Introducing Response Time Modeling (facebookresearch#361)

Summary:
Pull Request resolved: facebookresearch#361

Implements https://arxiv.org/abs/2306.06296 (Response Time Improves Choice Prediction and Function Estimation for Gaussian Process Models of Perception and Preferences, UAI 2024).

Code to make figures etc in the paper will be available at https://github.com/facebookresearch/response-time-gps.

Differential Revision: D59561138

Author: mshvartsman

aepsych/__init__.py

@@ -11,6 +11,7 @@
 
 from . import acquisition, config, factory, generators, models, strategy, utils
 from .config import Config
+from .distributions import RTDistWithUniformLapseRate, LogNormalDDMDistribution, ShiftedGammaDDMDistribution, ShiftedInverseGammaDDMDistribution, ShiftedLogNormalDDMDistribution
 from .likelihoods import BernoulliObjectiveLikelihood
 from .models import GPClassificationModel
 from .strategy import SequentialStrategy, Strategy
@@ -31,6 +32,11 @@
     "BernoulliObjectiveLikelihood",
     "BernoulliLikelihood",
     "GaussianLikelihood",
+    "RTDistWithUniformLapseRate",
+    "LogNormalDDMDistribution",
+    "ShiftedGammaDDMDistribution",
+    "ShiftedInverseGammaDDMDistribution",
+    "ShiftedLogNormalDDMDistribution",
 ]
 
 try: