Probabilistic and Logical Circuits for Tractable Causal Reasoning

Channel:
United States Simons Institute for the Theory of Computing
Subscribers:
68,700
Published on ● Video Link: https://www.youtube.com/watch?v=yLnoBET8O_I


Category:
Vlog
Duration: 19:00
777 views
20

Benjie Wang (University of Oxford)
https://simons.berkeley.edu/talks/benjie-wang-university-oxford-2023-09-08
Meet the Fellows Welcome Event Fall 2023

Probabilistic circuits are a unifying class of representations of probability distributions, which admit polynomial (often linear) time exact algorithms for many probabilistic queries. However, not all information about a system can be captured through its probability distribution: in particular, causal effects cannot be expressed in purely probabilistic language. To this end, Pearl's Causal Hierarchy defines a much richer query language through the semantics of intervention. In this talk I will describe recent progress in computational aspects of causal reasoning; focusing in particular on the identification of probabilistic and logical circuit representations that admit efficient algorithms for causal queries.



Other Videos By Simons Institute for the Theory of Computing


2023-09-19Parallel Batch-Dynamic Graph Representations
2023-09-19Dynamic Algorithms for 𝑘-center on Graphs
2023-09-19A Blackbox Reduction for Adaptive Adversaries using Differential Privacy
2023-09-19Recent Progress on Sublinear Time Algorithms for Maximum Matching: Upper Bounds
2023-09-18Parallelism in Dynamic Graph Algorithms
2023-09-18Optimizing Dynamic-Graph Data Structures on Multicores with the Locality-First Strategy
2023-09-18Recent Advances in Diversity Maximization in the Offline and Composable Coreset Models
2023-09-18Approximating Edit Distance in the Fully Dynamic Model
2023-09-18Lightning Talks
2023-09-12Dynamic Graph Algorithms: What We Know and What We Don’t | Richard M. Karp Distinguished Lecture
2023-09-08Probabilistic and Logical Circuits for Tractable Causal Reasoning
2023-09-08Probabilistic Reasoning and Learning for Trustworthy AI
2023-09-08Sampling and Approximately Counting CNF Formula Solutions in the Local Lemma Regime
2023-09-08Massively Parallel Join Algorithms
2023-09-08Relational Programming
2023-09-08Title: Research Interests and Latest Works
2023-09-08What can the demand analyst learn from machine learning?
2023-09-08From Robustness to Privacy and Back
2023-09-08Decoding the Maze: New Frontiers in Achieving Nash Equilibrium in AI Architectures
2023-09-08Risk Convergence and Algorithmic Regularization of Discrete-Stepsize (Stochastic) Gradient Descent
2023-09-08Quantum Cryptography in Algorithmica


Tags:
Simons Institute
theoretical computer science
UC Berkeley
Computer Science
Theory of Computation
Theory of Computing
Meet the Fellows Welcome Event Fall 2023
Benjie Wang