Abductive candidate retention improves learning with conflicting labels
Candidate Retention for Abductive Learning
Machine LearningArtificial Intelligence
Summary
When computers try to learn by combining neural networks with logical reasoning, they sometimes get multiple valid but conflicting explanations for the same data. The paper addresses the problem of choosing which explanations to keep for training, because picking only one can reinforce mistakes, and keeping all can dilute learning. The authors propose a method called Abductive Candidate Retention (ACR) that selectively keeps explanations by balancing uncertainty and confidence. Their experiments show this improves accuracy compared to previous methods.
What this means in practice
- •For machine learning engineers: Improve label selection for training models that combine neural networks with symbolic reasoning to reduce error from conflicting explanations.
- •For automated reasoning teams: Increase accuracy in systems that generate multiple logical explanations by selectively retaining the most useful ones during training.
Authors
Hao-Yuan He, Yu Liu, Ming Li
Abstract
Abductive learning combines neural perception with symbolic reasoning, using explanations generated by abduction to supervise the perception model. Multiple valid explanations of the same symbolic target can assign conflicting labels to the same inputs. Common policies select a single candidate as a pseudo-label, which may reinforce mistaken assignments, or weight all candidates, which may spread supervision across competing labels. These risks motivate selecting a retained subset to balance supervision sharpness and model-mass coverage. To guide this choice, we bound the coordinate-level supervision error using retained uncertainty, discarded model mass, and model mismatch. For a fixed model and training pair, only the first two terms depend on the retained set. We propose Abductive Candidate Retention (ACR), which uses these terms to guide greedy additions, accepting a candidate when its recovered mass exceeds the increase in retained uncertainty. Experiments show that ACR improves concept accuracy over single-candidate baselines and A3BL in most evaluated aggregated mod-addition settings. Objective ablations support the joint use of uncertainty and posterior mass.