Supplemental Materials

Stimuli, data, and model predictions for the paper titled "Modeling human intuitions about liquid flow with particle-based simulation" (Bates, Yildirim, Battaglia, Tenenbaum) forthcoming in PLOS Computational Biology.<br>Content Description<br><b>Stimuli:</b>Exp1_images: Images for all stimuli in Experiment 1Exp2_images: Images for all stimuli in Experiment 2<br>Exp1_movies: Videos used for practice trials in Experiment 1Exp2_movies: Videos used for practice trials in Experiment 2<br><b>Data:</b>responses.zip: Subject responses, indexed by stimulus name, in Pickle format. These can be loaded in Python using:<br>import picklewith open([file name], 'rb') as fid: data = pickle.load(fid)<br>The resulting format will be a Python dictionary.<br><b>Model Predictions:</b>(Note: load using Python NumPy, e.g.<br>import numpy as npdata = np.load('path_to_file')<br>All files are simple, numpy arrays containing model predictions for each set of parameter values. See notes below for description of each dimension. Parameter values along a dimension always go from low to high.)<br><i>SPH ground truth:</i>exp1_grnd_water_200.npy<br>exp1_grnd_stickyhoney_100.npyexp2_grnd_water_200.npyexp2_grnd_stickyhoney_100.npyDimensions: (scenes)<br><i>SPH Intuitive Fluids Engine:<br></i>exp1_damping_alpha_0.01_0.2_2.0_sticky_N_1_5_10_15_25_50_75_100.npyexp2_damping_alpha_0.01_0.2_2.0_sticky_N_1_5_10_15_25_50_75_100.npy<br>Dimensions: (scenes, alpha, zeta, # particles)<br><i>SPH Intuitive Fluids Engine (Position Noise, from Bates et al. 2015, referenced in text):</i>exp1_position_noise_alpha_0.01_0.2_2.0_sticky_noiselevels_5_N_1_5_10_15_25_50_75_100_samples_16.npy<br>exp2_position_noise_alpha_0.01_0.2_2.0_sticky_N_noiselevels_5_1_5_10_15_25_50_75_100_samples_16.npyDimensions: (scenes, alpha, noise magnitudes, # particles, noisy samples)<br><i>SimpleSim and Gravity Heuristic:</i>exp1_heuristic.npyexp2_heuristic.npyDimensions: (scenes, g, m)<br><i>Convolutional Neural Networks:</i>convnet_exp1_water.npyconvnet_exp1_honey.npyconvnet_exp2_water.npyconvnet_exp2_honey.npyDimensions: (scenes)<br>