On the Inference Calibration of Neural Machine Translation
This addresses inference calibration for neural machine translation, which is important for providing reliable error indicators, but it is incremental as it builds on existing label smoothing techniques.
The paper tackles the problem of miscalibration in neural machine translation during inference, where models trained with label smoothing are well-calibrated on training data but not at inference, and proposes a graduated label smoothing method that improves both calibration and translation performance, with experiments on three language pairs showing these gains.
Confidence calibration, which aims to make model predictions equal to the true correctness measures, is important for neural machine translation (NMT) because it is able to offer useful indicators of translation errors in the generated output. While prior studies have shown that NMT models trained with label smoothing are well-calibrated on the ground-truth training data, we find that miscalibration still remains a severe challenge for NMT during inference due to the discrepancy between training and inference. By carefully designing experiments on three language pairs, our work provides in-depth analyses of the correlation between calibration and translation performance as well as linguistic properties of miscalibration and reports a number of interesting findings that might help humans better analyze, understand and improve NMT models. Based on these observations, we further propose a new graduated label smoothing method that can improve both inference calibration and translation performance.