Search⌘ K
Home
Courses
Hands-On Quantum Machine Learning with Python

Calculating New Parameters

Explore how to calculate new parameters in Quantum Bayesian Networks using conditional probabilities derived from measurement results. Understand the recursive training process that refines these parameters over multiple iterations to improve the model's accuracy and convergence. This lesson teaches you to implement automated parameter updates and monitor log-likelihood scores through iterative training for effective quantum inference.

We'll cover the following...

Obtaining new parameter values from the results

Javascript (babel-node)
def to_params(results):
    states = results.items()
    
    def calc_norm(ischild_val, sex_val, surv_val):
        pop = filter_states(filter_states(filter_states(states, QPOS_SEX, sex_val), QPOS_ISCHILD, ischild_val), QPOS_SURV, surv_val)
        p_norm = sum(map(lambda item: item[1], filter_states(pop, QPOS_NORM, '1')))
        p_total = sum(map(lambda item: item[1], pop))
        return p_norm / p_total
    return {
        'p_norm_cms': calc_norm('1', '0', '1'),
        'p_norm_cmd': calc_norm('1', '0', '0'),
        'p_norm_cfs': calc_norm('1', '1', '1'),
        'p_norm_cfd': calc_norm('1', '1', '0'),
        'p_norm_ams': calc_norm('0', '0', '1'),
        'p_norm_amd': calc_norm('0', '0', '0'),
        'p_norm_afs': calc_norm('0', '1', '1'),
        'p_norm_afd': calc_norm('0', '1', '0'),
    }

The to_params function takes the results as a parameter in line 1. It returns a dictionary with the probabilities of being favored by a norm, given the set of values for the variables IsChild, Sex, and Survival in lines 12 to 21. These are conditional probabilities. We first filter all the states that have the specified values for the variables in line 5. From this set, we filter those states where the Norm has value 1 (filter_states(pop, QPOS_NORM, '1') ...