in the inductive framework followed by modern deep learning. The same holds for the manifold assumption-that similar examples should get the same prediction. In this work, we employ a transductive label propagation method that is based on the manifold assumption to make predictions on the entire dataset and use these predictions to generate pseudo-labels for the unlabeled data and train a deep neural network. At the core of the transductive method lies a nearest neighbor graph of the dataset that we create based on the embeddings of the same network. Therefore our learning process iterates between these two steps. We improve performance on several datasets especially in the few labels regime and show that our work is complementary to current state of the art.

We improve the understanding of the $\textit{golden ratio algorithm}$, which solves monotone variational inequalities (VI) and convex-concave min-max problems via the distinctive feature of adapting the step sizes to the local Lipschitz constants. Adaptive step sizes not only eliminate the need to pick hyperparameters, but they also remove the necessity of global Lipschitz continuity and can increase from one iteration to the next. We first establish the equivalence of this algorithm with popular VI methods such as reflected gradient, Popov or optimistic gradient descent-ascent in the unconstrained case with constant step sizes. We then move on to the constrained setting and introduce a new analysis that allows to use larger step sizes, to complete the bridge between the golden ratio algorithm and the existing algorithms in the literature. Doing so, we actually eliminate the link between the golden ratio $\frac{1+\sqrt{5}}{2}$ and the algorithm. Moreover, we improve the adaptive version of the algorithm, first by removing the maximum step size hyperparameter (an artifact from the analysis) to improve the complexity bound, and second by adjusting it to nonmonotone problems with weak Minty solutions, with superior empirical performance.

Researchers are doing intensive work on satellite images due to the information it contains with the development of computer vision algorithms and the ease of accessibility to satellite images. Building segmentation of satellite images can be used for many potential applications such as city, agricultural, and communication network planning. However, since no dataset exists for every region, the model trained in a region must gain generality. In this study, we trained several models in China and post-processing work was done on the best model selected among them. These models are evaluated in the Chicago region of the INRIA dataset. As can be seen from the results, although state-of-art results in this area have not been achieved, the results are promising. We aim to present our initial experimental results of a building segmentation from satellite images in this study.

This paper presents the preliminary findings of a semi-supervised segmentation method for extracting roads from sattelite images. Artificial Neural Networks and image segmentation methods are among the most successful methods for extracting road data from satellite images. However, these models require large amounts of training data from different regions to achieve high accuracy rates. In cases where this data needs to be of more quantity or quality, it is a standard method to train deep neural networks by transferring knowledge from annotated data obtained from different sources. This study proposes a method that performs path segmentation with semi-supervised learning methods. A semi-supervised field adaptation method based on pseudo-labeling and Minimum Class Confusion method has been proposed, and it has been observed to increase performance in targeted datasets.

In this paper, we introduce a novel optimization algorithm for machine learning model training called Normalized Stochastic Gradient Descent (NSGD) inspired by Normalized Least Mean Squares (NLMS) from adaptive filtering. When we train a high-complexity model on a large dataset, the learning rate is significantly important as a poor choice of optimizer parameters can lead to divergence. The algorithm updates the new set of network weights using the stochastic gradient but with $\ell_1$ and $\ell_2$-based normalizations on the learning rate parameter similar to the NLMS algorithm. Our main difference from the existing normalization methods is that we do not include the error term in the normalization process. We normalize the update term using the input vector to the neuron. Our experiments present that the model can be trained to a better accuracy level on different initial settings using our optimization algorithm. In this paper, we demonstrate the efficiency of our training algorithm using ResNet-20 and a toy neural network on different benchmark datasets with different initializations. The NSGD improves the accuracy of the ResNet-20 from 91.96\% to 92.20\% on the CIFAR-10 dataset.

Extracting building heights from satellite images is an active research area used in many fields such as telecommunications, city planning, etc. Many studies utilize DSM (Digital Surface Models) generated with lidars or stereo images for this purpose. Predicting the height of the buildings using only RGB images is challenging due to the insufficient amount of data, low data quality, variations of building types, different angles of light and shadow, etc. In this study, we present an instance segmentation-based building height extraction method to predict building masks with their respective heights from a single RGB satellite image. We used satellite images with building height annotations of certain cities along with an open-source satellite dataset with the transfer learning approach. We reached, the bounding box mAP 59, the mask mAP 52.6, and the average accuracy value of 70% for buildings belonging to each height class in our test set.

A large portion of today's world population suffer from vision impairments and wear prescription eyeglasses. However, eyeglasses causes additional bulk and discomfort when used with augmented and virtual reality headsets, thereby negatively impacting the viewer's visual experience. In this work, we remedy the usage of prescription eyeglasses in Virtual Reality (VR) headsets by shifting the optical complexity completely into software and propose a prescription-aware rendering approach for providing sharper and immersive VR imagery. To this end, we develop a differentiable display and visual perception model encapsulating display-specific parameters, color and visual acuity of human visual system and the user-specific refractive errors. Using this differentiable visual perception model, we optimize the rendered imagery in the display using stochastic gradient-descent solvers. This way, we provide prescription glasses-free sharper images for a person with vision impairments. We evaluate our approach on various displays, including desktops and VR headsets, and show significant quality and contrast improvements for users with vision impairments.