Understanding bird migration across full annual cycles is critical for effective conservation. The recent BirdFlow framework allows researchers to model population-scale migration without physical tracking tags by inferring likely flight paths from weekly population snapshots (such as eBird abundance maps). However, the original framework is Markovian, meaning it assumes a bird’s next move depends only on its current location, completely ignoring its history. Because of this short-term focus, the model cannot capture essential year-long biological behaviors, most notably annual site fidelity, where birds return to the exact same breeding or wintering grounds year after year.

Mechanisms for generating differentially private synthetic data based on marginals and graphical models have been successful in a wide range of settings. However, one limitation of these methods is their inability to incorporate public data. Initializing a data generating model by pre-training on public data has shown to improve the quality of synthetic data, but this technique is not applicable when model structure is not determined a priori. We develop the mechanism JAM-PGM, which expands the adaptive measurements framework to jointly select between measuring public data and private data. This technique allows for public data to be included in a graphical-model-based mechanism. We show that JAM-PGM is able to outperform both publicly assisted and non publicly assisted synthetic data generation mechanisms even when the public data distribution is biased.

Differential privacy is the dominant standard for formal and quantifiable privacy and has been used in major deployments that impact millions of people. Many differentially private algorithms for query release and synthetic data contain steps that reconstruct answers to queries from answers to other queries that have been measured privately. Reconstruction is an important subproblem for such mechanisms to economize the privacy budget, minimize error on reconstructed answers, and allow for scalability to high-dimensional datasets. In this paper, we introduce a principled and efficient postprocessing method ReM (Residuals-to-Marginals) for reconstructing answers to marginal queries. Our method builds on recent work on efficient mechanisms for marginal query release, based on making measurements using a residual query basis that admits efficient pseudoinversion, which is an important primitive used in reconstruction. An extension GReM-LNN (Gaussian Residuals-to-Marginals with Local Non-negativity) reconstructs marginals under Gaussian noise satisfying consistency and non-negativity, which often reduces error on reconstructed answers. We demonstrate the utility of ReM and GReM-LNN by applying them to improve existing private query answering mechanisms.

The shielding of electromagnetic noise is critical in obtaining magnetic resonance imaging measurements in the ultra-low magnetic field regime where the intrinsic signal-to-noise ratio is very small. The traditional approach of using an enclosure for electromagnetic shielding is expensive and hinders system portability. We describe here the use of a CNN-based software gradiometer to suppress the effect of electromagnetic ambient background noise sources that inductively couple into the signal detection coils. The system involves three ambient noise monitoring coils placed at a distance from the magnetic resonance signal detector. The three coils were used to synthesize the ambient noise captured by the signal detector; a convolutional neural network approach was used. Mathematical foundations are provided to justify the noise suppression framework. The results show that as much as 20-fold noise suppression can be achieved using an optimized convolutional neural network and simultaneous ambient noise measurements. The proposed approach has the potential to replace the requirement for magnetically shielded enclosures and make ultra-low field magnetic resonance imaging truly portable.