The interest in machine learning with tensor networks has been growing rapidly in recent years. The goal is to exploit tensor-structured basis functions in order to generate exponentially large feature spaces which are then used for supervised learning. We will propose two different tensor approaches for quantum-inspired machine learning. One is a kernel-based reformulation of the previously introduced MANDy, the other an alternating ridge regression in the tensor-train format. We will apply both methods to the MNIST and fashion MNIST data set and compare the results with state-of-the-art neural network-based classifiers.

Microstructures of a material form the bridge linking processing conditions - which can be controlled, to the material property - which is the primary interest in engineering applications. Thus a critical task in material design is establishing the processing-structure relationship, which requires domain expertise and techniques that can model the high-dimensional material microstructure. This work proposes a deep learning based approach that models the processing-structure relationship as a conditional image synthesis problem. In particular, we develop an auxiliary classifier Wasserstein GAN with gradient penalty (ACWGAN-GP) to synthesize microstructures under a given processing condition. This approach is free of feature engineering, requires modest domain knowledge and is applicable to a wide range of material systems. We demonstrate this approach using the ultra high carbon steel (UHCS) database, where each microstructure is annotated with a label describing the cooling method it was subjected to. Our results show that ACWGAN-GP can synthesize high-quality multiphase microstructures for a given cooling method.

Stochastic gradient descent (SGD) is the method of choice for distributed machine learning, by virtue of its light complexity per iteration on compute nodes, leading to almost linear speedups in theory. Nevertheless, such speedups are rarely observed in practice, due to high communication overheads during synchronization steps. We alleviate this problem by introducing independent subnet training: a simple, jointly model-parallel and data-parallel, approach to distributed training for fully connected, feed-forward neural networks. During subnet training, neurons are stochastically partitioned without replacement, and each partition is sent only to a single worker. This reduces the overall synchronization overhead, as each worker only receives the weights associated with the subnetwork it has been assigned to. Subnet training also reduces synchronization frequency: since workers train disjoint portions of the network, the training can proceed for long periods of time before synchronization, similar to local SGD approaches. We empirically evaluate our approach on real-world speech recognition and product recommendation applications, where we observe that subnet training i) results into accelerated training times, as compared to state of the art distributed models, and ii) often results into boosting the testing accuracy, as it implicitly combines dropout and batch normalization regularizations during training.

Prognostics and Health Management (PHM) is an emerging engineering discipline which is concerned with the analysis and prediction of equipment health and performance. One of the key challenges in PHM is to accurately predict impending failures in the equipment. In recent years, solutions for failure prediction have evolved from building complex physical models to the use of machine learning algorithms that leverage the data generated by the equipment. However, failure prediction problems pose a set of unique challenges that make direct application of traditional classification and prediction algorithms impractical. These challenges include the highly imbalanced training data, the extremely high cost of collecting more failure samples, and the complexity of the failure patterns. Traditional oversampling techniques will not be able to capture such complexity and accordingly result in overfitting the training data. This paper addresses these challenges by proposing a novel algorithm for failure prediction using Generative Adversarial Networks (GAN-FP). GAN-FP first utilizes two GAN networks to simultaneously generate training samples and build an inference network that can be used to predict failures for new samples. GAN-FP first adopts an infoGAN to generate realistic failure and non-failure samples, and initialize the weights of the first few layers of the inference network. The inference network is then tuned by optimizing a weighted loss objective using only real failure and non-failure samples. The inference network is further tuned using a second GAN whose purpose is to guarantee the consistency between the generated samples and corresponding labels. GAN-FP can be used for other imbalanced classification problems as well.

Generative Adversarial Network (GAN) and its variants serve as a perfect representation of the data generation model, providing researchers with a large amount of high-quality generated data. They illustrate a promising direction for research with limited data availability. When GAN learns the semantic-rich data distribution from a dataset, the density of the generated distribution tends to concentrate on the training data. Due to the gradient parameters of the deep neural network contain the data distribution of the training samples, they can easily remember the training samples. When GAN is applied to private or sensitive data, for instance, patient medical records, as private information may be leakage. To address this issue, we propose a Privacy-preserving Generative Adversarial Network (PPGAN) model, in which we achieve differential privacy in GANs by adding well-designed noise to the gradient during the model learning procedure. Besides, we introduced the Moments Accountant strategy in the PPGAN training process to improve the stability and compatibility of the model by controlling privacy loss. We also give a mathematical proof of the differential privacy discriminator. Through extensive case studies of the benchmark datasets, we demonstrate that PPGAN can generate high-quality synthetic data while retaining the required data available under a reasonable privacy budget.

