US 10,650,318 B2 · Systems and methods of determining sufficient causes from multiple outcomes

Abstract

Systems and methods are disclosed for question generation to infer the most probable cause from the observable outcome and known Noisy-OR causal relations. In embodiments, the outcomes are sorted by indices according to an order including but not limited to the outcomes' natural frequency order, expert-labeled order, machine-learning derived order, etc. According to their assigned indices, in embodiments, observed outcomes with lower indices are assigned for exact inference while observed outcomes with higher indices are assigned for variational inference. In embodiments, results of exact inference and variational inference are combined to predict the most probable cause. The unique combination of exact inference and variational inference according to outcome indices makes the probable cause inferring process faster.

Claims20

Dependent claims

A computer-implemented method for determining a set of probable causes from a set of observed outcomes, the method comprising:

obtaining the set of observed outcomes comprising one or more positive observed outcomes;
ordering at least some of the positive observed outcomes from the set of observed outcomes according to a pre-assigned index, in which each positive observed outcome of the at least some of the positive observed outcomes has an associated pre-assigned index;
dividing the ordered positive observed outcomes into a first set of outcomes and a second set of outcomes based on the ordering;
determining an inferred probability for each cause of a set of causes using both the first set of outcomes and the second set of outcomes, in which the first set of outcomes is used in a variational inference process and the second set of outcomes is used in an exact inference process;
outputting the set of probable causes comprising the causes with inferred probabilities above a threshold probability value.

2 more elements

A method for determining potential causes from multiple sufficient outcomes, the method comprising:

obtaining a set of observables comprising positive observables and negative observables, the negative observables being confirmations of not observing, the set of observations being related with a set of potential causes;
determining a single negative observable probability for each negative observable;
applying each single negative observable probability to determine a first joint probability of the negative observables;
dividing at least some of the positive observables into a first group and a second group based on pre-set index values, in which each positive observable of the at least some positive observables has an associated pre-set index;
selecting exact inference for the positive observables of the first group and variational inference for the positive observables of the second group;
determining a probability of each of the positive observables of the second group, which was selected for variational inference using variational transformation;
determining a second joint probability of the positive observables of the second group, which were selected for variational inference;
determining a third joint probability of the set of observables using the second joint probability of the second group of positive observables and the first joint probability of the negative observables;
and
determining an inferred probability for each cause of the set of potential causes using the third joint probability of the set of observables.

7 more elements

Dependencies are inferred from the claim text.

Systems and methods of determining sufficient causes from multiple outcomes

Life of this patent

Systems and methods of determining sufficient causes from multiple outcomes

Life of this patent

Abstract

Claims20