Modeling human intuitions about liquid flow with particle-based simulation

Humans can easily describe, imagine, and, crucially, predict a wide variety of behaviors of liquids—splashing, squirting, gushing, sloshing, soaking, dripping, draining, trickling, pooling, and pouring—despite tremendous variability in their material and dynamical properties. Here we propose and test a computational model of how people perceive and predict these liquid dynamics, based on coarse approximate simulations of fluids as collections of interacting particles. Our model is analogous to a “game engine in the head”, drawing on techniques for interactive simulations (as in video games) that optimize for efficiency and natural appearance rather than physical accuracy. In two behavioral experiments, we found that the model accurately captured people’s predictions about how liquids flow among complex solid obstacles, and was significantly better than several alternatives based on simple heuristics and deep neural networks. Our model was also able to explain how people’s predictions varied as a function of the liquids’ properties (e.g., viscosity and stickiness). Together, the model and empirical results extend the recent proposal that human physical scene understanding for the dynamics of rigid, solid objects can be supported by approximate probabilistic simulation, to the more complex and unexplored domain of fluid dynamics.

[1]  Hyeonkyeong Kim,et al.  Qualitative reasoning about fluids and mechanics , 1993 .

[2]  Takahiro Kawabe,et al.  Seeing liquids from static snapshots , 2015, Vision Research.

[3]  Jessica B. Hamrick,et al.  Inferring mass in complex scenes by mental simulation , 2016, Cognition.

[4]  Joshua B. Tenenbaum,et al.  Humans predict liquid dynamics using probabilistic simulation , 2015, CogSci.

[5]  Joshua B. Tenenbaum,et al.  Noisy Newtons: Unifying process and dependency accounts of causal attribution , 2012, CogSci.

[6]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[7]  Adam N. Sanborn Testing Bayesian and heuristic predictions of mass judgments of colliding objects , 2014, Front. Psychol..

[8]  N. J. Cook,et al.  Constructing naive theories of motion on the fly , 1994, Memory & cognition.

[9]  M. Kenward,et al.  An Introduction to the Bootstrap , 2007 .

[10]  Kevin A. Smith,et al.  Consistent physics underlying ballistic motion prediction , 2013, CogSci.

[11]  Joshua B. Tenenbaum,et al.  A Compositional Object-Based Approach to Learning Physical Dynamics , 2016, ICLR.

[12]  Alex Graves,et al.  Supervised Sequence Labelling with Recurrent Neural Networks , 2012, Studies in Computational Intelligence.

[13]  Patrick J. Hayes,et al.  The Naive Physics Manifesto , 1990, The Philosophy of Artificial Intelligence.

[14]  M. McCloskey,et al.  Intuitive physics: the straight-down belief and its origin. , 1983, Journal of experimental psychology. Learning, memory, and cognition.

[15]  Matthias Teschner,et al.  Eurographics/ Acm Siggraph Symposium on Computer Animation (2007) Weakly Compressible Sph for Free Surface Flows , 2022 .

[16]  John G. Mikhael,et al.  Functional neuroanatomy of intuitive physical inference , 2016, Proceedings of the National Academy of Sciences.

[17]  Rob Fergus,et al.  Learning Physical Intuition of Block Towers by Example , 2016, ICML.

[18]  F. Lacquaniti,et al.  Cognitive, perceptual and action-oriented representations of falling objects , 2005, Neuropsychologia.

[19]  R. Schubotz Opinion TRENDS in Cognitive Sciences Vol.11 No.5 Prediction , 2022 .

[20]  Razvan Pascanu,et al.  Interaction Networks for Learning about Objects, Relations and Physics , 2016, NIPS.

[21]  Vikash K. Mansinghka,et al.  Reconciling intuitive physics and Newtonian mechanics for colliding objects. , 2013, Psychological review.

[22]  S. T. Buckland,et al.  An Introduction to the Bootstrap , 1994 .

[23]  Kenneth D. Forbus Handbook of Knowledge Representation Edited Qualitative Modeling , 2022 .

