In this paper, we primarily explore the improvement of single stream audio systems using Angle of Arrival calculations in both simulation and real life gathered data. We wanted to learn how to discern the direction of an audio source from gathered signal data to ultimately incorporate into a multi modal security system. We focused on the MUSIC algorithm for the estimation of the angle of arrival but briefly experimented with other techniques such as Bartlett and Capo. We were able to implement our own MUSIC algorithm on stimulated data from Cornell. In addition, we demonstrated how we are able to calculate the angle of arrival over time in a real life scene. Finally, we are able to detect the direction of arrival for two separate and simultaneous audio sources in a real life scene. Eventually, we could incorporate this tracking into a multi modal system combined with video. Overall, we are able to produce compelling results for angle of arrival calculations that could be the stepping stones for a better system to detect events in a scene.
We develop a new model of insulin-glucose dynamics for forecasting blood glucose in type 1 diabetics. We augment an existing biomedical model by introducing time-varying dynamics driven by a machine learning sequence model. Our model maintains a physiologically plausible inductive bias and clinically interpretable parameters -- e.g., insulin sensitivity -- while inheriting the flexibility of modern pattern recognition algorithms. Critical to modeling success are the flexible, but structured representations of subject variability with a sequence model. In contrast, less constrained models like the LSTM fail to provide reliable or physiologically plausible forecasts. We conduct an extensive empirical study. We show that allowing biomedical model dynamics to vary in time improves forecasting at long time horizons, up to six hours, and produces forecasts consistent with the physiological effects of insulin and carbohydrates.
Predicting discharge medications right after a patient being admitted is an important clinical decision, which provides physicians with guidance on what type of medication regimen to plan for and what possible changes on initial medication may occur during an inpatient stay. It also facilitates medication reconciliation process with easy detection of medication discrepancy at discharge time to improve patient safety. However, since the information available upon admission is limited and patients' condition may evolve during an inpatient stay, these predictions could be a difficult decision for physicians to make. In this work, we investigate how to leverage deep learning technologies to assist physicians in predicting discharge medications based on information documented in the admission note. We build a convolutional neural network which takes an admission note as input and predicts the medications placed on the patient at discharge time. Our method is able to distill semantic patterns from unstructured and noisy texts, and is capable of capturing the pharmacological correlations among medications. We evaluate our method on 25K patient visits and compare with 4 strong baselines. Our methods demonstrate a 20% increase in macro-averaged F1 score than the best baseline.
Accessing sub-wavelength information about a scene from the far-field without invasive near-field manipulations is a fundamental challenge in wave engineering. Yet it is well understood that the dwell time of waves in complex media sets the scale for the waves' sensitivity to perturbations. Modern coded-aperture imagers leverage the degrees of freedom (DoF) offered by complex media as natural multiplexor but do not recognize and reap the fundamental difference between placing the object of interest outside or within the complex medium. Here, we show that the precision of localizing a sub-wavelength object can be improved by several orders of magnitude simply by enclosing it in its far field with a reverberant chaotic cavity. We identify deep learning as suitable noise-robust tool to extract sub-wavelength information encoded in multiplexed measurements, achieving resolutions well beyond those available in the training data. We demonstrate our finding in the microwave domain: harnessing the configurational DoF of a simple programmable metasurface, we localize a sub-wavelength object inside a chaotic cavity with a resolution of $\lambda/76$ using intensity-only single-frequency single-pixel measurements. Our results may have important applications in photoacoustic imaging as well as human-machine interaction based on reverberating elastic waves, sound or microwaves.
Generative adversarial networks (GANs) can be trained to generate 3D image data, which is useful for design optimisation. However, this conventionally requires 3D training data, which is challenging to obtain. 2D imaging techniques tend to be faster, higher resolution, better at phase identification and more widely available. Here, we introduce a generative adversarial network architecture, SliceGAN, which is able to synthesise high fidelity 3D datasets using a single representative 2D image. This is especially relevant for the task of material microstructure generation, as a cross-sectional micrograph can contain sufficient information to statistically reconstruct 3D samples. Our architecture implements the concept of uniform information density, which both ensures that generated volumes are equally high quality at all points in space, and that arbitrarily large volumes can be generated. SliceGAN has been successfully trained on a diverse set of materials, demonstrating the widespread applicability of this tool. The quality of generated micrographs is shown through a statistical comparison of synthetic and real datasets of a battery electrode in terms of key microstructural metrics. Finally, we find that the generation time for a $10^8$ voxel volume is on the order of a few seconds, yielding a path for future studies into high-throughput microstructural optimisation.
