Search for: All records

Hamiltonian Monte Carlo (HMC) is a powerful algorithm to sample latent variables from Bayesian models. The advent of probabilistic programming languages (PPLs) frees users from writing inference algorithms and lets users focus on modeling. However, many models are difficult for HMC to solve directly, and often require tricks like model reparameterization. We are motivated by the fact that many of those models could be simplified by marginalization. We propose to use automatic marginalization as part of the sampling process using HMC in a graphical model extracted from a PPL, which substantially improves sampling from real-world hierarchical models. 

As billions of nocturnal avian migrants traverse North America, twice a year they must contend with landscape changes driven by natural and anthropogenic forces, including the rapid growth of the artificial glow of the night sky. While airspaces facilitate migrant passage, terrestrial landscapes serve as essential areas to restore energy reserves and often act as refugia—making it critical to holistically identify stopover locations and understand drivers of use. Here, we leverage over 10 million remote sensing observations to develop seasonal contiguous United States layers of bird migrant stopover density. In over 70% of our models, we identify skyglow as a highly influential and consistently positive predictor of bird migration stopover density across the United States. This finding points to the potential of an expanding threat to avian migrants: peri-urban illuminated areas may act as ecological traps at macroscales that increase the mortality of birds during migration.

 

Tracking technologies have widely expanded our understanding of bird migration routes, destinations and underlying strategies. However, determining the entire trajectory of small birds equipped with lightweight geolocators remains a challenge.

We develop a highly optimized hidden Markov model (HMM) for reconstructing bird trajectories. The observation model is defined by pressure and, optionally, light measurements, while the movement model incorporates wind data to constrain consecutive positions based on realistic airspeeds. To reduce the computational costs associated with a large state space, we prune the HMM states and transitions based on flight and observation constraints to efficiently model the entire trajectory.

The approach presented is based on a mathematically exact procedure and is fast to compute. We demonstrate how to compute (1) the most likely trajectory, (2) the marginal probability map of each stationary period, (3) simulated trajectories and (4) the wind conditions (wind support/drift) encountered by the bird during each migratory flight.

We construct a version of an HMM optimized for reconstructing a bird's migration trajectory based on lightweight geolocator data. To render this approach easily accessible to researchers, we designed a dedicated R packageGeoPressureR(https://raphaelnussbaumer.com/GeoPressureR/).

 

We propose the use of U-statistics to reduce variance for gradient estimation in importance-weighted variational inference. The key observation is that, given a base gradient estimator that requires m > 1 samples and a total of n > m samples to be used for estimation, lower variance is achieved by averaging the base estimator on overlapping batches of size m than disjoint batches, as currently done. We use classical U-statistic theory to analyze the variance reduction, and propose novel approximations with theoretical guarantees to ensure computational efficiency. We find empirically that U-statistic variance reduction can lead to modest to significant improvements in inference performance on a range of models, with little computational cost. 

The exodus of flying animals from their roosting locations is often visible as expanding ring‐shaped patterns in weather radar data. The NEXRAD network, for example, archives more than 25 years of data across 143 contiguous US radar stations, providing opportunities to study roosting locations and times and the ecosystems of birds and bats. However, access to this information is limited by the cost of manually annotating millions of radar scans. We develop and deploy an AI‐assisted system to annotate roosts in radar data. We build datasets with roost annotations to support the training and evaluation of automated detection models. Roosts are detected, tracked, and incorporated into our developed web‐based interface for human screening to produce research‐grade annotations. We deploy the system to collect swallow and martin roost information from 12 radar stations around the Great Lakes spanning 21 years. After verifying the practical value of the system, we propose to improve the detector by incorporating both spatial and temporal channels from volumetric radar scans. The deployment on Great Lakes radar scans allows accelerated annotation of 15 628 roost signatures in 612 786 radar scans with 183.6 human screening hours, or 1.08 s per radar scan. We estimate that the deployed system reduces human annotation time by ~7×. The temporal detector model improves the average precision at intersection‐over‐union threshold 0.5 (AP^{IoU = .50}) by 8% over the previous model (48%→56%), further reducing human screening time by 2.3× in its pilot deployment. These data contain critical information about phenology and population trends of swallows and martins, aerial insectivore species experiencing acute declines, and have enabled novel research. We present error analyses, lay the groundwork for continent‐scale historical investigation about these species, and provide a starting point for automating the detection of other family‐specific phenomena in radar data, such as bat roosts and mayfly hatches.

 

We propose AIM, a new algorithm for differentially private synthetic data generation. AIM is a workload-adaptive algorithm within the paradigm of algorithms that first selects a set of queries, then privately measures those queries, and finally generates synthetic data from the noisy measurements. It uses a set of innovative features to iteratively select the most useful measurements, reflecting both their relevance to the workload and their value in approximating the input data. We also provide analytic expressions to bound per-query error with high probability which can be used to construct confidence intervals and inform users about the accuracy of generated data. We show empirically that AIM consistently outperforms a wide variety of existing mechanisms across a variety of experimental settings.