Astounding results from transformer models on natural language tasks have intrigued the vision community to study their application to computer vision problems. This has led to exciting progress on a number of tasks while requiring minimal inductive biases in the model design. This survey aims to provide a comprehensive overview of the transformer models in the computer vision discipline and assumes little to no prior background in the field. We start with an introduction to fundamental concepts behind the success of transformer models i.e., self-supervision and self-attention. Transformer architectures leverage self-attention mechanisms to encode long-range dependencies in the input domain which makes them highly expressive. Since they assume minimal prior knowledge about the structure of the problem, self-supervision using pretext tasks is applied to pre-train transformer models on large-scale (unlabelled) datasets. The learned representations are then fine-tuned on the downstream tasks, typically leading to excellent performance due to the generalization and expressivity of encoded features. We cover extensive applications of transformers in vision including popular recognition tasks (e.g., image classification, object detection, action recognition, and segmentation), generative modeling, multi-modal tasks (e.g., visual-question answering and visual reasoning), video processing (e.g., activity recognition, video forecasting), low-level vision (e.g., image super-resolution and colorization) and 3D analysis (e.g., point cloud classification and segmentation). We compare the respective advantages and limitations of popular techniques both in terms of architectural design and their experimental value. Finally, we provide an analysis on open research directions and possible future works.

Accurate and timely air quality and weather predictions are of great importance to urban governance and human livelihood. Though many efforts have been made for air quality or weather prediction, most of them simply employ one another as feature input, which ignores the inner-connection between two predictive tasks. On the one hand, the accurate prediction of one task can help improve another task's performance. On the other hand, geospatially distributed air quality and weather monitoring stations provide additional hints for city-wide spatiotemporal dependency modeling. Inspired by the above two insights, in this paper, we propose the Multi-adversarial spatiotemporal recurrent Graph Neural Networks (MasterGNN) for joint air quality and weather predictions. Specifically, we first propose a heterogeneous recurrent graph neural network to model the spatiotemporal autocorrelation among air quality and weather monitoring stations. Then, we develop a multi-adversarial graph learning framework to against observation noise propagation introduced by spatiotemporal modeling. Moreover, we present an adaptive training strategy by formulating multi-adversarial learning as a multi-task learning problem. Finally, extensive experiments on two real-world datasets show that MasterGNN achieves the best performance compared with seven baselines on both air quality and weather prediction tasks.

Principal Component Analysis (PCA) is the workhorse tool for dimensionality reduction in this era of big data. While often overlooked, the purpose of PCA is not only to reduce data dimensionality, but also to yield features that are uncorrelated. This paper focuses on this dual objective of PCA, namely, dimensionality reduction and decorrelation of features, which requires estimating the eigenvectors of a data covariance matrix, as opposed to only estimating the subspace spanned by the eigenvectors. The ever-increasing volume of data in the modern world often requires storage of data samples across multiple machines, which precludes the use of centralized PCA algorithms. Although a few distributed solutions to the PCA problem have been proposed recently, convergence guarantees and/or communications overhead of these solutions remain a concern. With an eye towards communications efficiency, this paper introduces a feedforward neural network-based one time-scale distributed PCA algorithm termed Distributed Sanger's Algorithm (DSA) that estimates the eigenvectors of a data covariance matrix when data are distributed across an undirected and arbitrarily connected network of machines. Furthermore, the proposed algorithm is shown to converge linearly to a neighborhood of the true solution. Numerical results are also shown to demonstrate the efficacy of the proposed solution.

Given an undirected measurement graph $G = ([n], E)$, the classical angular synchronization problem consists of recovering unknown angles $\theta_1,\dots,\theta_n$ from a collection of noisy pairwise measurements of the form $(\theta_i - \theta_j) \mod 2\pi$, for each $\{i,j\} \in E$. This problem arises in a variety of applications, including computer vision, time synchronization of distributed networks, and ranking from preference relationships. In this paper, we consider a generalization to the setting where there exist $k$ unknown groups of angles $\theta_{l,1}, \dots,\theta_{l,n}$, for $l=1,\dots,k$. For each $ \{i,j\} \in E$, we are given noisy pairwise measurements of the form $\theta_{\ell,i} - \theta_{\ell,j}$ for an unknown $\ell \in \{1,2,\ldots,k\}$. This can be thought of as a natural extension of the angular synchronization problem to the heterogeneous setting of multiple groups of angles, where the measurement graph has an unknown edge-disjoint decomposition $G = G_1 \cup G_2 \ldots \cup G_k$, where the $G_i$'s denote the subgraphs of edges corresponding to each group. We propose a probabilistic generative model for this problem, along with a spectral algorithm for which we provide a detailed theoretical analysis in terms of robustness against both sampling sparsity and noise. The theoretical findings are complemented by a comprehensive set of numerical experiments, showcasing the efficacy of our algorithm under various parameter regimes. Finally, we consider an application of bi-synchronization to the graph realization problem, and provide along the way an iterative graph disentangling procedure that uncovers the subgraphs $G_i$, $i=1,\ldots,k$ which is of independent interest, as it is shown to improve the final recovery accuracy across all the experiments considered.

