ELODI: Ensemble Logit Difference Inhibition for Positive-Congruent Training
This addresses the challenge of maintaining accuracy while reducing errors in classification system updates, which is crucial for practical deployment, though it is an incremental improvement over existing ensemble-based methods.
The paper tackles the problem of negative flips in model updates, where legacy model errors are introduced, and presents ELODI, a method that achieves state-of-the-art performance in both error rate and negative flip rate reduction at the inference cost of a single model.
Negative flips are errors introduced in a classification system when a legacy model is updated. Existing methods to reduce the negative flip rate (NFR) either do so at the expense of overall accuracy by forcing a new model to imitate the old models, or use ensembles, which multiply inference cost prohibitively. We analyze the role of ensembles in reducing NFR and observe that they remove negative flips that are typically not close to the decision boundary, but often exhibit large deviations in the distance among their logits. Based on the observation, we present a method, called Ensemble Logit Difference Inhibition (ELODI), to train a classification system that achieves paragon performance in both error rate and NFR, at the inference cost of a single model. The method distills a homogeneous ensemble to a single student model which is used to update the classification system. ELODI also introduces a generalized distillation objective, Logit Difference Inhibition (LDI), which only penalizes the logit difference of a subset of classes with the highest logit values. On multiple image classification benchmarks, model updates with ELODI demonstrate superior accuracy retention and NFR reduction.