For the robot I'm working on, we're using http://wiki.ros.org/wire http://wiki.ros.org/wire to make probabilistic world models. When the API would not only ret

For the robot I'm working on, we're using http://wiki.ros.org/wire http://wiki.ros.org/wire to make probabilistic world models. When the API would not only return a confidence score for one possible interpretation, it would be interesting to get multiple interpretations with varying confidence scores, so you can somehow determine one that makes the most sense given your world model.