Recent advances in Machine Learning (ML) have demonstrated that neural networks can exceed human performance in many tasks. While generalizing well over natural inputs, neural networks are vulnerable to adversarial inputs -an input that is ``similar'' to the original input, but misclassified by the model. Existing defenses focus on Lp-norm bounded adversaries that perturb ML inputs in the digital space. In the real world, however, attackers can generate adversarial perturbations that have a large Lp-norm in the digital space. Additionally, these defenses also come at a cost to accuracy, making their applicability questionable in the real world. To defend models against such a powerful adversary, we leverage one constraint on its power: the perturbation should not change the human's perception of the physical information; the physical world places some constraints on the space of possible attacks. Two questions follow: how to extract and model these constraints? and how to design a classification paradigm that leverages these constraints to improve robustness accuracy trade-off? We observe that an ML model is typically a part of a larger system with access to different input modalities. Utilizing these modalities, we introduce invariants that limit the attacker's action space. We design a hierarchical classification paradigm that enforces these invariants at inference time. As a case study, we implement and evaluate our proposal in the context of the real-world application of road sign classification because of its applicability to autonomous driving. With access to different input modalities, such as LiDAR, camera, and location we show how to extract invariants and develop a hierarchical classifier. Our results on the KITTI and GTSRB datasets show that we can improve the robustness against physical attacks at minimal harm to accuracy.

In many real-world applications, one has to consider the minimization of several loss functions simultaneously, which is, of course, an impossible mission. Therefore, one objective is chosen as the primary function to minimize, leaving the others to be bound by predefined thresholds. For example, in online portfolio selection [5], the ultimate goal is to maximize the wealth of the investor while keeping the risk bounded by a user-defined constant. In the Neyman-Pearson (NP) classification (see, e.g., [22]), an extension of the classical binary classification, the goal is to learn a classifier achieving low type-II error whose type-I error is kept below a given threshold. Another example is the online job scheduling in distributed data centers (see, e.g., [14]), in which a job router receives job tasks and schedules them to different servers to fulfill the service. Each server purchases power (within its capacity) from its zone market, used for serving the assigned jobs. Electricity market prices can vary significantly across time and zones, and the goal is to minimize the electricity cost subject to the constraint that incoming jobs must be served in time. It is indeed possible to adjust any training algorithms capable of dealing with one objective loss to deal with multiple objectives by assigning a positive weight to each loss function. However, this modification turns out to be a difficult problem, especially in the case where one has to maintain the constraints below a given threshold online.

We consider a family of problems that are concerned about making predictions for the majority of unlabeled, graph-structured data samples based on a small proportion of labeled examples. Relational information among the data samples, often encoded in the graph or network structure, is shown to be helpful for these semi-supervised learning tasks. Conventional graph-based regularization methods and recent graph neural networks do not fully leverage the interrelations between the features, the graph, and the labels. We propose a flexible generative framework for graph-based semi-supervised learning, which approaches the joint distribution of the node features, labels, and the graph structure. Borrowing insights from random graph models in network science literature, this joint distribution can be instantiated using various distribution families. For the inference of missing labels, we exploit recent advances of scalable variational inference techniques to approximate the Bayesian posterior. We conduct thorough experiments on benchmark datasets for graph-based semi-supervised learning. Results show that the proposed methods outperform state-of-the-art models under most settings.

Cross-domain recommendation has long been one of the major topics in recommender systems. Recently, various deep models have been proposed to transfer the learned knowledge across domains, but most of them focus on extracting abstract transferable features from auxilliary contents, e.g., images and review texts, and the patterns in the rating matrix itself is rarely touched. In this work, inspired by the concept of domain adaptation, we proposed a deep domain adaptation model (DARec) that is capable of extracting and transferring patterns from rating matrices {\em only} without relying on any auxillary information. We empirically demonstrate on public datasets that our method achieves the best performance among several state-of-the-art alternative cross-domain recommendation models.

Adaptive gradient approaches that automatically adjust the learning rate on a per-feature basis have been very popular for training deep networks. This rich class of algorithms includes Adagrad, RMSprop, Adam, and recent extensions. All these algorithms have adopted diagonal matrix adaptation, due to the prohibitive computational burden of manipulating full matrices in high-dimensions. In this paper, we show that block-diagonal matrix adaptation can be a practical and powerful solution that can effectively utilize structural characteristics of deep learning architectures, and significantly improve convergence and out-of-sample generalization. We present a general framework with block-diagonal matrix updates via coordinate grouping, which includes counterparts of the aforementioned algorithms, prove their convergence in non-convex optimization, highlighting benefits compared to diagonal versions. In addition, we propose an efficient spectrum-clipping scheme that benefits from superior generalization performance of Sgd. Extensive experiments reveal that block-diagonal approaches achieve state-of-the-art results on several deep learning tasks, and can outperform adaptive diagonal methods, vanilla Sgd, as well as a modified version of full-matrix adaptation proposed very recently.

