Deep one-class classification variants for anomaly detection learn a mapping that concentrates nominal samples in feature space causing anomalies to be mapped away. Because this transformation is highly non-linear, finding interpretations poses a significant challenge. In this paper we present an explainable deep one-class classification method, Fully Convolutional Data Description (FCDD), where the mapped samples are themselves also an explanation heatmap. FCDD yields competitive detection performance and provides reasonable explanations on common anomaly detection benchmarks with CIFAR-10 and ImageNet. On MVTec-AD, a recent manufacturing dataset offering ground-truth anomaly maps, FCDD sets a new state of the art in the unsupervised setting. Our method can incorporate ground-truth anomaly maps during training and using even a few of these (~5) improves performance significantly. Finally, using FCDD's explanations we demonstrate the vulnerability of deep one-class classification models to spurious image features such as image watermarks.

Despite its high search efficiency, differential architecture search (DARTS) often selects network architectures with dominated skip connections which lead to performance degradation. However, theoretical understandings on this issue remain absent, hindering the development of more advanced methods in a principled way. In this work, we solve this problem by theoretically analyzing the effects of various types of operations, e.g. convolution, skip connection and zero operation, to the network optimization. We prove that the architectures with more skip connections can converge faster than the other candidates, and thus are selected by DARTS. This result, for the first time, theoretically and explicitly reveals the impact of skip connections to fast network optimization and its competitive advantage over other types of operations in DARTS. Then we propose a theory-inspired path-regularized DARTS that consists of two key modules: (i) a differential group-structured sparse binary gate introduced for each operation to avoid unfair competition among operations, and (ii) a path-depth-wise regularization used to incite search exploration for deep architectures that often converge slower than shallow ones as shown in our theory and are not well explored during the search. Experimental results on image classification tasks validate its advantages.

Stochastic gradient descent (SGD) and its variants are mainstream methods to train deep neural networks. Since neural networks are non-convex, more and more works study the dynamic behavior of SGD and the impact to its generalization, especially the escaping efficiency from local minima. However, these works take the over-simplified assumption that the covariance of the noise in SGD is (or can be upper bounded by) constant, although it is actually state-dependent. In this work, we conduct a formal study on the dynamic behavior of SGD with state-dependent noise. Specifically, we show that the covariance of the noise of SGD in the local region of the local minima is a quadratic function of the state. Thus, we propose a novel power-law dynamic with state-dependent diffusion to approximate the dynamic of SGD. We prove that, power-law dynamic can escape from sharp minima exponentially faster than flat minima, while the previous dynamics can only escape sharp minima polynomially faster than flat minima. Our experiments well verified our theoretical results. Inspired by our theory, we propose to add additional state-dependent noise into (large-batch) SGD to further improve its generalization ability. Experiments verify that our method is effective.

This paper investigates the use of nonparametric kernel-regression to obtain a tasksimilarity aware meta-learning algorithm. Our hypothesis is that the use of tasksimilarity helps meta-learning when the available tasks are limited and may contain outlier/ dissimilar tasks. While existing meta-learning approaches implicitly assume the tasks as being similar, it is generally unclear how this task-similarity could be quantified and used in the learning. As a result, most popular metalearning approaches do not actively use the similarity/dissimilarity between the tasks, but rely on availability of huge number of tasks for their working. Our contribution is a novel framework for meta-learning that explicitly uses task-similarity in the form of kernels and an associated meta-learning algorithm. We model the task-specific parameters to belong to a reproducing kernel Hilbert space where the kernel function captures the similarity across tasks. The proposed algorithm iteratively learns a meta-parameter which is used to assign a task-specific descriptor for every task. The task descriptors are then used to quantify the task-similarity through the kernel function. We show how our approach conceptually generalizes the popular meta-learning approaches of model-agnostic meta-learning (MAML) and Meta-stochastic gradient descent (Meta-SGD) approaches. Numerical experiments with regression tasks show that our algorithm outperforms these approaches when the number of tasks is limited, even in the presence of outlier or dissimilar tasks. This supports our hypothesis that task-similarity helps improve the metalearning performance in task-limited and adverse settings.

