Reconciling intuitive physics and Newtonian mechanics for colliding objects.

By Adam N. Sanborn, Vikash K. Mansinghka, T. Griffiths. 2013

219 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Developmental and Educational Psychology, General Decision Sciences

How researchers see this paper

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

In Cognitive Neuroscience

In Developmental and Educational Psychology

In General Decision Sciences

Citations in the corpus

SE3-nets: Learning rigid body motion using deep neural networks (2016, neutral): “approximate, intuitive notions of physics [4], [5].”
Rational Use of Cognitive Resources: Levels of Analysis Between the Computational and the Algorithmic (2015, methodology): “This connection establishes a direct link between levels of analysis and a simple way to deﬁne a process model that implements a Bayesian model (even in complex settings such as intuitive physics, e.g., Sanborn, Mansinghka, & Grifﬁths, 2013).”
Simulation as an engine of physical scene understanding (2013, background): “Probabilistic simulations may help explain how physical knowledge influences perceived scene layouts (26– 28), movement planning (29), causal inferences (11, 12), language semantics, and syntax (e.”
Physion: Evaluating Physical Prediction from Vision in Humans and Machines (2021, methodology): “This testing protocol is aligned with how human studies query physical prediction ability [10, 51, 8], and allows the benchmark to be agnostic to particular model design choices.”
Mind Games: Game Engines as an Architecture for Intuitive Physics. (2017, background): “…of predictions across di ﬀ erent tasks and types of reasoning, including predicting how a scene will unfold over time [6, 74], interacting with a dynamic scene [75], reasoning about underlying physical properties [10, 11, 66], causal judgments [7], and quantitative infant physical reasoning [9].”
Inferring mass in complex scenes by mental simulation. (2016, unknown): “…works well for inferring properties like mass because the laws of Newtonian physics (such as conservation of momentum) can be encoded as distributions over random variables such as velocity, where the randomness comes from perceptual uncertainty (Sanborn et al., 2009, 2013; Sanborn, 2014).”
Learning physics from dynamical scenes (2014, result): “As Sanborn et al. (2013) show, such a ‘noisy-Newtonian’ representation of intuitive physics can account for previous findings in dynamical perception that have supported a heuristic account of physical reasoning (Gilden & Proffitt, 1989; Todd & Warren, 1982), or direct perception models (e.g.…”
The scope and limits of simulation in automated reasoning (2016, background): “[54] propose that “people’s judgments [about physical events such as colliding objects] are based on optimal statistical inference over a Newtonian physical model that incorporates sensory noise and intrinsic uncertainty about the physical properties of the objects being viewed.”
The Tools Challenge: Rapid Trial-and-Error Learning in Physical Problem Solving (2019, background): “…is bound to be imperfect and uncertain, so planning must be robust to those sources of uncertainty, and revisable in adaptive ways when the agent sees that their initial plan was not successful (Battaglia, Hamrick, & Tenenbaum, 2013; Smith & Vul, 2013; Sanborn, Mansinghka, & Grifﬁths, 2013).”
Rapid trial-and-error learning with simulation supports flexible tool use and physical reasoning (2019, background): “This engine is able to simulate the world forward in time with approximately correct but stochastic dynamics (16, 17).”
A Comparative Evaluation of Approximate Probabilistic Simulation and Deep Neural Networks as Accounts of Human Physical Scene Understanding (2016, background): “…class of models has been proposed based on stored memories of experienced scenes and physical outcomes, together with pattern recognition algorithms (such as neural networks) for accessing appropriate memory items to predict outcomes in a new scene context (Sanborn et al., 2013; Sanborn, 2014).”