论文标题
受控文本生成的分配方法
A Distributional Approach to Controlled Text Generation
论文作者
论文摘要
我们提出了一种分配方法，以解决预先训练的语言模型（LMS）的受控文本生成。这种方法允许在单个正式框架中指定目标LM的“点”和“分布”约束，据我们所知，这是第一个具有这样一般性的模型 - 同时最大程度地减少了KL与初始LM分布的差异。然后，最佳目标分布被唯一地确定为显式EBM（基于能量的模型）表示。然后，从该最佳表示形式中，我们通过策略梯度的自适应分配变体训练目标控制的自回旋LM。我们在圆点约束上进行了第一组实验，以显示我们方法比一组基线的优势，这是在获得受控的LM平衡约束满意度与初始LM差异的方面。然后，我们对分布约束进行实验，这是我们方法的独特特征，证明了它的潜力是对语言模型中偏见问题的补救措施。通过消融研究，我们显示了自适应技术获得更快收敛的有效性。 （代码可在https://github.com/naver/gdc上找到）
We propose a Distributional Approach for addressing Controlled Text Generation from pre-trained Language Models (LMs). This approach permits to specify, in a single formal framework, both "pointwise" and "distributional" constraints over the target LM -- to our knowledge, the first model with such generality -- while minimizing KL divergence from the initial LM distribution. The optimal target distribution is then uniquely determined as an explicit EBM (Energy-Based Model) representation. From that optimal representation we then train a target controlled Autoregressive LM through an adaptive distributional variant of Policy Gradient. We conduct a first set of experiments over pointwise constraints showing the advantages of our approach over a set of baselines, in terms of obtaining a controlled LM balancing constraint satisfaction with divergence from the initial LM. We then perform experiments over distributional constraints, a unique feature of our approach, demonstrating its potential as a remedy to the problem of Bias in Language Models. Through an ablation study, we show the effectiveness of our adaptive technique for obtaining faster convergence. (Code available at https://github.com/naver/gdc)
