Efficient neural causal discovery
AssigneeQUALCOMM Technologies, Inc.
InventorsPhillip LIPPE, Taco Sebastiaan COHEN, Efstratios GAVVES
Life of this patent
Estimating the term…
Abstract
A method for generating a causal graph includes receiving a data set including observation data and intervention data corresponding to multiple variables. A probability distribution is determined for each variable based on the observation data. A likelihood of including each edge in the graph is computed based on the probability distribution and the intervention data. Each edge is a causal connection between variables of the multiple variables. The graph is generated based on the likelihood of including each edge. The graph may be updated by iteratively repeating the determination of the probability distribution and the computing of the likelihood of including each edge.
Claims30
A method, comprising:
- receiving a data set including observation data and intervention data corresponding to a plurality of variables;
- determining a probability distribution for each variable of the plurality of variables based on the observation data;
- computing a likelihood of including each edge in a graph based on the probability distribution and the intervention data, each edge connecting variables of the plurality of variables;
- and
- generating the graph based on the likelihood of including each edge.
An apparatus, comprising:
- a memory;
- and
- at least one processor coupled to the memory, the at least one processor being configured:
- to receive a data set including observation data and intervention data corresponding to a plurality of variables;
- to determine a probability distribution for each variable of the plurality of variables based on the observation data;
- to compute a likelihood of including each edge in a graph based on the probability distribution and the intervention data, each edge connecting variables of the plurality of variables;
- and
- to generate the graph based on the likelihood of including each edge.
An apparatus, comprising:
- means for receiving a data set including observation data and intervention data corresponding to a plurality of variables;
- means for determining a probability distribution for each variable of the plurality of variables based on the observation data;
- means for computing a likelihood of including each edge in a graph based on the probability distribution and the intervention data, each edge connecting variables of the plurality of variables;
- and
- means for generating the graph based on the likelihood of including each edge.
A non-transitory computer readable medium having encoded thereon program code, the program code being executed by a processor and comprising:
- program code to receive a data set including observation data and intervention data corresponding to a plurality of variables;
- program code to determine a probability distribution for each variable of the plurality of variables based on the observation data;
- program code to compute a likelihood of including each edge in a graph based on the probability distribution and the intervention data, each edge connecting variables of the plurality of variables;
- and
- program code to generate the graph based on the likelihood of including each edge.
Dependencies are inferred from the claim text.