NeRF

We present a method that achieves state-of-the-art results for synthesizing novel views of complex scenes by optimizing an underlying continuous volumetric scene function using a sparse set of input views. Our algorithm represents a scene using a fully connected (nonconvolutional) deep network, whose input is a single continuous 5D coordinate (spatial location ( x , y , z ) and viewing direction ( θ, ϕ )) and whose output is the volume density and view-dependent emitted radiance at that spatial location. We synthesize views by querying 5D coordinates along camera rays and use classic volume rendering techniques to project the output colors and densities into an image. Because volume rendering is naturally differentiable, the only input required to optimize our representation is a set of images with known camera poses. We describe how to effectively optimize neural radiance fields to render photorealistic novel views of scenes with complicated geometry and appearance, and demonstrate results that outperform prior work on neural rendering and view synthesis.

Wind field reconstruction with adaptive random Fourier features

We investigate the use of spatial interpolation methods for reconstructing the horizontal near-surface wind field given a sparse set of measurements. In particular, random Fourier features is compared with a set of benchmark methods including kriging and inverse distance weighting. Random Fourier features is a linear model β ( x ) = ∑ k = 1 K β k e i ω k x approximating the velocity field, with randomly sampled frequencies ω k and amplitudes β k trained to minimize a loss function. We include a physically motivated divergence penalty | ∇ ⋅ β ( x ) | 2 , as well as a penalty on the Sobolev norm of β . We derive a bound on the generalization error and a sampling density that minimizes the bound. We then devise an adaptive Metropolis–Hastings algorithm for sampling the frequencies of the optimal distribution. In our experiments, our random Fourier features model outperforms the benchmark models.

Disregarding the Shoulders of Giants: Inferences from Innovation Research

Studies proposing new determinants of corporate innovation include previously identified factors in an ad hoc manner. We find that only a sparse set of recently proposed innovation determinants provide material, independent information about patents and citations. We document that inferences in recent empirical studies often change when we include previously discovered innovation determinants. Commonly used econometric methods, including fixed effects and plausible shocks, do not always mitigate the need to condition on previously identified innovation determinants. Rather than randomly selecting a subset of control variables from prior studies, our analysis offers researchers a framework to consider previously proposed variables.

Electrocardio Panorama: Synthesizing New ECG views with Self-supervision

Multi-lead electrocardiogram (ECG) provides clinical information of heartbeats from several fixed viewpoints determined by the lead positioning. However, it is often not satisfactory to visualize ECG signals in these fixed and limited views, as some clinically useful information is represented only from a few specific ECG viewpoints. For the first time, we propose a new concept, Electrocardio Panorama, which allows visualizing ECG signals from any queried viewpoints. To build Electrocardio Panorama, we assume that an underlying electrocardio field exists, representing locations, magnitudes, and directions of ECG signals. We present a Neural electrocardio field Network (Nef-Net), which first predicts the electrocardio field representation by using a sparse set of one or few input ECG views and then synthesizes Electrocardio Panorama based on the predicted representations. Specially, to better disentangle electrocardio field information from viewpoint biases, a new Angular Encoding is proposed to process viewpoint angles. Also, we propose a self-supervised learning approach called Standin Learning, which helps model the electrocardio field without direct supervision. Further, with very few modifications, Nef-Net can synthesize ECG signals from scratch. Experiments verify that our Nef-Net performs well on Electrocardio Panorama synthesis, and outperforms the previous work on the auxiliary tasks (ECG view transformation and ECG synthesis from scratch). The codes and the division labels of cardiac cycles and ECG deflections on Tianchi ECG and PTB datasets are available at https://github.com/WhatAShot/Electrocardio-Panorama.

Semi-supervised sequence modeling for improved behavioral segmentation

A popular approach to quantifying animal behavior from video data is through discrete behavioral segmentation, wherein video frames are labeled as containing one or more behavior classes such as walking or grooming. Sequence models learn to map behavioral features extracted from video frames to discrete behaviors, and both supervised and unsupervised methods are common. However, each approach has its drawbacks: supervised models require a time-consuming annotation step where humans must hand label the desired behaviors; unsupervised models may fail to accurately segment particular behaviors of interest. We introduce a semi-supervised approach that addresses these challenges by constructing a sequence model loss function with (1) a standard supervised loss that classifies a sparse set of hand labels; (2) a weakly supervised loss that classifies a set of easy-to-compute heuristic labels; and (3) a self-supervised loss that predicts the evolution of the behavioral features. With this approach, we show that a large number of unlabeled frames can improve supervised segmentation in the regime of sparse hand labels and also show that a small number of hand labeled frames can increase the precision of unsupervised segmentation.

Abstract 041: Cardiovascular Disease, Hypertension, Diabetes And Cystatin C Jointly Predict Covid-19 Mortality Alongside Age, Male Sex And Black Ethnicity

Introduction: Studies of risk factors for severe/fatal COVID-19 to date may not have identified the optimal set of informative predictors. Hypothesis: Use of penalized regression with stability analysis may identify new, sparse sets of risk factors jointly associated with COVID-19 mortality. Methods: We investigated demographic, social, lifestyle, biological (lipids, cystatin C, vitamin D), medical (comorbidities, medications) and air pollution data from UK Biobank (N=473,574) in relation to linked COVID-19 mortality, and compared with non-COVID-19 mortality. We used penalized regression models (LASSO) with stability analysis (80% selection threshold from 1,000 models with 80% subsampling) to identify a sparse set of variables associated with COVID-19 mortality. Results: Among 43 variables considered by LASSO stability selection, cardiovascular disease, hypertension, diabetes, cystatin C, age, male sex and Black ethnicity were jointly predictive of COVID-19 mortality risk at 80% selection threshold (Figure). Of these, Black ethnicity and hypertension contributed to COVID-19 but not non-COVID-19 mortality. Conclusions: Use of LASSO stability selection identified a sparse set of predictors for COVID-19 mortality including cardiovascular disease, hypertension, diabetes and cystatin C, a marker of renal function that has also been implicated in atherogenesis and inflammation. These results indicate the importance of cardiometabolic comorbidities as predisposing factors for COVID-19 mortality. Hypertension was differentially highly selected for risk of COVID-19 mortality, suggesting the need for continued vigilance with good blood pressure control during the pandemic.

Sparse balance: excitatory-inhibitory networks with small bias currents and broadly distributed synaptic weights

Cortical circuits generate excitatory currents that must be cancelled by strong inhibition to assure stability. The resulting excitatory-inhibitory (E-I) balance can generate spontaneous irregular activity but, in standard balanced E-I models, this requires that an extremely strong feedforward bias current be included along with the recurrent excitation and inhibition. The absence of experimental evidence for such large bias currents inspired us to examine an alternative regime that exhibits asynchronous activity without requiring unrealistically large feedforward input. In these networks, irregular spontaneous activity is supported by a continually changing sparse set of neurons. To support this activity, synaptic strengths must be drawn from high-variance distributions. Unlike standard balanced networks, these sparse balance networks exhibit robust nonlinear responses to uniform inputs and non-Gaussian statistics. In addition to simulations, we present a mean-field analysis to illustrate the properties of these networks.