Language is a valuable source of clinical information in Alzheimer's Disease, as it declines concurrently with neurodegeneration. Consequently, speech and language data have been extensively studied in connection with its diagnosis. This paper summarises current findings on the use of artificial intelligence, speech and language processing to predict cognitive decline in the context of Alzheimer's Disease, detailing current research procedures, highlighting their limitations and suggesting strategies to address them. We conducted a systematic review of original research between 2000 and 2019, registered in PROSPERO (reference CRD42018116606). An interdisciplinary search covered six databases on engineering (ACM and IEEE), psychology (PsycINFO), medicine (PubMed and Embase) and Web of Science. Bibliographies of relevant papers were screened until December 2019. From 3,654 search results 51 articles were selected against the eligibility criteria. Four tables summarise their findings: study details (aim, population, interventions, comparisons, methods and outcomes), data details (size, type, modalities, annotation, balance, availability and language of study), methodology (pre-processing, feature generation, machine learning, evaluation and results) and clinical applicability (research implications, clinical potential, risk of bias and strengths/limitations). While promising results are reported across nearly all 51 studies, very few have been implemented in clinical research or practice. We concluded that the main limitations of the field are poor standardisation, limited comparability of results, and a degree of disconnect between study aims and clinical applications. Attempts to close these gaps should support translation of future research into clinical practice.

This paper describes the NEMO submission to SIGTYP 2020 shared task which deals with prediction of linguistic typological features for multiple languages using the data derived from World Atlas of Language Structures (WALS). We employ frequentist inference to represent correlations between typological features and use this representation to train simple multi-class estimators that predict individual features. We describe two submitted ridge regression-based configurations which ranked second and third overall in the constrained task. Our best configuration achieved the micro-averaged accuracy score of 0.66 on 149 test languages.

When we train a machine learning algorithm we have to be credible and prove that it is better than another model. In this article we will see the techniques to do this. How do I split the data of my set in order to make correct estimates of the performance of my ML algorithm? We will see it also with respect to the confidence limits. We will see how to evaluate two classifiers. Since in general there is no better technique than another, you have to compare the different approaches. Depending on the problem under examination, we have preferable approaches compared to others. In general we have no guarantee that one technique works better than others if we do not use statistical tests. For example, if we have a binary problem, our choice will probably fall on a binary classifier. Another important aspect is that, in many contexts, not all errors weigh in the same way, for example in the medical field saying that a sick patient is well means that the error has a higher cost than the opposite error. We initially assume constant costs for each type of error, which means not considering the costs. We will also see the cost-sensitive performance evaluation in our scheme (model). It is also possible to make a numeric evaluation, for example, to predict the performance of our processor. While the classification works only with labels or probabilities, it is also possible to evaluate a numerical value. Although in theory I could do it, if I evaluate the performance on the errors of the training set I make a mistake. Evaluation of the training set error is not a good indicator of performance on future data.

This article provides an introduction to forecasting insurance claims payouts. We focus specifically on the claims arising from weather events (events) that result in large scale destruction such as hurricanes, wildfires, floods, etc. We first provide a general overview of the traditional methodology and then discuss potential use of machine learning (ML) techniques to enhance the forecasting process. While the examples and website references provided in this article are US-centric, the ideas presented herein are general and can be applied to all locations. In other regions and countries, the analyst will need to substitute the appropriate data sources for event data.

One of the most well-established applications of machine learning is in deciding what content to show website visitors. When observation data comes from high-velocity, user-generated data streams, machine learning methods perform a balancing act between model complexity, training time, and computational costs. Furthermore, when model freshness is critical, the training of models becomes time-constrained. Parallelized batch offline training, although horizontally scalable, is often not time-considerate or cost-effective. In this paper, we propose Lambda Learner, a new framework for training models by incremental updates in response to mini-batches from data streams. We show that the resulting model of our framework closely estimates a periodically updated model trained on offline data and outperforms it when model updates are time-sensitive. We provide theoretical proof that the incremental learning updates improve the loss-function over a stale batch model. We present a large-scale deployment on the sponsored content platform for a large social network, serving hundreds of millions of users across different channels (e.g., desktop, mobile). We address challenges and complexities from both algorithms and infrastructure perspectives, and illustrate the system details for computation, storage, and streaming production of training data.

Clustering is a fundamental task in unsupervised learning, one that targets to group a dataset into clusters of similar objects. There has been recent interest in embedding normative considerations around fairness within clustering formulations. In this paper, we propose 'local connectivity' as a crucial factor in assessing membership desert in centroid clustering. We use local connectivity to refer to the support offered by the local neighborhood of an object towards supporting its membership to the cluster in question. We motivate the need to consider local connectivity of objects in cluster assignment, and provide ways to quantify local connectivity in a given clustering. We then exploit concepts from density-based clustering and devise LOFKM, a clustering method that seeks to deepen local connectivity in clustering outputs, while staying within the framework of centroid clustering. Through an empirical evaluation over real-world datasets, we illustrate that LOFKM achieves notable improvements in local connectivity at reasonable costs to clustering quality, illustrating the effectiveness of the method.