[24]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[25]  James L. McClelland Integrating probabilistic models of perception and interactive neural networks: a historical and tutorial review , 2013, Front. Psychol..

[26]  A. Caramazza,et al.  Naive beliefs in “sophisticated” subjects: misconceptions about trajectories of objects , 1981, Cognition.

[27]  Geoffrey E. Hinton,et al.  NeuroAnimator: fast neural network emulation and control of physics-based models , 1998, SIGGRAPH.

[28]  Ernest Davis Pouring liquids: A study in commonsense physical reasoning , 2008, Artif. Intell..

[29]  Kenneth D. Forbus,et al.  Reasoning about Fluids via Molecular Collections , 1987, AAAI.

[30]  Kevin A. Smith,et al.  Sources of uncertainty in intuitive physics , 2012, CogSci.

[31]  Markus H. Gross,et al.  SPH Based Shallow Water Simulation , 2011, VRIPHYS.

[32]  Bernard Meltzer,et al.  Analogical Representations of Naive Physics , 1989, Artif. Intell..

[33]  Susan J. Hespos,et al.  PSYCHOLOGICAL SCIENCE Research Article Five-Month-Old Infants Have Different Expectations for Solids and Liquids , 2022 .

[34]  Jessica B. Hamrick,et al.  Simulation as an engine of physical scene understanding , 2013, Proceedings of the National Academy of Sciences.

[35]  Chenfanfu Jiang,et al.  Probabilistic Simulation Predicts Human Performance on Viscous Fluid-Pouring Problem , 2016, CogSci.

[36]  Matthias Teschner,et al.  Versatile surface tension and adhesion for SPH fluids , 2013, ACM Trans. Graph..

[37]  J. Jonides,et al.  Intuitive reasoning about abstract and familiar physics problems , 1986, Memory & cognition.

[38]  Kenneth D. Forbus Qualitative modeling. , 2011, Wiley interdisciplinary reviews. Cognitive science.

[39]  Ken Perlin,et al.  Accelerating Eulerian Fluid Simulation With Convolutional Networks , 2016, ICML.

[40]  Alex Graves,et al.  Neural Turing Machines , 2014, ArXiv.

[41]  H. Hecht,et al.  Influence of animation on dynamical judgments. , 1992, Journal of experimental psychology. Human perception and performance.

[42]  Tae-Yong Kim,et al.  Unified particle physics for real-time applications , 2014, ACM Trans. Graph..

[43]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[44]  Trevor Darrell,et al.  Caffe: Convolutional Architecture for Fast Feature Embedding , 2014, ACM Multimedia.

[45]  Kenneth D. Forbus,et al.  Qualitative Mental Models: Simulations or Memories? , 1997 .

[46]  Roland Memisevic,et al.  Modeling Deep Temporal Dependencies with Recurrent "Grammar Cells" , 2014, NIPS.

[47]  Chenfanfu Jiang,et al.  Consistent Probabilistic Simulation Underlying Human Judgment in Substance Dynamics , 2017, CogSci.

[48]  Susan J. Hespos,et al.  Five-Month-Old Infants Have General Knowledge of How Nonsolid Substances Behave and Interact , 2016, Psychological science.

[49]  Kazushi Maruya,et al.  Seeing liquids from visual motion , 2015, Vision Research.

[50]  Ernest Davis,et al.  The Scope and Limits of Simulation in Cognitive Models , 2015, ArXiv.

[51]  J. Monaghan Smoothed particle hydrodynamics , 2005 .

[52]  R. Hertwig,et al.  Heuristics: The Foundations of Adaptive Behavior , 2015 .

[53]  Drew H. Abney,et al.  Journal of Experimental Psychology : Human Perception and Performance Influence of Musical Groove on Postural Sway , 2015 .

[54]  M. McCloskey,et al.  Naive physics: the curvilinear impetus principle and its role in interactions with moving objects. , 1983, Journal of experimental psychology. Learning, memory, and cognition.

[55]  Ha Hong,et al.  Performance-optimized hierarchical models predict neural responses in higher visual cortex , 2014, Proceedings of the National Academy of Sciences.