Mixture Density Network

Mixture Density Networks (MDNs) are a type of neural network designed to model probability distributions, particularly those with multiple modes (peaks), offering a powerful tool for regression tasks where uncertainty is significant. Current research focuses on applying MDNs to diverse fields, including wireless communication channel modeling, astrophysics (galaxy shape recovery), robotics (kinematic modeling), and material science (nanophotonics and catalyst design), often incorporating them into larger architectures like recurrent networks or graph neural networks. The ability of MDNs to capture uncertainty and multimodality makes them valuable for applications requiring robust predictions and informed decision-making under conditions of high variability or incomplete information.

Papers

October 4, 2022

Modular Approach to Machine Reading Comprehension: Mixture of Task-Aware Experts
Anirudha Rayasam, Anusha Kamath, Gabriel Bayomi Tinoco Kalejaiye
Mixture Component Commonsense Reasoning Machine Reading Comprehension Modular Approach Mixture Density Network Expert Network Task Expert

September 15, 2022

Multi-Task Mixture Density Graph Neural Networks for Predicting Cu-based Single-Atom Alloy Catalysts for CO2 Reduction Reaction
Chen Liang, Bowen Wang, Shaogang Hao, Guangyong Chen, Pheng-Ann Heng, Xiaolong Zou
Human Prediction First Principle Mixture Density Network Catalyst Discovery Adsorption Energy

August 23, 2022

Survival Mixture Density Networks
Xintian Han, Mark Goldstein, Rajesh Ranganath
Neural ODE Mixture Density Network Deep Survival

August 2, 2022

Compound Density Networks for Risk Prediction using Electronic Health Records
Yuxi Liu, Shaowen Qin, Zhenhao Zhang, Wei Shao
Electronic Health Record Latent Variable Latent Variable Model Risk Prediction Imputation Method Knowledge Enhanced Conditional Imputation Mixture Density Network

June 7, 2022

Learning-Based Motion Planning with Mixture Density Networks
Yinghan Wang, Xiaoming Duan, Jianping He
Motion Planning Optimal Path Motion Planning Algorithm Neural Planner Mixture Density Network Optimal Motion Planning Multimodal Planning

May 3, 2022

B\'ezier Curve Gaussian Processes
Ronny Hug, Stefan Becker, Wolfgang Hübner, Michael Arens, Jürgen Beyerer
Gaussian Process Probabilistic Model Mixture Density Network Model Driven B\'ezier Curve

January 25, 2022

A deep mixture density network for outlier-corrected interpolation of crowd-sourced weather data
Charlie Kirkwood, Theo Economou, Henry Odbert, Nicolas Pugeault
Spatio Temporal Observational Data Climate Model Weather Data Mixture Density Network Crowd Sourced Data

November 29, 2021

Direct multi-modal inversion of geophysical logs using deep learning
Sergey Alyaev, Ahmed H. Elsheikh
Deep Learning Gradient Inversion Mixture Density Network

May 27, 2021

Phone-Level Prosody Modelling with GMM-Based MDN for Diverse and Controllable Speech Synthesis
Chenpeng Du, Kai Yu
Prosodic Feature Diverse Set Mixture Density Network Controllable Speech Synthesis Multi Speaker Tt