Federated Learning (FL) is currently the most widely adopted framework for collaborative training of (deep) machine learning models under privacy constraints. Albeit it's popularity, it has been observed that Federated Learning yields suboptimal results if the local clients' data distributions diverge. To address this issue, we present Clustered Federated Learning (CFL), a novel Federated Multi-Task Learning (FMTL) framework, which exploits geometric properties of the FL loss surface, to group the client population into clusters with jointly trainable data distributions. In contrast to existing FMTL approaches, CFL does not require any modifications to the FL communication protocol to be made, is applicable to general non-convex objectives (in particular deep neural networks) and comes with strong mathematical guarantees on the clustering quality. CFL is flexible enough to handle client populations that vary over time and can be implemented in a privacy preserving way. As clustering is only performed after Federated Learning has converged to a stationary point, CFL can be viewed as a post-processing method that will always achieve greater or equal performance than conventional FL by allowing clients to arrive at more specialized models. We verify our theoretical analysis in experiments with deep convolutional and recurrent neural networks on commonly used Federated Learning datasets.

Learning network representations is a fundamental task for many graph applications such as link prediction, node classification, graph clustering, and graph visualization. Many real-world networks are interpreted as dynamic networks and evolve over time. Most existing graph embedding algorithms were developed for static graphs mainly and cannot capture the evolution of a large dynamic network. In this paper, we propose Dynamic joint V ariational Graph Au-toencoders (Dyn-VGAE) that can learn both local structures and temporal evolutionary patterns in a dynamic network. Dyn-VGAE provides a joint learning framework for computing temporal representations of all graph snapshots simultaneously. Each auto-encoder embeds a graph snapshot based on its local structure and can also learn temporal dependencies by collaborating with other autoen-coders. We conduct experimental studies on dynamic real-world graph datasets and the results demonstrate the effectiveness of the proposed method.

Deep neural networks (DNNs) have been shown to over-fit a dataset when being trained with noisy labels for a long enough time. To overcome this problem, we present a simple and effective method self-ensemble label filtering (SELF) to progressively filter out the wrong labels during training. Our method improves the task performance by gradually allowing supervision only from the potentially non-noisy (clean) labels and stops learning on the filtered noisy labels. For the filtering, we form running averages of predictions over the entire training dataset using the network output at different training epochs. We show that these ensemble estimates yield more accurate identification of inconsistent predictions throughout training than the single estimates of the network at the most recent training epoch. While filtered samples are removed entirely from the supervised training loss, we dynamically leverage them via semi-supervised learning in the unsupervised loss. We demonstrate the positive effect of such an approach on various image classification tasks under both symmetric and asymmetric label noise and at different noise ratios. It substantially outperforms all previous works on noise-aware learning across different datasets and can be applied to a broad set of network architectures. The acquisition of large quantities of a high-quality human annotation is a frequent bottleneck in applying DNNs. There are two cheap but imperfect alternatives to collect annotation at large scale: crowdsourcing from non-experts and web annotations, particularly for image data where the tags and online query keywords are treated as valid labels. Both these alternatives typically introduce noisy (wrong) labels. While Rolnick et al. (2017) empirically demonstrated that DNNs can be surprisingly robust to label noise under certain conditions, Zhang et al. (2017) has shown that DNNs have the capacity to memorize the data and will do so eventually when being confronted with too many noisy labels. Consequently, training DNNs with traditional learning procedures on noisy data strongly deteriorates their ability to generalize - a severe problem.

In this work, we present a novel upper bound of target error to address the problem for unsupervised domain adaptation. Recent studies reveal that a deep neural network can learn transferable features which generalize well to novel tasks. Furthermore, a theory proposed by Ben-David et al. (2010) provides a upper bound for target error when transferring the knowledge, which can be summarized as minimizing the source error and distance between marginal distributions simultaneously. However, common methods based on the theory usually ignore the joint error such that samples from different classes might be mixed together when matching marginal distribution. And in such case, no matter how we minimize the marginal discrepancy, the target error is not bounded due to an increasing joint error. To address this problem, we propose a general upper bound taking joint error into account, such that the undesirable case can be properly penalized. In addition, we utilize constrained hypothesis space to further formalize a tighter bound as well as a novel cross margin discrepancy to measure the dissimilarity between hypotheses which alleviates instability during adversarial learning. Extensive empirical evidence shows that our proposal outperforms related approaches in image classification error rates on standard domain adaptation benchmarks.

In recent years, machine learning has made leaps and bounds enabling applications with high recognition accuracy for speech and images. However, other types of data to which these models can be applied have not yet been explored as thoroughly. In particular, it can be relatively challenging to accurately classify single or multi-model, real-time sensor data. Labelling is an indispensable stage of data pre-processing that can be even more challenging in real-time sensor data collection. Currently, real-time sensor data labelling is an unwieldly process with limited tools available and poor performance characteristics that can lead to the performance of the machine learning models being compromised. In this paper, we introduce new techniques for labelling at the point of collection coupled with a systematic performance comparison of two popular types of Deep Neural Networks running on five custom built edge devices. These state-of-the-art edge devices are designed to enable real-time labelling with various buttons, slide potentiometer and force sensors. This research provides results and insights that can help researchers utilising edge devices for real-time data collection select appropriate labelling techniques. We also identify common bottlenecks in each architecture and provide field tested guidelines to assist developers building adaptive, high performance edge solutions.