通过推理时自适应优化实现语言生成中的统一去毒与去偏
Unified Detoxifying and Debiasing in Language Generation via Inference-time Adaptive Optimization
AI总结：
针对去毒与去偏彼此割裂且可能相互加剧的问题，本文提出无需训练数据的推理时自适应优化框架UDDIA，统一矫正输出空间并兼顾生成质量、效率与有效性。
摘要：
警告：本文包含表现出冒犯性和偏见的模型输出。近来，预训练语言模型（PLM）因其能够生成相当流畅的文本而在各类自然语言生成（NLG）任务中蓬勃发展。然而，人们观察到这些模型会捕捉并复现训练语料库中的有害内容，典型表现为有毒语言和社会偏见，从而引发严重的道德问题。以往关于合乎伦理的NLG的研究分别处理去毒和去偏，这存在问题，因为我们发现去偏后的模型仍表现出毒性，而去毒后的模型甚至会加剧社会偏见。为应对这一挑战，我们提出了首个统一去毒与去偏框架，名为UDDIA，该框架将这两个问题共同形式化为对输出空间的矫正。我们在理论上将该框架解释为学习一种混合加权属性的文本分布。此外，UDDIA基于参数高效微调机制，在解码期间仅对少量参数进行自适应优化，且无需任何训练数据。这使得生成质量损失最小化，并以可接受的计算成本提升了矫正性能。实验结果表明，与若干强基线相比，UDDIA同时实现了去偏和去毒，并更好地平衡了效率与有效性，朝着实用的合乎伦理的NLG迈出了进一步。
英文摘要：
Warning: this paper contains model outputs exhibiting offensiveness and biases. Recently pre-trained language models (PLMs) have prospered in various natural language generation (NLG) tasks due to their ability to generate fairly fluent text. Nevertheless, these models are observed to capture and reproduce harmful contents in training corpora, typically toxic language and social biases, raising severe moral issues. Prior works on ethical NLG tackle detoxifying and debiasing separately, which is problematic since we find debiased models still exhibit toxicity while detoxified ones even exacerbate social biases. To address such a challenge, we propose the first unified framework of detoxifying and debiasing called UDDIA, which jointly formalizes these two problems as rectifying the output space. We theoretically interpret our framework as learning a text distribution mixing weighted attributes. Besides, UDDIA conducts adaptive optimization of only a few parameters during decoding based on a parameter-efficient tuning schema without any training data. This leads to minimal generation quality loss and improved rectification performance with acceptable computational cost. Experimental results demonstrate that compared to several strong baselines, UDDIA achieves debiasing and detoxifying simultaneously and better balances efficiency and effectiveness, taking a further step towards practical ethical NLG.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
