In this work, we explore joint energy-based model (EBM) training during the finetuning of pretrained text encoders (e.g., Roberta) for natural language understanding (NLU) tasks. Our experiments show that EBM training can help the model reach a better calibration that is competitive to strong baselines, with little or no loss in accuracy. We discuss three variants of energy functions (namely scalar, hidden, and sharp-hidden) that can be defined on top of a text encoder, and compare them in experiments. Due to the discreteness of text data, we adopt noise contrastive estimation (NCE) to train the energy-based model. To make NCE training more effective, we train an auto-regressive noise model with the masked language model (MLM) objective.
@article{arxiv.2101.06829,
title = {Joint Energy-based Model Training for Better Calibrated Natural Language Understanding Models},
author = {Tianxing He and Bryan McCann and Caiming Xiong and Ehsan Hosseini-Asl},
journal= {arXiv preprint arXiv:2101.06829},
year = {2021}
}