In this paper, taking the advantage of multiple source domains, we propose a novel approach for visual Domain Generalization (DG). The three key ideas underlying our formulation are (1) leveraging disentangled representations of the images to define different factors of variations, (2) generating perturbed images by changing such factors composing the representations of the images, (3) enforcing the learner (classifier) to be invariant to such changes in the images. We demonstrate the effectiveness of our approach on several widely used datasets for the domain generalization problem, on all of which we achieve competitive results with state-of-the-art models.

Despite remarkable success in a variety of computer vision applications, it is well-known that deep learning can fail catastrophically when presented with out-of-distribution data, where there are usually style differences between the training and test images. Toward addressing this challenge, we consider the domain generalization problem, wherein predictors are trained using data drawn from a family of related training (source) domains and then evaluated on a distinct and unseen test domain. Naively training a model on the aggregate set of data (pooled from all source domains) has been shown to perform suboptimally, since the information learned by that model might be domain-specific and generalizes imperfectly to test domains. Data augmentation has been shown to be an effective approach to overcome this problem. However, its application has been limited to enforcing invariance to simple transformations like rotation, brightness change, etc. Such perturbations do not necessarily cover plausible real-world variations that preserve the semantics of the input (such as a change in the image style). In this paper, taking the advantage of multiple source domains, we propose a novel approach to express and formalize robustness to these kind of real-world image perturbations. The three key ideas underlying our formulation are (1) leveraging disentangled representations of the images to define different factors of variations, (2) generating perturbed images by changing such factors composing the representations of the images, (3) enforcing the learner (classifier) to be invariant to such changes in the images. We use image-to-image translation models to demonstrate the efficacy of this approach. Based on this, we propose a domain-invariant regularization (DIR) loss function that enforces invariant prediction of targets (class labels) across domains which yields improved generalization performance. We demonstrate the effectiveness of our approach on several widely used datasets for the domain generalization problem, on all of which our results are competitive with the state-of-the-art.

Unsupervised domain adaptation (uDA) models focus on pairwise adaptation settings where there is a single, labeled, source and a single target domain. However, in many real-world settings one seeks to adapt to multiple, but somewhat similar, target domains. Applying pairwise adaptation approaches to this setting may be suboptimal, as they fail to leverage shared information among multiple domains. In this work, we propose an information theoretic approach for domain adaptation in the novel context of multiple target domains with unlabeled instances and one source domain with labeled instances. Our model aims to find a shared latent space common to all domains, while simultaneously accounting for the remaining private, domain-specific factors. Disentanglement of shared and private information is accomplished using a unified information-theoretic approach, which also serves to establish a stronger link between the latent representations and the observed data. The resulting model, accompanied by an efficient optimization algorithm, allows simultaneous adaptation from a single source to multiple target domains. We test our approach on three challenging publicly-available datasets, showing that it outperforms several popular domain adaptation methods.

For unsupervised domain adaptation, the target domain error can be provably reduced by having a shared input representation that makes the source and target domains indistinguishable from each other. Very recently it has been shown that it is not only critical to match the marginal input distributions, but also align the output class distributions. The latter can be achieved by minimizing the maximum discrepancy of predictors. In this paper, we take this principle further by proposing a more systematic and effective way to achieve hypothesis consistency using Gaussian processes (GP). The GP allows us to induce a hypothesis space of classifiers from the posterior distribution of the latent random functions, turning the learning into a large-margin posterior separation problem, significantly easier to solve than previous approaches based on adversarial minimax optimization. We formulate a learning objective that effectively influences the posterior to minimize the maximum discrepancy. This is shown to be equivalent to maximizing margins and minimizing uncertainty of the class predictions in the target domain. Empirical results demonstrate that our approach leads to state-to-the-art performance superior to existing methods on several challenging benchmarks for domain adaptation.

Subspace clustering is a common modeling paradigm used to identify constituent modes of variation in data with locally linear structure. These structures are common to many problems in computer vision, including modeling time series of complex human motion. However classical subspace clustering algorithms learn the relationships within a set of data without considering the temporal dependency and then use a separate clustering step (e.g., spectral clustering) for final segmentation. Moreover, these, frequently optimization-based, algorithms assume that all observations have complete features. In contrast in real-world applications, some features are often missing, which results in incomplete data and substantial performance degeneration of these approaches. In this paper, we propose a unified non-parametric generative framework for temporal subspace clustering to segment data drawn from a sequentially ordered union of subspaces that deals with the missing features in a principled way. The non-parametric nature of our generative model makes it possible to infer the number of subspaces and their dimension automatically from data. Experimental results on human action datasets demonstrate that the proposed model consistently outperforms other state-of-the-art subspace clustering approaches.

This paper introduces a probabilistic latent variable model to address unsupervised domain adaptation problems. Specifically, we tackle the task of categorization of visual input from different domains by learning projections from each domain to a latent (shared) space jointly with the classifier in the latent space, which simultaneously minimizes the domain disparity while maximizing the classifier's discriminative power. Furthermore, the non-parametric nature of our adaptation model makes it possible to infer the latent space dimension automatically from data. We also develop a novel regularized Variational Bayes (VB) algorithm for efficient estimation of the model parameters. We compare the proposed model with the state-of-the-art methods for the tasks of visual domain adaptation using both handcrafted and deep-net features. Our experiments show that even with a simple softmax classifier, our model outperforms several state-of-the-art methods that take advantage of more sophisticated classification schemes.

Bayesian models provide a framework for probabilistic modelling of complex datasets. Many such models are computationally demanding, especially in the presence of large datasets. In sensor network applications, statistical (Bayesian) parameter estimation usually relies on decentralized algorithms, in which both data and computation are distributed across the nodes of the network. In this paper we propose a framework for decentralized Bayesian learning using Bregman Alternating Direction Method of Multipliers (B-ADMM). We demonstrate the utility of our framework, with Mean Field Variational Bayes (MFVB) as the primitive for distributed affine structure from motion (SfM).

Hashing-based approaches have gained popularity for large-scale image retrieval in recent years. It has been shown that semi-supervised hashing, which incorporates similarity/dissimilarity information into hash function learning could improve the hashing quality. In this paper, we present a novel kernel-based semi-supervised binary hashing model for image retrieval by taking into account auxiliary information, i.e., similar and dissimilar data pairs in achieving high quality hashing. The main idea is to map the data points into a highly non-linear feature space and then map the non-linear features into compact binary codes such that similar/dissimilar data points have similar/dissimilar hash codes. Empirical evaluations on three benchmark datasets demonstrate the superiority of the proposed method over several existing unsupervised and semi-supervised hash function learning methods.