Withoutrequiring repeatable trials, itcanflexibly capture covariate-dependent jointSCDs, andprovide interpretable latent causes underlying the statistical dependencies between neurons.

Our tight mechanism-specific bounds outperform tight mechanism-agnostic bounds and classic group privacy results.

A common approach for private training of differentiable models, such as neural networks, is to apply first-order methods with noisy gradients.

Trusted execution environments (TEEs) canbeused to protect the content of the data during query computation, while supporting differential-private (DP) queries in TEEs provides record privacy when query output isrevealed.

Neural Information Processing Systems http://nips.cc/

This study provides a normative theory for how Bayesian causal inference can be implemented in neural circuits. In both cognitive processes such as causal reasoning and perceptual inference such ascue integration, the nervous systems need to choose different models representing the underlying causal structures when making inferences on external stimuli.

While existing approaches only model finite, discrete fidelities, in practice, the feasible fidelity choice is often infinite, which can correspond to a continuous mesh spacing orfinite element length.

Tzu-Mao Li, Michaël Gharbi, Andrew Adams, Frédo Durand, and Jonathan Ragan-Kelley. Yuanming Hu, Luke Anderson, Tzu-Mao Li, Qi Sun, Nathan Carr, Jonathan Ragan-Kelley, and Frédo Durand.