Interactive: Manipulate parameters and interpret This leads to the Drift Diffusion Model (DDM) where evidence accumulates until reaching a stopping criterion.īy the end of this tutorial, you should be able to:ĭefine and implement the Sequential Probability Ratio Test for a series of measurementsĭefine what drift and diffusion mean in a drift-diffusion modelĮxplain the speed-accuracy trade-off in a drift diffusion modelīonus (math): derive the Drift Diffusion Model mathematically from SPRTĬode: Accumulate evidence and make a decision (DDM) We will use the Sequential Probability Ratio Test (SPRT) to infer which state is true. In Tutorial 1 we will assume that the world state is binary ( \(\pm 1\)) and constant over time, but allow for multiple observations over time. Today we will allow for dynamic world states and measurements. This produced a posterior probability distribution \(p(s|m)\). On Bayes Day, we learned how to combine the sensory measurement \(m\) about a latent variable \(s\) with our prior knowledge, using Bayes’ Theorem. Interactive Demo 2.2: Accuracy versus stop-timeīonus Section 1: DDM with fixed thresholds on confidenceīonus Coding Exercise 1.1, Coding: Simulating the DDM with fixed confidence thresholdsīonus Interactive Demo 1.2: DDM with fixed confidence thresholdīonus Coding Exercise 1.3: Speed/Accuracy Tradeoff Revisitedīonus Interactive demo 1.4: Speed/Accuracy with a threshold rule Section 2: Analyzing the DDM: accuracy vs stopping timeĬoding Exercise 2.1: The Speed/Accuracy Tradeoff Interactive Demo 1.2: Trajectories under the fixed-time stopping rule Video 2: Sequential Probability Ratio TestĬoding Exercise 1.1: Simulating an SPRT model Section 1: Sequential Probability Ratio Test as a Drift Diffusion Model Video 1: Overview of Tutorials on Hidden Dynamics Tutorial 3: Simultaneous fitting/regressionĮxample Model Project: the Train Illusion Tutorial 4: Model-Based Reinforcement Learning Tutorial 2: Learning to Act: Multi-Armed Bandits Tutorial 2: Optimal Control for Continuous State Tutorial 1: Optimal Control for Discrete States Tutorial 1: Sequential Probability Ratio Testīonus Tutorial 4: The Kalman Filter, part 2īonus Tutorial 5: Expectation Maximization for spiking neurons Tutorial 2: Bayesian inference and decisions with continuous hidden state Tutorial 1: Bayes with a binary hidden state Tutorial 3: Synaptic transmission - Models of static and dynamic synapsesīonus Tutorial: Spike-timing dependent plasticity (STDP)īonus Tutorial: Extending the Wilson-Cowan Model Tutorial 1: The Leaky Integrate-and-Fire (LIF) Neuron Model Tutorial 3: Combining determinism and stochasticity Tutorial 3: Building and Evaluating Normative Encoding Modelsīonus Tutorial: Diving Deeper into Decoding
0 Comments
Leave a Reply. |