Partial domain adaptation (PDA) extends standard domain adaptation to a more realistic scenario where the target domain only has a subset of classes from the source domain. The key challenge of PDA is how to select the relevant samples in the shared classes for knowledge transfer. Previous PDA methods tackle this problem by re-weighting the source samples based on the prediction of classifier or discriminator, thus discarding the pixel-level information. In this paper, to utilize both high-level and pixel-level information, we propose a reinforced transfer network (RTNet), which is the first work to apply reinforcement learning to address the PDA problem. The RTNet simultaneously mitigates the negative transfer by adopting a reinforced data selector to filter out outlier source classes, and promotes the positive transfer by employing a domain adaptation model to minimize the distribution discrepancy in the shared label space. Extensive experiments indicate that RTNet can achieve state-of-the-art performance for partial domain adaptation tasks on several benchmark datasets. Codes and datasets will be available online.

Symbol detection plays an important role in the implementation of digital receivers. In this work, we propose ViterbiNet, which is a data-driven symbol detector that does not require channel state information (CSI). ViterbiNet is obtained by integrating deep neural networks (DNNs) into the Viterbi algorithm. We identify the specific parts of the Viterbi algorithm that are channel-model-based, and design a DNN to implement only those computations, leaving the rest of the algorithm structure intact. We then propose a meta-learning based approach to train ViterbiNet online based on recent decisions, allowing the receiver to track dynamic channel conditions without requiring new training samples for every coherence block. Our numerical evaluations demonstrate that the performance of ViterbiNet, which is ignorant of the CSI, approaches that of the CSI-based Viterbi algorithm, and is capable of tracking time-varying channels without needing instantaneous CSI or additional training data. Moreover, unlike conventional Viterbi detection, ViterbiNet is robust to CSI uncertainty, and it can be reliably implemented in complex channel models with constrained computational burden. More broadly, our results demonstrate the conceptual benefit of designing communication systems to that integrate DNNs into established algorithms.

Machine learning promises methods that generalize well from finite labeled data. However, the brittleness of existing neural net approaches is revealed by notable failures, such as the existence of adversarial examples that are misclassified despite being nearly identical to a training example, or the inability of recurrent sequence-processing nets to stay on track without teacher forcing. We introduce a method, which we refer to as \emph{state reification}, that involves modeling the distribution of hidden states over the training data and then projecting hidden states observed during testing toward this distribution. Our intuition is that if the network can remain in a familiar manifold of hidden space, subsequent layers of the net should be well trained to respond appropriately. We show that this state-reification method helps neural nets to generalize better, especially when labeled data are sparse, and also helps overcome the challenge of achieving robust generalization with adversarial training.

The knowledge base completion problem is the problem of inferring missing information from existing facts in knowledge bases. Path-ranking based methods use sequences of relations as general patterns of paths for prediction. However, these patterns usually lack accuracy because they are generic and can often apply to widely varying scenarios. We leverage type hierarchies of entities to create a new class of path patterns that are both discriminative and generalizable. Then we propose an attention-based RNN model, which can be trained end-to-end, to discover the new path patterns most suitable for the data. Experiments conducted on two benchmark knowledge base completion datasets demonstrate that the proposed model outperforms existing methods by a statistically significant margin. Our quantitative analysis of the path patterns shows that they balance between generalization and discrimination.

We are interested in learning models of intuitive physics similar to the ones that animals use for navigation, manipulation and planning. In addition to learning general physical principles, however, we are also interested in learning ``on the fly'', from a few experiences, physical properties specific to new environments. We do all this in an unsupervised manner, using a meta-learning formulation where the goal is to predict videos containing demonstrations of physical phenomena, such as objects moving and colliding with a complex background. We introduce the idea of summarizing past experiences in a very compact manner, in our case using dynamic images, and show that this can be used to solve the problem well and efficiently. Empirically, we show via extensive experiments and ablation studies, that our model learns to perform physical predictions that generalize well in time and space, as well as to a variable number of interacting physical objects.