knowledge-base-1

Probabilistic Deep Learning

3260 papers • 126 benchmarks • 313 datasets

This task has no description! Would you like to contribute one?

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in probabilistic-deep-learning-3

Trend

Dataset

Best Model

Actions

No benchmarks available.

Libraries

i

Use these libraries to find probabilistic-deep-learning-3 models and implementations

Datasets

Subtasks

No subtasks available.

Most implemented papers

BreastScreening: On the Use of Multi-Modality in Medical Imaging Diagnosis

F. M. Calisto, N. Nunes, J. Nascimento•Mon Apr 06 2020

The field research, design and comparative deployment of a multimodal medical imaging user interface for breast screening is described and recommendations from the radiologists for guiding the future design of medical imaging interfaces are summarized.

44

Content

0

Paper Graph

Estimating and Evaluating Regression Predictive Uncertainty in Deep Object Detectors

Steven L. Waslander, Ali Harakeh•Tue Jan 12 2021

It is shown that in the context of object detection, training variance networks with negative log likelihood (NLL) can lead to high entropy predictive distributions regardless of the correctness of the output mean, so it is proposed to use the energy score as a non-local proper scoring rule and find that when used for training, it leads to better calibrated and lower entropy predictive distribution than NLL.

44 0

Paper Graph

Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems

P. Perdikaris, Yibo Yang•Mon Jan 14 2019

A probabilistic deep learning methodology that enables the construction of predictive data-driven surrogates for stochastic systems and can accommodate high-dimensional inputs and outputs and are able to return predictions with quantified uncertainty is presented.

62 0

Paper Graph

Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shift

Jasper Snoek, Zachary Nado, D. Sculley, Balaji Lakshminarayanan, Sebastian Nowozin, Yaniv Ovadia, Emily Fertig, Jie Jessie Ren, Joshua V. Dillon•Wed Jun 05 2019

A large-scale benchmark of existing state-of-the-art methods on classification problems and the effect of dataset shift on accuracy and calibration is presented, finding that traditional post-hoc calibration does indeed fall short, as do several other previous methods.

1891 0

Paper Graph

Detecting Model Misspecification in Amortized Bayesian Inference with Neural Networks

Marvin Schmitt, Paul-Christian Burkner, Ullrich Kothe, Stefan T. Radev•Wed Dec 15 2021

This paper proposes an augmented optimization objective which imposes a probabilistic structure on the latent data space and utilizes maximum mean discrepancy (MMD) to detect potentially catastrophic misspecifications during inference undermining the validity of the obtained results.

50 0

Paper Graph

Short-Term Density Forecasting of Low-Voltage Load Using Bernstein-Polynomial Normalizing Flows

Marcel Arpogaus, Marcus Voss, Beate Sick, Mark Nigge-Uricher, O. Dürr•Thu Apr 28 2022

D density predictions for 24h-ahead load forecasting compare favorably against Gaussian and Gaussian mixture densities and outperform a non-parametric approach based on the pinball loss, especially in low-data scenarios.

21 0

Paper Graph

A Simple Approach to Improve Single-Model Deep Uncertainty via Distance-Awareness

Jasper Snoek, Ghassen Jerfel, Zachary Nado, Balaji Lakshminarayanan, Dustin Tran, Jie Jessie Ren, Zi Lin, J. Liu, Shreyas Padhy, Yeming Wen•Sat Apr 30 2022

Spectral-normalized Neural Gaussian Process (SNGP), a simple method that improves the distance-awareness ability of modern DNNs with two simple changes: applying spectral normalization to hidden weights to enforce bi-Lipschitz smoothness in representations and replacing the last output layer with a Gaussian process layer.

68 0

Paper Graph

Deep Directional Statistics: Pose Estimation with Uncertainty Quantification

Sebastian Nowozin, Peter Gehler, S. Prokudin•Tue May 08 2018

A novel probabilistic deep learning model is proposed for the task of angular regression using von Mises distributions to predict a distribution over object pose angle and it is demonstrated on a number of challenging pose estimation datasets that the model produces calibrated probability predictions and competitive or superior point estimates compared to the current state of the art.

98 0

Paper Graph

DeepTract: A Probabilistic Deep Learning Framework for White Matter Fiber Tractography

Itay Benou, Tammy Riklin-Raviv•Tue Dec 11 2018

DeepTract, a deep-learning framework for estimating white matter fibers orientation and streamline tractography, adopts a data-driven approach for fiber reconstruction from diffusion-weighted images (DWI), which does not assume a specific diffusion model.

47 0

Paper Graph

Hybrid Models with Deep and Invertible Features

Balaji Lakshminarayanan, Y. Teh, Eric T. Nalisnick, Akihiro Matsukawa, Dilan Görür•Wed Feb 06 2019

The hybrid model, despite the invertibility constraints, achieves similar accuracy to purely predictive models, and the generative component remains a good model of the input features despite the hybrid optimization objective.

101 0

Paper Graph

Adding a benchmark result helps the community track progress.