Acronyms and abbreviations are the short-form of longer phrases and they are ubiquitously employed in various types of writing. Despite their usefulness to save space in writing and reader's time in reading, they also provide challenges for understanding the text especially if the acronym is not defined in the text or if it is used far from its definition in long texts. To alleviate this issue, there are considerable efforts both from the research community and software developers to build systems for identifying acronyms and finding their correct meanings in the text. However, none of the existing works provide a unified solution capable of processing acronyms in various domains and to be publicly available. Thus, we provide the first web-based acronym identification and disambiguation system which can process acronyms from various domains including scientific, biomedical, and general domains. The web-based system is publicly available at http://iq.cs.uoregon.edu:5000 and a demo video is available at https://youtu.be/IkSh7LqI42M. The system source code is also available at https://github.com/amirveyseh/MadDog.
As an innovative solution for privacy-preserving machine learning (ML), federated learning (FL) is attracting much attention from research and industry areas. While new technologies proposed in the past few years do evolve the FL area, unfortunately, the evaluation results presented in these works fall short in integrity and are hardly comparable because of the inconsistent evaluation metrics and the lack of a common platform. In this paper, we propose a comprehensive evaluation framework for FL systems. Specifically, we first introduce the ACTPR model, which defines five metrics that cannot be excluded in FL evaluation, including Accuracy, Communication, Time efficiency, Privacy, and Robustness. Then we design and implement a benchmarking system called FedEval, which enables the systematic evaluation and comparison of existing works under consistent experimental conditions. We then provide an in-depth benchmarking study between the two most widely-used FL mechanisms, FedSGD and FedAvg. The benchmarking results show that FedSGD and FedAvg both have advantages and disadvantages under the ACTPR model. For example, FedSGD is barely influenced by the none independent and identically distributed (non-IID) data problem, but FedAvg suffers from a decline in accuracy of up to 9% in our experiments. On the other hand, FedAvg is more efficient than FedSGD regarding time consumption and communication. Lastly, we excavate a set of take-away conclusions, which are very helpful for researchers in the FL area.
In this work we propose a new supervised learning method for temporally-encoded multilayer spiking networks to perform classification. The method employs a reinforcement signal that mimics backpropagation but is far less computationally intensive. The weight update calculation at each layer requires only local data apart from this signal. We also employ a rule capable of producing specific output spike trains; by setting the target spike time equal to the actual spike time with a slight negative offset for key high-value neurons the actual spike time becomes as early as possible. In simulated MNIST handwritten digit classification, two-layer networks trained with this rule matched the performance of a comparable backpropagation based non-spiking network.
We investigate whether post-hoc model explanations are effective for diagnosing model errors--model debugging. In response to the challenge of explaining a model's prediction, a vast array of explanation methods have been proposed. Despite increasing use, it is unclear if they are effective. To start, we categorize \textit{bugs}, based on their source, into:~\textit{data, model, and test-time} contamination bugs. For several explanation methods, we assess their ability to: detect spurious correlation artifacts (data contamination), diagnose mislabeled training examples (data contamination), differentiate between a (partially) re-initialized model and a trained one (model contamination), and detect out-of-distribution inputs (test-time contamination). We find that the methods tested are able to diagnose a spurious background bug, but not conclusively identify mislabeled training examples. In addition, a class of methods, that modify the back-propagation algorithm are invariant to the higher layer parameters of a deep network; hence, ineffective for diagnosing model contamination. We complement our analysis with a human subject study, and find that subjects fail to identify defective models using attributions, but instead rely, primarily, on model predictions. Taken together, our results provide guidance for practitioners and researchers turning to explanations as tools for model debugging.
Generative adversarial networks (GANs) exhibit training pathologies that can lead to convergence-related degenerative behaviors, whereas spatially-distributed, coevolutionary algorithms (CEAs) for GAN training, e.g. Lipizzaner, are empirically robust to them. The robustness arises from diversity that occurs by training populations of generators and discriminators in each cell of a toroidal grid. Communication, where signals in the form of parameters of the best GAN in a cell propagate in four directions: North, South, West, and East, also plays a role, by communicating adaptations that are both new and fit. We propose Lipi-Ring, a distributed CEA like Lipizzaner, except that it uses a different spatial topology, i.e. a ring. Our central question is whether the different directionality of signal propagation (effectively migration to one or more neighbors on each side of a cell) meets or exceeds the performance quality and training efficiency of Lipizzaner Experimental analysis on different datasets (i.e, MNIST, CelebA, and COVID-19 chest X-ray images) shows that there are no significant differences between the performances of the trained generative models by both methods. However, Lipi-Ring significantly reduces the computational time (14.2%. . . 41.2%). Thus, Lipi-Ring offers an alternative to Lipizzaner when the computational cost of training matters.