On probabilistic inference by weighted model counting

By Mark Chavira, Adnan Darwiche. 2008

486 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Management Science and Operations Research, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Management Science and Operations Research

In Signal Processing

Citations in the corpus

Logic meets Probability: Towards Explainable AI Systems for Uncertain Worlds (2017, background): “A variety of exact and approximate solver techniques have since been developed that are known to be very competitive, especially in the presence of hard constraints, deterministic reasoning and logical dependencies [Chavira and Darwiche, 2008; Ermon et al. , 2013].”
Hashing-Based Approximate Probabilistic Inference in Hybrid Domains (2015, methodology): “Weighted model counting (Chavira and Darwiche, 2008) is an extension of model counting (Gomes et al., 2009).”
Three Modern Roles for Logic in AI (2020, background): “(can be critical for the efficient computation of weighted model counts on the resulting Boolean formula). A detailed treatment of reduction techniques and various practical tradeoffs can be found in [10] and [27, Chapter 13].4 3 TRACTABLE CIRCUITS We now turn to a systematic approach for solving pr”
On the Tractability of SHAP Explanations (2020, methodology): “WMC has been extensively studied both in the theory and the AI communities, and the precise complexity of E ( F ) is known for many families of Boolean functions F .”
A Semantic Loss Function for Deep Learning with Symbolic Knowledge (2017, methodology): “Examining Deﬁnition 1 of semantic loss, we see that the right-hand side is a well-known automated reasoning task called weighted model counting (WMC) (Chavira & Darwiche, 2008; Sang et al., 2005).”
Open-Universe Weighted Model Counting (2017, methodology): “Weighted model counting (WMC) has recently emerged as an effective and general approach to probabilistic inference, offering a computational framework for encoding a variety of formalisms, including factor graphs and Bayesian networks (Choi, Kisa, and Darwiche 2013; Chavira and Darwiche 2008).”
Satisfiability modulo counting: a new approach for analyzing privacy properties (2014, unknown): “In the artiﬁcial intelligence community, some have proposed [13, 41] reducing Bayesian inference problems to weighted #SAT, but parameters are not used.”
Component Caching in Hybrid Domains with Piecewise Polynomial Densities (2016, methodology): “4Readers may observe significant similarities between this encoding scheme and that of multi-valued discrete Bayesian networks (Sang, Beame, and Kautz 2005; Chavira and Darwiche 2008).”
Importance sampling-based estimation over AND/OR search spaces for graphical models (2012, background): “Another advantage of using AND/OR sample graph mean in such networks is that it can take advantage of many implicit conditional independencies that are not elucidated by the primal graph (Chavira and Darwiche, 2008; Gogate and Domingos, 2010).”
Weighted Model Counting With Function Symbols (2017, background): “Let L FUN N be a ﬁrst-order language with function symbols { f ( x ) , . . . , g ( x , y ) , . . . , h ( x , y , z ) , . . . } , variables, equality, 2 Alternatively, a weight function could map atoms themselves to positive reals [Chavira and Darwiche, 2008].”
Reasoning about Deterministic Actions with Probabilistic Prior and Application to Stochastic Filtering (2010, methodology): “We compute the probability of a logical formula δ t over state variables at time t by applying model counting (e.g., (Gomes, Sabhar-wal, and Selman 2006; Chavira and Darwiche 2008)).”
Discriminative Learning of Sum-Product Networks (2012, background): “Sum-product networks (SPNs) [23] overcome this by exploiting context-specific independence [7] and determinism [8].”
The Symbolic Interior Point Method (2016, methodology): “ilistic inference, local structure has enabled tractability in high-treewidth models, culminating in the promising direction of using circuits and decision-diagrams for the underlying graphical model [10]. Such data structures are a more powerful representation, in admitting compositionality and reﬁ”