Multi-label ranking (MLR) is the problem of predicting and ranking multiple labels for a single instance. The predicted labels are known as the instance's labelset. MLR can be typically reduced to two sub-problems: The first is multi-label classification, where the task is to bipartite the data into relevant labels (the labelset) and irrelevant labels. The second is label ranking classification, where the task is to rank labels for each instance. A label ranking may contain ties; in the extreme case relevant labels hold a tie on first place, and irrelevant labels hold a tie on second place, thus turning the label ranking classification into a multi-label one.

In recent years, ride-hailing services have been increasingly prevalent as they provide huge convenience for passengers. As a fundamental problem, the timely prediction of passenger demands in different regions is vital for effective traffic flow control and route planning. As both spatial and temporal patterns are indispensable passenger demand prediction, relevant research has evolved from pure time series to graph-structured data for modeling historical passenger demand data, where a snapshot graph is constructed for each time slot by connecting region nodes via different relational edges (e.g., origin-destination relationship, geographical distance, etc.). Consequently, the spatiotemporal passenger demand records naturally carry dynamic patterns in the constructed graphs, where the edges also encode important information about the directions and volume (i.e., weights) of passenger demands between two connected regions. However, existing graph-based solutions fail to simultaneously consider those three crucial aspects of dynamic, directed, and weighted (DDW) graphs, leading to limited expressiveness when learning graph representations for passenger demand prediction. Therefore, we propose a novel spatiotemporal graph attention network, namely Gallat (Graph prediction with all attention) as a solution. In Gallat, by comprehensively incorporating those three intrinsic properties of DDW graphs, we build three attention layers to fully capture the spatiotemporal dependencies among different regions across all historical time slots. Moreover, the model employs a subtask to conduct pretraining so that it can obtain accurate results more quickly. We evaluate the proposed model on real-world datasets, and our experimental results demonstrate that Gallat outperforms the state-of-the-art approaches.

Given the inherent class imbalance issue within student performance datasets, samples belonging to the edges of the target class distribution pose a challenge for predictive machine learning algorithms to learn. In this paper, we introduce a general framework for synthetic embedding-based data generation (SEDG), a search-based approach to generate new synthetic samples using embeddings to correct the detriment effects of class imbalances optimally. We compare the SEDG framework to past synthetic data generation methods, including deep generative models, and traditional sampling methods. In our results, we find SEDG to outperform the traditional re-sampling methods for deep neural networks and perform competitively for common machine learning classifiers on the student performance task in several standard performance metrics.

With the prevailing of live video streaming, establishing an online pixelation method for privacy-sensitive objects is an urgency. Caused by the inaccurate detection of privacy-sensitive objects, simply migrating the tracking-by-detection structure into the online form will incur problems in target initialization, drifting, and over-pixelation. To cope with the inevitable but impacting detection issue, we propose a novel Privacy-sensitive Objects Pixelation (PsOP) framework for automatic personal privacy filtering during live video streaming. Leveraging pre-trained detection networks, our PsOP is extendable to any potential privacy-sensitive objects pixelation. Employing the embedding networks and the proposed Positioned Incremental Affinity Propagation (PIAP) clustering algorithm as the backbone, our PsOP unifies the pixelation of discriminating and indiscriminating pixelation objects through trajectories generation. In addition to the pixelation accuracy boosting, experiments on the streaming video data we built show that the proposed PsOP can significantly reduce the over-pixelation ratio in privacy-sensitive object pixelation.

Self-supervised pre-training (SSP) employs random image transformations to generate training data for visual representation learning. In this paper, we first present a modeling framework that unifies existing SSP methods as learning to predict pseudo-labels. Then, we propose new data augmentation methods of generating training examples whose pseudo-labels are harder to predict than those generated via random image transformations. Specifically, we use adversarial training and CutMix to create hard examples (HEXA) to be used as augmented views for MoCo-v2 and DeepCluster-v2, leading to two variants HEXA_{MoCo} and HEXA_{DCluster}, respectively. In our experiments, we pre-train models on ImageNet and evaluate them on multiple public benchmarks. Our evaluation shows that the two new algorithm variants outperform their original counterparts, and achieve new state-of-the-art on a wide range of tasks where limited task supervision is available for fine-tuning. These results verify that hard examples are instrumental in improving the generalization of the pre-trained models.

We propose a new framework for 2-D interpreting (features and samples) black-box machine learning models via a metamodeling technique, by which we study the output and input relationships of the underlying machine learning model. The metamodel can be estimated from data generated via a trained complex model by running the computer experiment on samples of data in the region of interest. We utilize a Gaussian process as a surrogate to capture the response surface of a complex model, in which we incorporate two parts in the process: interpolated values that are modeled by a stationary Gaussian process Z governed by a prior covariance function, and a mean function mu that captures the known trends in the underlying model. The optimization procedure for the variable importance parameter theta is to maximize the likelihood function. This theta corresponds to the correlation of individual variables with the target response. There is no need for any pre-assumed models since it depends on empirical observations. Experiments demonstrate the potential of the interpretable model through quantitative assessment of the predicted samples.