The scope and limits of simulation in automated reasoning

63 citations recorded by the source corpus.

How researchers see this paper

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

Learning Physical Intuition of Block Towers by Example (2016, methodology): “Compared to top-down, simulation-based models such as (Battaglia et al., 2013 Bottom-up and top-down approaches each have their advantages, and the precise combination of these systems in human reasoning is the subject of debate (e.g. (Davis & Marcus, 2016) and (Goodman et al., 2015)).”
A Comparative Evaluation of Approximate Probabilistic Simulation and Deep Neural Networks as Accounts of Human Physical Scene Understanding (2016, background): “Results and Discussions As shown in Table 1, when no geometric error or external force is added to the IPE model (σ = 0,φ = 0), its results almost always match ground-truths (94.2% accuracy).”
Commonsense reasoning about containers using radically incomplete information (2017, background): “The challenges of radically incomplete information are often far beyond the scope of existing automated reasoners based on simulation (Davis & Marcus, to appear); rather they require alternative reasoning techniques specifically designed for incomplete information.”