论文标题
软对准目标，以适应语言生成
Soft Alignment Objectives for Robust Adaptation of Language Generation
论文作者
论文摘要
域的适应性允许生成语言模型解决其应用程序的域移动引起的特定缺陷。但是，通过进一步培训对内域数据的传统改编迅速削弱了该模型将其推广到其他领域的能力，从而使改编的模型的开放式部署容易出现错误。这项工作介绍了基于预测令牌与参考的语义相似性建立的新颖培训目标。 我们的结果表明，（1）通过从代币的语义相似性中构建训练目标来避免对单个正确预测的共同假设可以减轻域适应过程中的灾难性遗忘，而（2）保留适应性的质量，（3）具有可忽略的补充以增加计算成本。 在更广泛的背景下，基于连续令牌相似性的目标是探索高效但na \“ıve精确匹配的令牌级别的目标，表达但计算和资源密集型的顺序目标之间的中间立场。
Domain adaptation allows generative language models to address specific flaws caused by the domain shift of their application. However, the traditional adaptation by further training on in-domain data rapidly weakens the model's ability to generalize to other domains, making the open-ended deployments of the adapted models prone to errors. This work introduces novel training objectives built upon a semantic similarity of the predicted tokens to the reference. Our results show that (1) avoiding the common assumption of a single correct prediction by constructing the training target from tokens' semantic similarity can mitigate catastrophic forgetting during domain adaptation, while (2) preserving the quality of the adaptation, (3) with negligible additions to compute costs. In the broader context, the objectives grounded in a continuous token similarity pioneer the exploration of the middle ground between the efficient but na\"ıve exact-match token-level objectives and expressive but computationally- and resource-intensive sequential objectives.
