no code implementations • 8 Feb 2022 • Meet P. Vadera, Adam D. Cobb, Brian Jalaian, Benjamin M. Marlin
In this paper, we investigate the potential of sparse network structures to flexibly trade-off model storage costs and inference run time against predictive performance and uncertainty quantification ability.
no code implementations • 3 Dec 2021 • Meet P. Vadera, Benjamin M. Marlin
Approximate Bayesian deep learning methods hold significant promise for addressing several issues that occur when deploying deep learning components in intelligent systems, including mitigating the occurrence of over-confident errors and providing enhanced robustness to out of distribution examples.
no code implementations • 13 Jun 2021 • Meet P. Vadera, Soumya Ghosh, Kenney Ng, Benjamin M. Marlin
Bayesian decision theory provides an elegant framework for acting optimally under uncertainty when tractable posterior distributions are available.
1 code implementation • 8 Jul 2020 • Meet P. Vadera, Adam D. Cobb, Brian Jalaian, Benjamin M. Marlin
In this paper, we describe initial work on the development ofURSABench(the Uncertainty, Robustness, Scalability, and Accu-racy Benchmark), an open-source suite of bench-marking tools for comprehensive assessment of approximate Bayesian inference methods with a focus on deep learning-based classification tasks
no code implementations • 16 May 2020 • Meet P. Vadera, Brian Jalaian, Benjamin M. Marlin
In this paper, we present a general framework for distilling expectations with respect to the Bayesian posterior distribution of a deep neural network classifier, extending prior work on the Bayesian Dark Knowledge framework.
no code implementations • 7 Feb 2020 • Meet P. Vadera, Satya Narayan Shukla, Brian Jalaian, Benjamin M. Marlin
In this paper, we consider the problem of assessing the adversarial robustness of deep neural network models under both Markov chain Monte Carlo (MCMC) and Bayesian Dark Knowledge (BDK) inference approximations.
no code implementations • 4 Jun 2019 • Meet P. Vadera, Benjamin M. Marlin
Bayesian Dark Knowledge is a method for compressing the posterior predictive distribution of a neural network model into a more compact form.