Learned Prior

Learned priors represent a significant advancement in various fields by incorporating prior knowledge into models, improving performance and generalization. Current research focuses on integrating learned priors into diverse applications, including 3D scene reconstruction, robot motion planning, and image classification, often utilizing diffusion models, variational autoencoders, and deep neural networks. This approach enhances model accuracy, uncertainty estimation, and efficiency, particularly in scenarios with limited data or complex tasks, impacting fields ranging from computer vision to robotics and privacy-preserving machine learning. The ability to effectively learn and incorporate prior knowledge is proving crucial for advancing the capabilities of many machine learning systems.

Papers

October 21, 2024

FrugalNeRF: Fast Convergence for Few-shot Novel View Synthesis without Learned Priors
Chin-Yang Lin, Chung-Ho Wu, Chang-Han Yeh, Shih-Han Yen, Cheng Sun, Yu-Lun Liu
Neural Radiance Field View Synthesis Radiance Field Ground Truth Depth High Fidelity Rendering Exponential Convergence Rate Learned Prior

July 3, 2024

VEGS: View Extrapolation of Urban Scenes in 3D Gaussian Splatting using Learned Priors
Sungwon Hwang, Min-Jung Kim, Taewoong Kang, Jayeon Kang, Jaegul Choo
Gaussian Splatting Novel View Synthesis Extrapolation Framework Camera Trajectory Urban Scene Urban Scene Reconstruction Learned Prior Automotive LiDAR

August 3, 2023

Motion Planning Diffusion: Learning and Planning of Robot Motions with Diffusion Models
Joao Carvalho, An T. Le, Mark Baierl, Dorothea Koert, Jan Peters
LeArning Abstract Diffusion Model Task Planning Robot Motion Diffusion Planner Trajectory Distribution Learned Prior

July 15, 2023

Learning Expressive Priors for Generalization and Uncertainty Estimation in Neural Networks
Dominik Schnaus, Jongseok Lee, Daniel Cremers, Rudolph Triebel
Neural Network Strong Generalization Uncertainty Estimation Generalization Bound Generalization Guarantee Expressive Neural Learned Prior Non Vacuous Generalization Bound

June 8, 2023

Differentially Private Image Classification by Learning Priors from Random Processes
Xinyu Tang, Ashwinee Panda, Vikash Sehwag, Prateek Mittal
Stochastic Process Private Learning Privacy Utility Private Stochastic Gradient Descent Privacy Preserving Image Privacy Preserving Machine Learning Learned Prior

May 6, 2023

SST-ReversibleNet: Reversible-prior-based Spectral-Spatial Transformer for Efficient Hyperspectral Image Reconstruction
Zeyu Cai, Jian Yu, Ziyu Zhang, Chengqian Jin, Feipeng Da
Hyperspectral Image Reconstruction Spectral Image Learned Prior Spectral Spatial Transformer

April 9, 2023

PriorCVAE: scalable MCMC parameter inference with Bayesian deep generative modelling
Elizaveta Semenova, Prakhar Verma, Max Cairney-Leeming, Arno Solin, Samir Bhatt, Seth Flaxman
Gaussian Process Variational Autoencoders Learned Prior Bayesian Generative Model

March 25, 2023

Unsupervised Inference of Signed Distance Functions from Single Sparse Point Clouds without Learning Priors
Chao Chen, Yu-Shen Liu, Zhizhong Han
3D Point Cloud Signed Distance Function Sparse Point Cloud Unsupervised Algorithm Neural SDFs Learned Prior Unoriented Point Cloud Surface Estimation

March 20, 2023

Convergence Guarantees of Overparametrized Wide Deep Inverse Prior
Nathan Buskulic, Yvain Quéau, Jalal Fadili
Inverse Problem Convergence Guarantee Two Layer Neural Network Overparametrization Bound Deep Prior Learned Prior

July 27, 2022

Towards the Probabilistic Fusion of Learned Priors into Standard Pipelines for 3D Reconstruction
Tristan Laidlow, Jan Czarnowski, Andrea Nicastro, Ronald Clark, Stefan Leutenegger
Multi View Stereo Dense Depth Map Probabilistic Fusion Single View Depth Energy Pipeline Learned Prior

April 21, 2022

Physics vs. Learned Priors: Rethinking Camera and Algorithm Design for Task-Specific Imaging
Tzofi Klinghoffer, Siddharth Somasundaram, Kushagra Tiwary, Ramesh Raskar
Computer Vision End to End Theoretical Physic Camera Model Algorithm Design Learned Prior Imaging Task

Learned Prior

Papers

FrugalNeRF: Fast Convergence for Few-shot Novel View Synthesis without Learned Priors

VEGS: View Extrapolation of Urban Scenes in 3D Gaussian Splatting using Learned Priors

Motion Planning Diffusion: Learning and Planning of Robot Motions with Diffusion Models

Learning Expressive Priors for Generalization and Uncertainty Estimation in Neural Networks

Differentially Private Image Classification by Learning Priors from Random Processes

SST-ReversibleNet: Reversible-prior-based Spectral-Spatial Transformer for Efficient Hyperspectral Image Reconstruction

PriorCVAE: scalable MCMC parameter inference with Bayesian deep generative modelling

Unsupervised Inference of Signed Distance Functions from Single Sparse Point Clouds without Learning Priors

Convergence Guarantees of Overparametrized Wide Deep Inverse Prior

Towards the Probabilistic Fusion of Learned Priors into Standard Pipelines for 3D Reconstruction

Physics vs. Learned Priors: Rethinking Camera and Algorithm Design for Task-Specific Imaging