BackgroundThe Epi-Info software suite, built and maintained by the Centers for Disease Control and Prevention (CDC), is widely used by epidemiologists and public health researchers to collect and analyze public health data, especially in the event of outbreaks such as Ebola and Zika. As it exists today, Epi-Info Desktop runs only on the Windows platform, and the larger Epi-Info Suite of products consists of separate codebases for several different devices and use-cases. Software portability has become increasingly important over the past few years as it offers a number of obvious benefits. These include reduced development time, reduced cost, and simplified system architecture. Thus, there is a blatant need for continued research. Specifically, it is critical to fully understand any underlying negative performance issues which arise from platform-agnostic systems. Such understanding should allow for improved design, and thus result in substantial mitigation of reduced performance. In this paper, we present a viable cross-platform architecture for Epi-Info which solves many of these problems.ResultsWe have successfully generated executables for Linux, Mac, and Windows from a single code-base, and we have shown that performance need not be completely sacrificed when building a cross-platform application. This has been accomplished by using Electron as a wrapper for an AngularJS app, a Python analytics module, and a local, browser-based NoSQL database.ConclusionsPromising results warrant future research. Specifically, the design allows for cross-platform form-design, data-collection, offline/online modes, scalable storage, automatic local-to-remote data sync, and fast analytics which rival more traditional approaches.

Learning a set of tasks over time, also known as continual learning (CL), is one of the most challenging problems in artificial intelligence. While recent approaches achieve some degree of CL in deep neural networks, they either (1) store a new network (or an equivalent number of parameters) for each new task, (2) store training data from previous tasks, or (3) restrict the network's ability to learn new tasks. To address these issues, we propose a novel framework, Self-Net, that uses an autoencoder to learn a set of low-dimensional representations of the weights learned for different tasks. We demonstrate that these low-dimensional vectors can then be used to generate high-fidelity recollections of the original weights. Self-Net can incorporate new tasks over time with little retraining, minimal loss in performance for older tasks, and without storing prior training data. We show that our technique achieves over 10X storage compression in a continual fashion, and that it outperforms state-of-the-art approaches on numerous datasets, including continual versions of MNIST, CIFAR10, CIFAR100, Atari, and task-incremental CORe50. To the best of our knowledge, we are the first to use autoencoders to sequentially encode sets of network weights to enable continual learning.

In many applications, data is easy to acquire but expensive and time-consuming to label, prominent examples include medical imaging and NLP. This disparity has only grown in recent years as our ability to collect data improves. Under these constraints, it makes sense to select only the most informative instances from the unlabeled pool and request an oracle (e.g., a human expert) to provide labels for those samples. The goal of active learning is to infer the informativeness of unlabeled samples so as to minimize the number of requests to the oracle. Here, we formulate active learning as an open-set recognition problem. In this paradigm, only some of the inputs belong to known classes; the classifier must identify the rest as unknown. More specifically, we leverage variational neural networks (VNNs), which produce high-confidence (i.e., low-entropy) predictions only for inputs that closely resemble the training data. We use the inverse of this confidence measure to select the samples that the oracle should label. Intuitively, unlabeled samples that the VNN is uncertain about contain features that the network has not been exposed to; thus they are more informative for future training. We carried out an extensive evaluation of our novel, probabilistic formulation of active learning, achieving state-of-the-art results on MNIST, CIFAR-10, CIFAR-100, and FashionMNIST. Additionally, unlike current active learning methods, our algorithm can learn even in the presence of out-of-distribution outliers. As our experiments show, when the unlabeled pool consists of a mixture of samples from multiple datasets, our approach can automatically distinguish between samples from seen vs. unseen datasets. Overall, our results show that high-quality uncertainty measures are key for pool-based active learning.

Neurons in real brains are enormously complex computational units. Among other things, they're responsible for transforming inbound electro-chemical vectors into outbound action potentials, updating the strengths of intermediate synapses, regulating their own internal states, and modulating the behavior of other nearby neurons. One could argue that these cells are the only things exhibiting any semblance of real intelligence. It is odd, therefore, that the machine learning community has, for so long, relied upon the assumption that this complexity can be reduced to a simple sum and fire operation. We ask, might there be some benefit to substantially increasing the computational power of individual neurons in artificial systems? To answer this question, we introduce Deep Artificial Neurons (DANs), which are themselves realized as deep neural networks. Conceptually, we embed DANs inside each node of a traditional neural network, and we connect these neurons at multiple synaptic sites, thereby vectorizing the connections between pairs of cells. We demonstrate that it is possible to meta-learn a single parameter vector, which we dub a neuronal phenotype, shared by all DANs in the network, which facilitates a meta-objective during deployment. Here, we isolate continual learning as our meta-objective, and we show that a suitable neuronal phenotype can endow a single network with an innate ability to update its synapses with minimal forgetting, using standard backpropagation, without experience replay, nor separate wake/sleep phases. We demonstrate this ability on sequential non-linear regression tasks. * first author Preprint. Under review.

Continual learning (CL) is one of the most challenging problems in artificial intelligence. While several recent approaches achieve some degree of CL in deep neural networks, they are generally marred by unscalable storage requirements, inefficient training regimes, or model saturation. In this paper, we present a scalable approach to continual learning that offers a practical solution to these problems. Motivated by the biological mechanisms responsible for consolidating knowledge and encoding experiences for long term storage, we present Self-Net, a novel framework which auto-encodes its own networks in a continual fashion. We show that a modified contractive autoencoder can efficiently integrate entire networks into a compact latent space, and we demonstrate that the latent representations can be used to generate high-fidelity recollections of their original counterparts. The result is a single, compact model capable of generating the an entire set of task-specific networks, each individually trained on a different task during the lifetime of the system. Our technique outperforms other state-of-the-art approaches on numerous datasets, including continual versions of MNIST, CIFAR10, CIFAR100, and Atari. To the best of our knowledge, we are the first to demonstrate the efficacy of using autoencoders to sequentially encode entire sets of networks in order to facilitate continual learning.