Tractable Probabilistic Inference

Tractable probabilistic inference focuses on developing methods for efficiently performing probabilistic reasoning in complex systems, overcoming the computational limitations often encountered with traditional approaches. Current research emphasizes the development of novel model architectures like probabilistic circuits and sum-product-set networks, along with algorithms such as those for detecting commutative and exchangeable factors in factor graphs, to enable efficient inference in various settings, including those with tree-structured data and offline reinforcement learning. These advancements are significant because they allow for exact and scalable inference in models previously considered intractable, improving the accuracy and applicability of probabilistic methods in diverse fields like machine learning, robotics, and causal inference. The resulting improvements in efficiency and scalability are crucial for handling large-scale datasets and complex real-world problems.

Papers

September 3, 2024

Planning to avoid ambiguous states through Gaussian approximations to non-linear sensors in active inference agents
Wouter M. Kouw
Task Planning Active Inference Human Sensing Gaussian Approximation Tractable Probabilistic Inference

August 14, 2024

Sum-Product-Set Networks: Deep Tractable Models for Tree-Structured Graphs
Milan Papež, Martin Rektoris, Tomáš Pevný, Václav Šmídl
Probabilistic Circuit Structured Graph Sum Product Network Tractable Probabilistic Inference

July 23, 2024

Efficient Detection of Commutative Factors in Factor Graphs
Malte Luttermann, Johann Machemer, Marcel Gehrke
Probabilistic Graphical Model Factor Graph Efficient Detection Tractable Probabilistic Inference Probabilistic Inference Exploit Symmetry

March 15, 2024

Efficient Detection of Exchangeable Factors in Factor Graphs
Malte Luttermann, Johann Machemer, Marcel Gehrke
Probabilistic Graphical Model Factor Graph Efficient Detection Tractable Probabilistic Inference Probabilistic Inference Exploit Symmetry

October 31, 2023

A Tractable Inference Perspective of Offline RL
Xuejie Liu, Anji Liu, Guy Van den Broeck, Yitao Liang
Offline Reinforcement Learning Sequence Model Moral Tractability Tractable Probabilistic Model Tractable Probabilistic Inference

September 20, 2023

Colour Passing Revisited: Lifted Model Construction with Commutative Factors
Malte Luttermann, Tanya Braun, Ralf Möller, Marcel Gehrke
Scientific Inference Inference Algorithm Lifted Inference Tractable Probabilistic Inference Probabilistic Inference Exploit Symmetry

June 14, 2023

Scalable Neural-Probabilistic Answer Set Programming
Arseny Skryagin, Daniel Ochs, Devendra Singh Dhami, Kristian Kersting
Probabilistic Logic Probabilistic Programming Neural Predicate Tractable Probabilistic Inference

February 13, 2023

Probabilistic Circuits That Know What They Don't Know
Fabrizio Ventola, Steven Braun, Zhongjie Yu, Martin Mundt, Kristian Kersting
Model Uncertainty Probabilistic Circuit Free Uncertainty Tractable Probabilistic Inference Inference Based Adaptive Dropout

September 21, 2022

Continuous Mixtures of Tractable Probabilistic Models
Alvaro H. C. Correia, Gennaro Gala, Erik Quaeghebeur, Cassio de Campos, Robert Peharz
Mixture Model Tractable Probabilistic Model Tractable Probabilistic Inference Continuous Latent Space

July 15, 2022

Algorithms to estimate Shapley value feature attributions
Hugh Chen, Ian C. Covert, Scott M. Lundberg, Su-In Lee
Practical Algorithm Feature Attribution Model Agnostic Tractable Probabilistic Inference Shapley Value Feature

February 7, 2022

Causal Inference Using Tractable Circuits
Adnan Darwiche
Causal Inference Circuit Complexity Tractable Probabilistic Inference Parametric Causal