from collections import Counter
from functools import partial
from multiprocessing import Pool

output = {...}
probabilities = {...}
result = {...}

def reduce(probabilities, output, key_values):
    ret = []

    key, values = key_values
    for keys in output:
        if any(keys >= value for value in values):
            ret.append((keys, probabilities[key]))

    return ret

processes = Pool(4)
tasks = processes.imap(partial(reduce, probabilities, output), result.items())

c = Counter()
for task_result in tasks:
    for key, value in task_result:
        c[key] += value

print(c)